Dated Checks and Corrections

These dated records describe what was checked at the time, the observed results, material corrections and what remained unverified. They include AI-operated website checks and product demonstrations; they do not establish independent human review. Older findings are historical, not a guarantee of current behavior. The main editorial standards explain the evidence labels and correction process.

About this archive: Existing test inputs, recorded responses, dates, access conditions and limitations are retained. Internal project counts and release bookkeeping have been separated from the public records. Moving the records did not retest the guides or update their evidence dates.

Reviewing changed details in the email builder

Reviewing changed details in the email builder

Checked October 26, 2026, Eastern time.

Copying and recovery

Review scope: complete text that works in an empty chat, verified keyboard activation, feedback and phone fallback. If a reader changes a field after building a prompt, Copy, Select and Save stay unavailable even when the prompt is edited. The reader can rebuild or update the prompt and confirm that they checked the changed details. A desktop check verified keyboard actions, exact clipboard text, a saved text file, further field changes, and Undo. Actual phone manual-copy fallback remains unverified.

Prompt builder

Review scope: a useful improvement over the fixed example for a test user and production browser/network evidence for the stated data behavior. The added review step reduces the chance of accidentally using an older prompt. It does not automatically update or verify the reader’s facts; a reader can still select text directly. A comparable reader-benefit test and network check remain open.

Clearer builder recovery and saved-file timestamps

Clearer builder recovery and saved-file timestamps

Verified October 26, 2026, Eastern time.

Clear steps and difficult controls

Review scope: numbered steps, helpful screenshots at difficult controls, relevant device differences, and a fresh-start walkthrough without unexplained clicks. The email builder now shows three starting actions, explains the default tone, and tells a reader how to paste the resulting prompt into an AI chat to request a draft. It distinguishes that chat action from sending an email. Actual phone/device differences and remaining difficult-control screenshots are still required.

Complete prompt recovery and copying

Review scope: complete instructions and source material copy into an empty chat, with keyboard activation, feedback and phone fallback verified. An AI-operated desktop check found that editing a previously generated prompt can enable copying while changes made to the fields are still absent from the prompt. The live status and nearby guidance now tell readers to include those changes themselves or rebuild. The prompt remains editable and copyable; the warning does not automatically insert the changed facts. Rebuilding, example replacement, restart and Undo were checked. Actual phone fallback and remaining exercise-use checks remain open.

Email prompt builder

Review scope: evidence of improvement over the fixed prompt for a test user, plus browser/network verification of the stated data behavior. The builder explains that loading the example replaces current work and that Undo is available before the next edit or leaving. In a same-information review, both the fixed Morgan prompt and the builder contained the essential deadline, proposed Tuesday, open question and restrictions. The builder offered structured editing and an explicit do-not-send instruction; this single AI-operated comparison does not establish a better result for a reader. No production network capture was available, so the recommendation remains open.

Date accuracy: maintained

The email-prompt, time-check and cost-check downloads now label their saved timestamps as UTC, with nearby explanations that the date may differ locally. All three files were downloaded and inspected with fictional input. The time-check page needed one reload before the updated script appeared. Original 3 was already completed; this maintenance does not add another completion.

Checks and limits: desktop keyboard example, build, changed-field warning, rebuilding, restart and Undo passed. Missing required fields announced their messages and moved focus to the first field. Complete selection and three text downloads were verified, and the builder steps remained readable at 31.875 pixels without desktop overflow. No physical phone, actual browser zoom, screen-reader, network capture or independent-reader trial was conducted.

Try the updated email prompt builder.

Clearer prompt limits and accurate update dates

Clearer prompt limits and accurate update dates

Verified October 26, 2026, Eastern time.

Prompt builder

Review scope: evidence that the builder improves a test user’s prompt over the fixed example and browser/network checks that verify the stated data behavior. The editable prompt now explains its 10,000-character limit and asks readers to check the end of a long pasted prompt. The guidance is connected to the field’s accessible description.

In an AI-operated desktop test with fictional workshop details, the built prompt preserved Wednesday, Friday at 6 p.m., pending room approval, 12 chairs, the open questions, the requested word limit and the instruction not to send. The fixed Morgan example is useful for its own facts; the builder accommodates a different situation. This was not an independent human test or a network capture, and

Copy and save a complete prompt

Review scope: complete copied exercises work in an empty chat, with keyboard activation, clear feedback and phone fallback verified. A keyboard-operated builder check copied the full 896-character fictional prompt exactly, selected the entire prompt for manual copying, and saved a text file containing the full prompt and guide link. Phone fallback and remaining exercise-use checks remain open.

Update-date accuracy: corrected

Forty guide notes and the three-assistant comparison had displayed October 27 even though those changes and tests occurred on October 26 in Eastern time. Their visible dates and relevant machine-readable dates now use October 26. Existing link anchors keep their original spelling so prior links continue to work.

All 40 guide-note changes were checked in the saved content; four representative guide pages and the comparison were inspected live. The builder’s prompt and accessible instruction were checked in the live desktop browser. No physical-phone, actual browser-zoom, screen-reader, data-transmission or independent-reader verification is claimed.

Open the prompt builder.

Three assistants, two tasks: the recorded comparison

Three assistants, two tasks: the recorded comparison

Verified October 26, 2026.

Original assistant comparison

Review scope: a small original comparison of three assistants on identical everyday tasks, retaining plans, prompts, outputs and limitations, followed by a dated recommendation that explains who should skip an option and what available access allowed.

The updated alternatives guide records an invitation and a workshop-planning task in ChatGPT, Perplexity and Gemini. All three completed both initial tasks and the invitation revision. ChatGPT and Gemini completed the planning revision; a Perplexity sign-up overlay prevented that follow-up. Eleven actual answers are retained and displayed in readable disclosures. The exact original requests and revision requests are available to copy, select or save.

ChatGPT and Perplexity were signed out; Gemini used an existing signed-in Auto session in Temporary Chat. Gemini’s standard access is an interface-based inference, and billing entitlement was not independently checked. The guide makes that limitation explicit. No paid plan was bought, no account was created, and no private documents were submitted. Copilot’s Life → Go to Chat route led to sign-in before a prompt was sent; it has no writing-quality result.

Each invitation preserved the limit and placeholders; every revised invitation met the word limit. Gemini’s initial unsupported promise was removed in its revision. All initial planning answers preserved the supplied quantities and calculated four remaining places and 60 volunteer minutes. ChatGPT and Gemini produced six-bullet revisions. The recommendation follows those limited results and includes reasons to skip an option. No general winner, paid-feature value, typical performance, measured time saving or independent human verification is claimed.

Complete exercise copying

Review scope: complete requests work in an empty chat, with keyboard activation, feedback and phone fallback verified. This release adds three complete request sources and extends the reusable comparison record. All five selections matched their source, with readonly labels, full selection, Escape recovery and focus return. Four clipboard transfers were verified; the first planning clipboard reading was stale, and an isolated keyboard retry matched all 524 characters. Four saved files contained the complete source and correct guide link or source anchor. Actual external runs now support the two tasks. Phone fallback and remaining exercise-use coverage stay open.

Accessibility and readability

50 requires keyboard and actual screen-reader checks across the core journey and forms. Five quoted planning answers now use real lists, preserving their original wording. All six result disclosures opened with Enter and closed with Space. The updated route scored 100/100 in the automated mobile accessibility check. Actual screen-reader and remaining theme/form checks are still missing.

51 requires contrast, actual browser zoom, narrow reflow and touch checks across required states. The five exercise sources remained readable at the Largest article setting (31.875 px) without horizontal overflow at the available 1363 px desktop viewport; Standard returned to 21.25 px. This is article text enlargement, not actual browser zoom. No physical-phone, narrow-viewport or screen-reader completion is claimed.

Image framing

The remaining phone-width crop/layout evidence could not be obtained with the available browser controls. Earlier desktop checks remain valid. No image edits or new device test are claimed in this release.

The white, navy and deep-green identity is preserved. Original emoji symbols are retained privately but omitted from the displayed answer transcripts, with an explicit note. No generic image or decorative card layout was added.

Complete fictional inputs and worked checks in ten guides

Complete fictional inputs and worked checks in ten guides

Verified October 26, 2026. Ten guides previously offered a practice prompt that referred to notes, data or an outline the reader still had to create. Each now supplies a complete fictional practice input, a worked answer check and copying/selection/saving controls.

Complete exercise copying

The review scope requires that copied text works in an empty chat, with keyboard activation, feedback and phone fallback verified. The new inputs contain the exercise instructions and all intended fictional source material. The budgeting example now includes its transactions, period, currency and exclusion rules; the dashboard includes a job register and week definition; the comic includes every scene needed to identify its deliberate faults. The worked checks explain what a useful result preserves and which assumptions to reject.

All ten complete selections matched the live source text. Four representative isolated keyboard copy actions and four downloaded files matched the full source and correct guide anchor: budgeting, job register, notice data and comic outline. The initial budgeting clipboard check was stale when saving immediately followed copying; a separate keyboard-copy retest matched all 2,062 characters. No new external AI run or independent reader test is claimed. Phone fallback and remaining exercise-use evidence remain open.

The arithmetic and data checks were calculated independently from the supplied samples: two unique in-week jobs, $45 in April consumption, five notice rows with four unique IDs, and the comic’s missing HALL destination and unreachable END-B. The examples and repairs are prepared teaching material, not captured AI responses, client projects or product tests.

Accessible controls and answer reveals

The review scope requires useful structure, labels, image alternatives and form-error information, with actual screen-reader verification. Each complete request opens in a labelled read-only selection field with instructions and expanded state. Enter activates selection; Close and Escape return focus to the selection control. Every worked check opens with Enter and closes with Space through a native disclosure control.

All ten guides passed those desktop interactions and scored 100/100 in automated mobile accessibility checks. Actual screen-reader output, the previously recorded theme-generated search/comment issues and remaining form-error coverage are still required. Automated results do not close those gaps.

Readability, contrast and enlargement

The review scope requires actual contrast and state checks, enlarged text, browser zoom, narrow-screen reflow and usable touch spacing. All ten long practice inputs remained within the tested desktop page at the Largest text setting, 31.875px. The representative notice-data check measured text/control contrast from 8.01:1 to 13.98:1. Its open answer reveal measured 12.62:1, with a visible 3px focus outline at 7.23:1 against the background.

These are desktop and built-in text-size checks. Actual browser zoom, phone-width reflow, touch behavior and the remaining site-wide state coverage stay open.

Ten changed guides

  1. Start an AI Side Hustle Without an Audience — Supplied an invented groomer’s notes and a check for unsupported policies. Copy, selection and saving controls cover the complete input.
  2. Turn Work Experience Into a Useful AI Service — Supplied handover notes and a check separating agreed tasks from purchasing authority. Copy, selection and saving controls cover the complete input.
  3. Use AI to Examine a Decision From Different Angles — Supplied both event options and a check separating known availability from unknown preferences. Copy, selection and saving controls cover the complete input.
  4. Could AI Help You Earn From a Hobby? — Supplied a card-pack product sheet and a check for unsupported product or delivery claims. Copy, selection and saving controls cover the complete input.
  5. Keep Learning Opportunities When Your Team Uses AI — Supplied source notes and a flawed draft, with an explained repair. Copy, selection and saving controls cover the complete input.
  6. Could Your Data Dashboard Become a Useful Product? — Supplied notice rows and field definitions, with duplicate, missing-date and category checks. Copy, selection and saving controls cover the complete input.
  7. Build a Dashboard Service Clients Can Trust — Supplied the job register and counting rule, with the two-job answer and exclusions. Copy, selection and saving controls cover the complete input.
  8. AI Jingles and Taglines: Start With a Clear Brief — Supplied a complete tagline brief and a check for unsupported promises and actual timing. Copy, selection and saving controls cover the complete input.
  9. Prototype a Budgeting App Without Bank Data — Supplied transactions, period, currency and handling rules, with the $45 baseline and correction tests. Copy, selection and saving controls cover the complete input.
  10. Create an Interactive Comic With AI — Supplied a complete faulty comic outline and a check for missing routes and unsupported knowledge. Copy, selection and saving controls cover the complete input.

Earlier publication and source-review dates, page addresses, excerpts, photographs and unrelated article content were preserved. New change notes describe the new teaching material and controls without implying a new product or source review. White, navy and deep green, readable type and restrained borders retain the site’s visual identity. No decorative emoji, generic AI image or personal publisher identity was added.

Complete practice requests in 30 guides

Complete practice requests in 30 guides

Verified October 26, 2026. Thirty guide interfaces now provide copying, manual selection, text-file saving and article text-size choices for 31 complete practice requests.

Complete, usable exercise text

The review scope requires all exercise instructions and fictional source material, selectable text and manual-copy guidance, with empty-chat usefulness, keyboard activation, feedback and phone fallback verified. Each selected request was checked for the information it actually needs. The two family-care requests include their example notes or draft. The request-tracker selection includes its preceding workshop scenario. Blank-organizer requests explicitly leave personal entries empty. The budgeting-app prompt was excluded because the reader-created sample dataset and handling rules are missing.

All 31 selected blocks passed live exact manual-selection checks. Four representative keyboard copy actions matched the complete request text, and all four downloaded text files contained that exact text and the correct source anchor. These covered a care-meeting exercise, a scenario-plus-prompt brief, a blank household worksheet and a concept brief. The first mouse clipboard read in the care-meeting check was stale; the keyboard retest matched the source. No external AI output or independent reader test is claimed. Phone fallback and remaining exercise-use evidence are still needed.

Accessible controls

The review scope includes useful labels, structure, image alternatives and form-error information, verified with an actual screen reader. Each new selection control opens a labelled read-only text field, exposes the correct expanded state and instructions, and selects the complete text. Close and Escape both return keyboard focus to the selection control. Reopening preserves the source.

All 30 guides passed those live desktop checks. Six representative pages scored 100/100 in automated mobile accessibility checks: family care, lab questions, the request tracker, pendant concepts, the emergency worksheet and painted-pot design. These results do not replace an actual screen-reader test or resolve the previously recorded theme-generated search and comment-field issues and remaining form-error coverage.

Readability and enlargement

The review scope requires measured contrast and states, enlarged text, actual browser zoom, narrow-screen reflow and usable touch spacing. All 30 guides now offer Standard, Larger and Largest article text. Largest rendered the request text at 31.875px without horizontal page overflow at the tested desktop width; Standard restored 21.25px. This is built-in text enlargement, not browser zoom. Actual zoom, narrow-width and touch checks remain open.

Changed guides

Thirty guide improvements; each retains its original request wording
GuideComplete selections enabled
Use AI to Prepare a Family Care Meeting2
Using AI After a Hospital Stay1
Use AI to Organize Care for an Aging Parent1
Start a Useful Newsletter With AI1
Use AI to Organize Your Blood Pressure Log1
Track Menopause Symptoms Without Letting AI Diagnose1
Use AI to Prepare Questions About Lab Results1
Practice a Negotiation With AI Role-Play1
Offer a Research Brief People Can Check1
Test a Small No-Code App Before Charging1
Make a Nature Audio Script with AI1
Plan a Small Party With AI1
Use AI to Improve a Handmade Etsy Product1
AI for Coaches: Draft, Review, Discuss1
Make an Audio Story Children Can Choose1
Write Better Comedy With AI1
Plan Holiday Decorations With AI1
Plan a Family Santa Workshop With AI Help1
Make a Family Christmas Song With AI1
Turn an AI Jewelry Idea Into a Useful Brief1
Use AI to Prepare a Better Game Night1
Practice Difficult Conversations with AI1
AI Fragrance Tools: Preference Is Not Treatment1
Use AI to Prepare Questions for a Lawyer1
Use AI to Organize Chronic Care1
Use AI to Reconnect With an Old Friend1
Prepare for an Estate Planning Conversation1
Use AI for Dating Without Pretending1
Use AI to Organize Your Home Emergency Plan1
Use AI to Cut Unwanted Phone Time1

The original request wording, prior editorial/source dates, publication dates, addresses and excerpts were preserved. Each new date note describes only this website-control update. The existing white, navy and deep-green identity remains; no decorative emoji, generic AI imagery or personal publisher identity was added.

Clearer exercise selection and reliable keyboard return

Clearer exercise selection and reliable keyboard return

Verified October 26, 2026. None is newly complete.

Copy the complete exercise

The original requirement is to copy all exercise instructions and fictional source material, preserve selectable text and manual-copy guidance, and verify empty-chat usefulness, keyboard activation, feedback and phone fallback. Five guides already had complete exercises and manual selection. Their selection panels now have explicit Close actions, Escape recovery and returned keyboard focus; missing state and instruction associations were added. Their exercise text was preserved.

The affected guides cover useful context, comparing options, summarizing a document, planning a week and checking an AI answer. Live keyboard checks selected the exact complete text in each: 345, 561, 636, 386 and 793 characters respectively. Close and Escape collapsed the panels and returned focus correctly. Native disclosure controls kept the external button state synchronized where present.

Test an AI Tool Before You Pay gained copying, manual selection and text saving for its complete action-list exercise and blank trial record. Both 391-character and 245-character sources passed exact selection checks. Clipboard transfer matched both sources; the first immediate exercise clipboard read was empty and a deliberate recheck matched all 391 characters. Both downloads retained the full source and correct guide anchor. The trial-record message directs readers to their own notes.

Phone fallback and the remaining exercise-use evidence are still required. No new external AI-product test or independent reader test is claimed.

Accessible controls and structure

The review scope includes useful labels, structure, image alternatives and error information, verified with an actual screen reader. The revised controls expose their target panel and expanded state, associate copying instructions with the text field, distinguish selection from copying and provide a predictable keyboard return. Reopening retained the complete source. Existing weekly-planning and claim-checking exercises still worked and reset correctly.

All six changed pages scored 100 out of 100 in automated mobile accessibility checks. These automated checks and desktop keyboard tests do not substitute for screen-reader output, complete form-error coverage or the unresolved theme-generated search and comment-field issues.

Readability, enlargement and reflow

The review scope requires contrast and control checks, enlarged text, actual browser zoom, narrow-screen reflow and usable touch spacing. The tool-testing guide now includes the shared reading-and-print view. Both source blocks remained visible and complete in that view; returning restored the prior disclosure states. Its Largest text option rendered both sources at 31.875 pixels with no horizontal page overflow at the tested desktop width. This was a built-in text-size check, not browser zoom.

The existing whole-task calculator still produced the correct fictional results: 22 minutes for the first attempt and 17 minutes for the later-attempt scenario. Actual zoom, phone-width and touch checks remain open.

Privacy at collection points: reviewed, still open

A public-comment privacy explanation is still needed beside the theme’s comment form, together with the previously outstanding provider and retention confirmations. Related articles separate the editable article ending from the comment form on the checked route. No misplaced notice or unsupported theme-code workaround was added, and no contact or comment form was submitted.

Saved article bodies matched the prepared changes exactly. Original publication dates, exercise content, page addresses and excerpts were preserved. The white, navy and deep-green identity remains, and no personal publisher identity, decorative emojis or generic AI imagery was added.

Larger article lists and uncropped photographs

Larger article lists and uncropped photographs

Verified October 26, 2026.

Consistent images without cutting off the subject

The requirement is to preserve important faces, hands and activities, with consistent image treatment at desktop and phone widths. Seven topic lists now use 400-pixel desktop photo frames instead of 320 pixels. All eight lists, including Latest Articles, use a consistent 400 by 267-pixel desktop frame, contained photographs and restrained 4-pixel corners. Contain keeps the whole source photograph visible.

An existing article-image rule now contains the photograph instead of cropping it. All 20 retained photographs affected by that selector passed live desktop checks; 19 received the new treatment and one retained its stronger existing rule. Native settings and the two style changes matched the prepared values exactly. Phone-width visual checks across layout types remain required before completion.

Readable text, enlargement and reflow

The completion requirement covers measured text and control contrast, enlarged text, browser zoom, narrow-screen reflow, tap spacing and cues that do not depend on color alone. All eight article lists now display 30-pixel desktop title links, 20-pixel summaries and 16-pixel dates. The existing summary rule was corrected so it no longer overrides the larger native text setting.

Live desktop checks covered 96 list rows with no document horizontal overflow. All 35 retained list photographs loaded without cropping, and 61 removed-photo placeholders occupied no space. Representative title, summary and date contrast against white measured 14.05:1, 7.01:1 and 6.38:1; the observed keyboard-focus outline measured 8.01:1. These are recorded samples, not a claim of complete site-wide contrast coverage. All eight updated list pages scored 100 out of 100 in an automated mobile accessibility audit. These results do not establish successful phone-width visual testing, browser zoom, tap-spacing testing or screen-reader use. Those completion checks remain open.

Structure and form accessibility: reviewed, still open

The original requirement includes useful headings, landmarks, image alternatives, form labels and errors, with actual screen-reader checks. Read-only checks found correctly associated fields on Contact and the Help search. Article comments still have a vague comment label and incomplete native field semantics. Repeated search-field identifiers generated by the theme also remain unresolved. No supported native setting for these theme-generated issues was available, and no screen-reader or form-delivery test is claimed.

Accurate search and share metadata: reviewed, still open

Ten articles whose decorative photographs were removed passed live checks for matching titles, descriptions, canonical addresses and share addresses. Structured data parsed, and no removed photograph remained in their share-image tags. No accurate metadata was rewritten. The brand author is still represented as a Person in the structured data; a supported correction and subsequent validation remain necessary.

Exact-problem reporting and accountable resolution: reviewed, still open

The requirement is for a reader to report the exact problem and for a real owner to track it through resolution. The existing homepage readability report has implementation and desktop-check records. A publisher decision closing that specific report has not been recorded. This round’s article-list improvements are recorded separately; general appreciation is not treated as acceptance of that earlier issue.

The white, navy and deep-green identity, readable typography and purposeful photography are preserved. No personal publisher identity, decorative emojis or generic AI imagery were added.

Image-purpose review completed and seven exercises improved

Image-purpose review completed and seven exercises improved

Verified October 26, 2026.

Give every image a clear purpose

The requirement was to explain what each image helps a reader understand or find, and remove duplicate decoration. All 396 starting featured-image assignments were accounted for: 386 were reviewed in this round and 10 earlier purpose decisions were reconciled against the current site. We removed 88 decorative featured-image assignments and their obsolete credits, retaining 308 assignments with recorded reasons. The 87 distinct removed image files remain in the media library.

We also removed 10 redundant photo placements from six topic introductions and the Projects page, and cleaned decorative emojis from two retained photographer credits. Topic introductions now use a readable single text column. Useful lesson-choice photos, instructional screenshots, result previews and the site icon were retained. Active page, template, widget and social-image settings were checked for other placements.

Native content checks and live desktop review passed for the changed articles and pages. Eight article lists were checked after publication: removed photos left no visible placeholder frames, and retained list photos matched the current assignments.

Copy the complete exercise

Seven guides gained consistent exercise-selection and reading controls. Two exercises now include their fictional source notes in the copied text. Desktop keyboard activation, full manual selection, focus recovery and reading-view restoration passed in all seven; exact clipboard transfer was also checked for both repaired exercises. Completion still requires phone fallback and full fresh-chat evidence.

Accessible structure and controls

Five core routes scored 100 in automated mobile accessibility checks before the later image changes. These checks do not replace actual screen-reader testing. Remaining work includes screen-reader checks and unresolved form/landmark/error coverage, including duplicate search IDs and labels generated by the Astra theme.

Readability when enlarged

All seven guides passed their built-in desktop text-size increase and reset, with content and controls remaining available and no horizontal overflow at the tested viewport. Actual browser zoom, phone reflow, tap spacing and complete contrast evidence remain required. This record does not claim those tests have passed.

The established white, navy and deep-green design, readable typography and purposeful photography remain the design direction. The publisher’s personal identity is not disclosed.

Thirty guide improvements: image purpose and practical controls

Thirty guide improvements: image purpose and practical controls

October 25, 2026. Maintenance and evidence records are separate.

#43 — Give every image a purpose

Review scope: review each image against its role in explaining an activity, clarifying a step or showing a result, and remove weak duplicate decoration. The full site inventory is not yet finished.

Twenty-nine featured photographs were visually reviewed in their live article context. Twenty-two decorative placements were removed, together with their obsolete credits. Examples include blank notebooks on communication and planning lessons, generic equipment on moderation and partnership guides, and flowers beside blank paper on a work-instructions lesson. These images did not explain the specific task. No media attachment was deleted.

Four useful photographs remain: a conventional scale supporting a measured-weight comparison, headphones identifying an audio-story project, physical documents beside a laptop for a document-reading task, and recipe-related objects for the guide’s specific recipe-card product example. Three device-article image decisions were left unchanged for a later product-specific choice; these reviews are not counted as improvements.

The language-practice guide’s hero was already hidden before this round; its unused featured assignment and obsolete credit were removed without claiming a new visible hero removal. Existing October 3 review dates were preserved in both combined credit paragraphs. Educational guidance about taking photographs and checking image permissions was preserved.

Fresh inventory confirms 396 featured-image assignments across published articles, using 395 distinct media items. These are remaining assignments, not a claim that every retained image has completed its purpose review.

#19, #50 and #51 — Complete exercises and usable controls

Eight guides gained the shared complete-exercise controls, manual text-selection fallback, article text-size choices and reading view. Each already contained a complete fictional exercise or a self-contained blank-template request. Their instructional text, source and test dates, titles, URLs and image assignments were preserved.

  • #19: the complete copied instructions and supplied material must work with keyboard feedback and phone fallback. All eight passed exact readonly selection with Enter and Space, Close and Escape focus recovery. Four also passed exact keyboard and click clipboard comparisons using a different test value before each attempt. Existing core empty-chat evidence is retained; no new AI-product trial or phone fallback test is claimed.
  • #50: meaningful structure, accessible names, keyboard behavior and screen-reader operation must be verified. Unique controls and keyboard recovery passed here. Actual screen-reader testing and shared-header issues remain open.
  • #51: readable contrast, enlargement, actual browser zoom, narrow reflow and touch spacing must be verified. Exercise text measured 21.25px at Standard and 31.875px at Largest, with no document overflow at the tested 1363 × 936 desktop viewport. Actual zoom, narrow-screen and physical-device checks remain open.

What was verified

Native read-back matched every intended article change. All 29 reviewed media attachments and their metadata remain unchanged; the seven retained or postponed article/image combinations are unchanged. All 22 changed image placements were checked live for intact headings and introductions, removed credits, normal spacing and no desktop overflow. Seven representative changed layouts were also visually inspected. Reading view restored existing answer-disclosure states where present. Some first visits showed cached content; one normal reload revealed the published changes.

Limits: no new health, finance or product-source review, independent human-reader test, phone/tablet, screen-reader, browser-zoom, printed-page or downloaded-file verification is claimed. Save controls use unchanged shared code; file delivery was not retested. No form, message or account action was performed. The white, navy and deep-green identity and publisher anonymity are preserved.

Eight guides with new exercise controls

Twenty-two guides with image-purpose corrections

Twenty guides: complete exercises and readable controls

Twenty guides: complete exercises and readable controls

October 25, 2026.

What changed

Nineteen guides gained the shared exercise controls, selectable-text fallback, article text-size choices and reading view. The four-week learning plan gained the exercise controls and reading view while retaining its existing text-size choices. Each guide has one control set for its marked exercise; on the four-week plan this is the Week 2 book-swap exercise.

The voice-writing guide now copies the rewrite instruction and fictional note together while preserving the separate dictation practice. The skill-practice prompt now asks which spreadsheet app, version and reference the reader uses before giving software-specific instructions, with an arithmetic-only fallback. Both material revisions have accurate October 25 update notes; their original publication and evidence dates are preserved.

The 25 edits comprise 19 shared-control additions, one replacement of a text-size-only block, two complete-prompt repairs, two update-note revisions and one reading-view placement repair found during live checking. Existing photographs, article titles and addresses remain unchanged. The controls use the established white, navy and deep-green design.

Verification and limits

  • #19 — Complete exercises: instructions and fictional input must be copied together; keyboard feedback, manual selection and phone fallback must work. All 20 guides passed exact full-text selection with keyboard recovery. The two revised prompts also passed keyboard and click clipboard comparisons. Earlier core empty-chat trials remain valid evidence; phone fallback is still unverified.
  • #50 — Accessible controls: clear names, keyboard operation, focus recovery and screen-reader behavior must be checked. Unique controls and focus return passed in this desktop batch. Actual screen-reader testing and the known shared-header issues remain open.
  • #51 — Readability: contrast, enlarged text, actual browser zoom, narrow reflow and touch spacing must be checked. Exercise text measured 21.25px at Standard and 31.875px at Largest, with no document overflow at the tested 1363 × 936 desktop viewport. Actual zoom, narrow-screen and physical-device checks remain open.

Reading view opened existing answer disclosures and restored their prior states where disclosures were present. The learning plan initially lacked its toolbar; a guide-only placement repair was applied and rechecked. Native read-back matched the planned edits, with titles, URLs, publication status and featured-image assignments preserved.

Limits: no new AI-product trial, health or finance source review, independent human-reader test, phone/tablet, screen-reader, zoom, printed-page or downloaded-file verification is claimed. Save controls use the existing shared code; file delivery was not retested. No message, form submission or account action was performed.

The 20 updated guides

Practical exercise controls and screenshot recovery

Practical exercise controls and screenshot recovery

October 25, 2026. Private queue, release and public evidence-record updates are separate.

#67 — Official-source monitoring and ranked review

Review scope: monitor official product guidance and reader reports, keep a ranked review queue, review volatile claims monthly and stable lessons quarterly, and bring a relevant finding to a real reviewer without excessive low-value alerts.

The due screen-help and sharing-data source checks were completed, with an additional check of the manual’s PDF guidance. One actionable finding became a verified update: the screen-help guide lacked a recovery route for a rejected screenshot file. The conditional PNG route described below was recorded in the ranked private queue and reported to the publisher in this round’s conversation. Existing reader-feedback tracking, weekly triage, monthly/quarterly cadence, three-item briefing limit, eight-call scheduled-review limit and twenty-call reserve remain in place.

The sharing-data explanation, manual PDF guidance, existing screenshot shortcuts and scam caution required no rewrite. These no-change findings are recorded without repeat alerts. The next volatile-claim review remains November 3, 2026; stable-lesson review remains January 5, 2027, with earlier review after a material change or relevant report. This closes the implemented source-review workflow using a specific delivered finding, not schedule configuration alone. Automated notification delivery, human read receipts and independent reader outcomes are not claimed.

#18 — Clear device-specific steps

Review scope: one main action per numbered step, current annotated screenshots at difficult controls and verified device differences, so a fresh-start reader is not left with unexplained clicks.

The screen-help guide now has a conditional “My screenshot file is rejected” answer. It connects the image formats documented by OpenAI with the PNG/HEIF capture choices documented by Apple, gives four steps to obtain and inspect a PNG on supported Macs, and keeps the typed-label fallback. It does not claim that all HEIF uploads fail.

The panel starts closed, opens with Enter and closes with Space while keeping visible focus. Its four steps, official links, privacy check and fallback were verified in the live page. Source/update entries now identify the October 25 documentation work; the original October 19 product-test date remains unchanged. No actual Mac capture, image upload or new screenshot-recognition test was performed. Actual device walkthroughs and remaining annotated-control evidence stay open.

#19 — Complete copying and manual selection

Review scope: complete exercise instructions and source material usable in an empty chat, reliable keyboard copying and feedback, selectable manual text and verified phone fallback. Nine guides with complete prompts but no exercise controls now have one shared Copy/Select/Save control set, plus reading/print view and text-size choices.

All nine passed keyboard selection of the entire readonly prompt, Close/Escape and focus return. The habit guide copied all 481 characters exactly with Enter and click; a distinct clipboard value was set before the keyboard check to avoid a false pass. Automatic copying on the other eight was not newly tested. Saved-file inspection was not repeated after prior browser download failures, and is not claimed as passed. The shared script is unchanged. Phone fallback and remaining fresh-chat checks stay open.

#50 and #51 — Accessible, readable controls

Review scope: meaningful structure, labels, errors and focus verified with screen readers; measured contrast, usable targets, enlarged text, actual zoom and narrow-screen reflow.

Each of the nine guides has one control set, reading toolbar and text-size group. Prompt text, buttons, feedback and selection fields measure 21.25px at Standard and 31.875px at Largest; controls are approximately 60px and 84px high. No horizontal page overflow was found at the 1363px desktop viewport. Meal-planning and digital-paperwork tables also fit at Largest and in reading view.

Reading view opened existing disclosures and restored their previous states; four guides also passed a mixed open/closed-state check. Pages without disclosures passed only the view toggle and focus-return checks. All pages were left in Standard text and normal view. The new screenshot panel also enlarged without overflow. White, navy and deep-green styling, square controls and readable labels follow the standing design brief. No new imagery or personal owner identity was added. Screen-reader testing, the known shared-header label defect, actual zoom, narrow-screen and phone/tablet checks remain open. Browser printing was not tested.

Nine guides with new exercise controls

  1. Use AI to Organize Your Money Decisions.
  2. Use AI to Compare Insurance Documents.
  3. Use AI for Memory Practice and Everyday Reminders.
  4. Use AI to Review a Medical Bill.
  5. Use AI to Prepare a Clear Refund or Warranty Request.
  6. Plan a Week of Meals With AI.
  7. Make a Better Grocery List With AI.
  8. Use AI to Organize a Small Batch of Digital Paperwork.
  9. Fabulous: Try One Habit Before Buying More.

Verification

  • Nine shared-control placements, one per guide.
  • One new screenshot-recovery disclosure.
  • Two scoped date-entry updates in the screen guide: substantive change and source review.

Native read-back matched all ten changed article bodies exactly; titles, URLs, published status and featured-image assignments were preserved. The nine control placements preserve all original article bytes, and the screen guide preserves existing scripts and its recorded product-test evidence.

Complete exercise controls on ten more guides

Complete exercise controls on ten more guides

October 25, 2026.

#59 — Feedback and review

Review scope: low-friction feedback about finishing a task, useful reasons such as unclear instructions or poor results, no request for private prompt contents, and weekly review that turns specific reports into tracked fixes without treating clicks as completed learning.

The implemented process has been demonstrated with the genuine October 19 report that homepage photos and adjacent text were too small. The private issue record links that report to the enlargement, October 20 verification and October 24 follow-through. The same issue remains in the maintenance queue; it is not counted as a second report. The weekly review process is defined and its Tuesday schedule is enabled.

This round rechecked the live feedback route: “partway through” and “text or photos that are hard to read” reached an editable Contact message with the exact guide address and a privacy link. The test message was not submitted, and no new reader report was created. Existing dated reviews show a report becoming a tracked, verified implementation; the closure is based on that evidence and the original scope, not a new claimed reader outcome.

Limits remain: Contact-email delivery, unattended notification delivery, reporter acceptance and physical-device retesting are unverified. The enabled schedule has no recorded run/result available and its next-run field is unset; this record does not claim that an automated review has run or reached the publisher. Those operational limits remain tracked.

#19 — Complete exercise copying

Review scope: full instructions and source material that work in an empty chat, reliable keyboard copying and feedback, selectable text and verified phone fallback. Each of the ten guides already contained one complete fictional practice prompt but had no Copy/Select/Save controls. Each now uses the existing shared controls once, preserving all original instructions and source material.

All ten live guides passed complete readonly selection, keyboard Select, Close/Escape and focus-return checks. The tutoring exercise copied all 753 characters exactly with Enter and click. The keyboard check began with distinct clipboard text, avoiding a false pass from a previous copy. Earlier clipboard inconsistencies remain unexplained and are not retroactively counted as passes. The tutoring download request produced no file path in this browser, so its saved file could not be inspected. The shared download implementation is unchanged. Phone fallback, wider keyboard clipboard coverage and remaining fresh-chat checks stay open.

#50 and #51 — Accessible, readable practice controls

Review scope: meaningful structure, labels, errors and focus verified with screen readers; measured text/control contrast, usable targets, enlarged text, actual zoom and narrow-screen reflow. Every changed guide has one labelled control set, one reading/print-view control and one text-size group. Prompt text, buttons, feedback and selection fields measure 21.25px at Standard and 31.875px at Largest. Buttons measure about 60px and 84px high respectively. No horizontal page overflow was observed at the 1363px desktop viewport.

The reading view opened existing answer disclosures and restored their previous states when closed. The useful-task guide also passed restoration of a mixed open/closed state. Guides with no disclosures passed the view toggle and focus return, without a claim that disclosures were tested there. All guides were left in Standard text and normal view. The unchanged white, navy and deep-green control palette retains measured green/white contrast of 8.01:1. Actual browser printing, phone/tablet use, zoom, narrow-screen and screen-reader checks remain unverified; the known shared-header label issue remains open.

Ten individual guide edits

  1. Put AI to Work on One Useful Task: one shared-control placement.
  2. Use AI to Build Clear Answers to Customer Questions: one shared-control placement.
  3. Find Useful Patterns in Customer Feedback: one shared-control placement.
  4. Use AI to Find Your Notes: one shared-control placement.
  5. Turn Voice Notes Into a First Draft: one shared-control placement.
  6. Use AI to Find Better Secondhand Shopping Options: one shared-control placement.
  7. Turn Tutoring Notes into Useful Practice with AI: one shared-control placement.
  8. Choose Golf Technology for One Clear Goal: one shared-control placement.
  9. Use AI to Organize Family Memories: one shared-control placement.
  10. Practice a Language With AI: A Simple Routine: one shared-control placement.

The source checks confirmed one complete prompt per guide, no competing scripts or duplicate shared references, balanced block markup and preservation of all original article bytes. Native read-back matched all ten expected bodies exactly; titles, URLs, published status and featured-image assignments were preserved. No shared script, photograph or personal owner identity was changed. The new controls also provide reading/print and text-size choices; existing optional guide-sharing behavior only appears where its required ending is present, so sharing is not claimed for every guide.

Clearer worked examples and copy feedback

Clearer worked examples and copy feedback

October 25, 2026.

#19 — Complete copying with a usable fallback

Review scope: complete instructions and fictional source material usable in an empty chat, reliable keyboard copying and feedback, selectable manual text, and verified phone fallback. The shared exercise controls used by 35 published guides now ignore obsolete copy responses after a newer action, a closed section or a changed source. A delayed failure cannot reopen an old selection box or take focus from a newer exercise. A superseded pending message says the result is unconfirmed. Synchronous clipboard errors use the full-text fallback.

All 40 local regression tests passed; 27 exposed failures in the previous script. These tests use a mock DOM and deferred promises, not live clipboard-race testing. Manual selection, readonly text, complete selection, Close/Escape and focus return passed on the manual, household and creative guides. The manual guide’s 725-character exercise and 544-character reference card copied exactly by click. Enter activation reported success, but the test browser’s virtual clipboard stayed empty and could not paste; keyboard transfer remains unverified and its cause is not established. Download requests on household and creative guides returned no file to inspect. All four download payload variants were preserved and passed local tests; this is not a live download pass. Phone fallback and remaining fresh-chat checks stay open.

#50 — Meaningful structure, labels and focus

Review scope: headings, landmarks, labels, errors and meaningful image information verified with screen readers. Four existing worked answers now use readable HTML figures: a three-step discount calculation, a completed craft-planner example, a labelled fruit-count chart and a source-evidence comparison. The chart keeps apples 8, pears 5 and oranges 7 on one zero-based 0–8 scale; visible numbers convey the values without color. All remain explicitly fictional teaching examples, not recorded AI responses. Original exercises and answer-reveal labels are preserved. The three existing answer reveals start closed and support Enter/Space with visible focus. Actual screen-reader checks and the known shared-header label defect remain open.

#51 — Contrast, enlarged text and readable controls

Review scope: measured text/control contrast, usable targets and non-color cues, enlarged text, actual zoom and narrow-screen reflow. Shared copy feedback now measures 21.25px at Standard, up from 19.125px. Square deep-green controls and ruled white selection panels match the navy/green identity. On three sampled guides, controls, selected text and feedback measure 31.875px at Largest with no desktop overflow; this Largest size already worked before this round.

The four worked examples use native selectable text, purposeful rules and clear labels, with default body text at 21.25px. Sampled navy/white contrast is 12.57:1, green/white 8.01:1 and divider/white 4.22:1. The chart bars have exact 100%, 62.5% and 87.5% proportions. All four layouts passed desktop overflow checks at a 1363px viewport. These four pages have no text-size controls, so no Largest-mode result is claimed for them. Actual browser zoom, narrow-screen reflow and phone testing remain unverified.

Verification

  1. Shared copy-script response and error handling.
  2. Shared copy-control and selection-panel styling.
  3. Tutoring calculation styling.
  4. Tutoring calculation structure.
  5. Craft-planner example styling.
  6. Craft-planner example structure.
  7. Fruit-count chart styling.
  8. Fruit-count chart structure.
  9. Source-evidence comparison styling.
  10. Source-evidence comparison structure.

The eight article edits produce four clearer worked examples; they are not eight completed audit recommendations. Native read-back matched all five changed assets exactly, preserving titles, URLs and published status. All 35 shared-block references were checked in source; live interaction checks sampled three guides. No new images or personal owner identity were added. Image assignments are unchanged, so this pass does not claim progress on the remaining full image inventory.

Purposeful images and readable practice records

Purposeful images and readable practice records

#43 — Give every image a useful role

Review scope: an image-by-image purpose review across articles, pages, templates and social imagery, with unnecessary decoration removed. Forty-eight assigned photos were visually inspected in their article context. Six useful photos were retained; 42 repeated or poorly matched assignments and their 42 obsolete standalone credits were removed. Examples included the same dark-screen laptop used across unrelated lessons, empty stationery without a connection to the task, and an unrelated coffee mug. Retained uses include a person actively working, headphones for listening tests, and three sticky notes for a three-screen paper prototype.

All media files remain available for restoration. Two article heroes were already hidden, so those changes clean assigned/archive/social imagery rather than newly removing visible article photos. Source-confirmed totals are now 418 featured assignments using 417 distinct media items. This targeted review does not complete the wider inventory.

#19 — Complete copying with a usable fallback

Review scope: complete copied material, keyboard activation, feedback, manual selection and verified phone fallback. The inbox guide’s blank sorting record now has a labelled read-only selection box, Close and Escape, focus return and protection against obsolete copy callbacks. Its unchanged 624-character record copied exactly with Enter; complete selection and reopening checks passed. Twenty-six local asynchronous tests passed, including rejection and closing the record while a copy is pending; these are mocked tests, not live clipboard-race tests.

Isolated rechecks also copied all 862 meeting-exercise characters with Enter and Space, and the 76-character home-maintenance guide link with Enter. No code change was made for those rechecks. Earlier mismatches remain unexplained and were not reproduced. Phone fallback remains unverified.

#50 and #51 — Accessible, readable controls

Review scope: meaningful labels, structure, focus and errors verified with screen readers; measured contrast, enlarged text, actual zoom and narrow-width reflow. New inbox controls are labelled, keyboard-operable and return focus predictably. Weekly planning and context-sharing guides each gained one instance of the existing text-size choices.

Inbox copy controls, notebook download links and weekly quiz buttons now measure 21.25px at Standard and 31.875px at Largest. Notebook download links had remained at 17px at both sizes. Household rows and notebook downloads change from two columns to one at Largest. The household saved-plan text already enlarged correctly; this pass does not claim a new font-size improvement there. Sampled green/white text contrast measured 8.01:1 and navy/white 12.57:1; other sampled component text and control borders also passed their applicable checks. These measurements do not establish full-site accessibility compliance.

The five changed interactive guides passed the recorded desktop checks without horizontal overflow at 1363px. Screen-reader testing, the known shared-header label defect, actual browser zoom and narrow-viewport verification remain open. A keyboard zoom shortcut did not change the test browser’s scale and is not counted as a zoom test.

Individual edit count and verification

  • 42 featured-image assignment removals.
  • 42 corresponding standalone-credit removals.
  • Four inbox edits: control styling, worked-example styling, full-record selection markup, and guarded copy/selection behavior.
  • Three component styling edits: weekly quiz, notebook downloads and household plan rows.
  • Two additions of existing text-size choices: weekly planning and context sharing.

Native read-back matched all 45 affected article bodies exactly; titles, URLs and published status were preserved. All 42 image changes passed live heading/content, image, credit and social-image checks after refreshing older cached views where needed. The fictional examples, calculations and embedded download payloads were preserved. White, navy and deep-green surfaces, ruled sections and square controls follow the standing design brief. No new imagery or personal owner identity was added.

Complete exercise selection and readable choices

Complete exercise selection and readable choices

#19 — complete copying and manual fallback

Review scope: full exercise instructions and source material, reliable keyboard copying, selectable fallback, clear feedback and remaining phone/fresh-chat checks. The meeting guide now offers Select full exercise, a labelled read-only box, Close and Escape. The box contains the unchanged 862-character exercise and all five fictional notes. Its static text remains in the HTML if JavaScript is unavailable. Automatic-copy failure selects the complete text instead of only telling the reader to find it in the article.

The meeting and whole-task time calculator ignore obsolete clipboard callbacks after newer actions. Eighteen meeting and 34 time-calculator local mock-DOM checks passed, covering current failure, newer selection/closing and delayed success or failure. These are not live clipboard-race tests.

Live click-copy matched all 862 meeting characters and all 861 time-report characters. Full selection, Close/Escape and focus return passed. The time download contained the exact report in a 927-byte file. Initial clipboard reads were inconsistent. In a final check without other clipboard tests running, time copying passed by keyboard and click, and meeting click-copy passed; the first meeting keyboard-copy read still did not match. Its cause is unestablished, so keyboard clipboard transfer for that guide remains unresolved. Phone fallback and remaining fresh-chat checks remain open.

#50 — labelled controls and understandable feedback

Review scope: meaningful structure, labels, errors and media alternatives verified with screen readers. Meeting selection is labelled, read-only and keyboard-operable. Close and Escape hide the panel, update its expanded state and return focus to Select full exercise. The new text-size groups appear once on the meeting and time guides. The privacy picker still distinguishes four useful details from three unnecessary details, clears outdated feedback after a change, and resets all seven choices with focus returned to the first checkbox.

Screen-reader verification and the previously recorded shared-header search-label defect remain unresolved. No theme file or PHP change was made.

#51 — readable controls and adaptive layout

Review scope: measured contrast, enlarged text, actual browser zoom, narrow-width reflow and usable control spacing, with demonstrated failures fixed. Meeting controls and status text increased from 17px to 21.25px Standard and 31.875px Largest. Time-calculator hints increased from 19.125px to 21.25px Standard and reach 31.875px at Largest. The new size controls let readers enlarge those guides without changing the rest of the site.

In the privacy exercise, the legend previously stayed at 17px even with Largest selected. It now grows with the choices to 31.875px. The two 368px columns become one 760px column at Largest; checkboxes grow with the labels and the explanation control measured about 76px high. Time fields and results also form one column at Largest. No horizontal overflow appeared at the tested 1363px desktop width. White, navy and deep-green styling, ruled sections and square controls follow the standing design brief. Actual zoom, narrow-screen, physical-device and screen-reader checks remain open.

#61 — clean guide links in more lessons

Review scope: accurate share/print previews, correct destinations, a reused lesson excerpt and useful engagement evidence. The existing Keep or share this guide component now also follows an existing next-step panel. This supports 11 additional guides, bringing source-confirmed coverage from 24 to 35. The original finishing-panel placement takes priority and duplicate controls are prevented.

Three newly supported guides were exercised live: home maintenance, assistant alternatives and work instructions. Each showed one component immediately after its next-step panel, full URL selection and a matching canonical URL after click-copy. The first home-maintenance keyboard-copy read differed; a click retry matched, and that initial discrepancy is retained in the evidence. All 35 guides were not individually live-tested. No link was sent to another person and no engagement result is claimed.

Ten individual edits

  1. Meeting copy and selection-control styling.
  2. Meeting worked-example typography and ruled styling.
  3. Labelled complete-exercise selection panel with static fallback.
  4. Meeting copy, Select, Close, Escape and delayed-callback behavior.
  5. Meeting article text-size controls.
  6. Time-calculator typography and adaptive layout.
  7. Time-calculator delayed-copy callback guard.
  8. Time article text-size controls.
  9. Privacy-picker typography, choice layout and explanation target.
  10. Shared guide-link placement after next-step panels.

Verification: native readback matched all three article bodies and the shared block exactly; titles, URLs and publication status were preserved. The time example still gives 22 minutes for the first attempt, 17 for a later scenario, and a two-task projection of 39 minutes versus 40. These are fictional entered values, not measured reader savings.

Larger practice text and visible keyboard focus

Larger practice text and visible keyboard focus

#51 — readable text, contrast and reflow

Review scope: measured contrast, enlarged text, actual browser zoom, narrow-screen reflow, usable control spacing and non-color cues, with demonstrated failures fixed. Six practice sections now inherit readable article typography: checking claims, subscription costs, trip planning, document summaries, comparing options and task lists. The claim, budget and trip guides also gained the existing Standard, Larger and Largest controls.

Claim hints increased from 13px to 21.25px at Standard and 31.875px at Largest. Budget fields, result labels and summary text now use 21.25px Standard instead of fixed 18px or 16px. Trip labels and recorded text, summary/comparison evidence and tables, and the task prompt also scale to 31.875px at Largest. The trip grid changes from two 368px columns to one 760px column. No horizontal document, tested table or field overflow appeared at the available 1363px desktop viewport.

White, navy and deep green replace cream and repeated rounded practice panels. Controls retain their functional grouping; selected claim answers gain an underline as well as color. Palette calculations give approximately 12.57:1 for navy on white, 8.01:1 for green on white and 4.22:1 for the input border on white. These measurements do not certify every site state. Physical-device, actual browser-zoom and narrow-width checks remain open.

#50 — focus and understandable controls

Review scope: meaningful structure, labels, errors and media alternatives verified with screen-reader checks. At Largest, Next claim originally focused a heading about 266px above the visible screen. The claim script now permits normal focus scrolling for new claims and completion. Retesting placed new-claim headings visibly around 425–452px down the viewport; the completion heading reached the top. Restart resets the choices and returns focus to claim one. All four answers, wrong-answer feedback and Enter/Space operation passed. Budget invalid-input feedback and focus, and selection-panel Close/Escape behavior also passed.

Screen-reader verification and the previously recorded shared-header search-label issue remain unresolved. This round changed article components, not theme files or PHP.

#19 — reliable copy feedback

Review scope: complete exercise inputs, selectable fallback, keyboard activation and clear feedback, plus remaining phone fallback and fresh-chat checks. The budget calculator now ignores obsolete clipboard callbacks after amounts change or a newer action occurs. A delayed failure can no longer reopen an empty or replaced summary and steal focus. Twenty-seven local before/after checks cover success, failure, invalidation and latest-attempt behavior; these are mock-DOM tests, not browser race injection.

In the live browser, the fictional 50/20/180 inputs still produced $35 monthly equivalent, $420 yearly charges, $200 first-month scenario and $20 for each other month. Actual clipboard contents matched the 868-character summary; the 936-byte text download contained that exact summary. Keyboard selection retained all 743/405 characters for trip exercises, 636 for summaries, 561 for comparisons and 624 for tasks. Phone fallback remains unverified.

Eleven individual edits

  1. Claim-practice typography, source panel and choice layout.
  2. Budget fields, readable result rows and summary styling.
  3. Trip demonstration, evidence labels and selection-panel styling.
  4. Document-summary evidence, table and selection-panel styling.
  5. Options-comparison evidence, table and selection-panel styling.
  6. Task exercise and selection-control styling.
  7. Budget delayed-copy callback guard.
  8. Text-size controls on the claim-checking guide.
  9. Text-size controls on the budget guide.
  10. Text-size controls on the trip guide.
  11. Visible focus scrolling for new claims and completion.

Verification and limits: all six native article bodies matched the reviewed changes exactly, with titles, URLs, publication status, lesson facts and embedded practice sheets preserved. Standard text was restored after testing. No independent reader trial, screen-reader test, physical phone/tablet test, actual zoom test or network capture is claimed. No owner identity or new imagery was added.

Recover builder work and keep clean guide links

Recover builder work and keep clean guide links

October 24, 2026.

#36 — a useful local email-prompt builder

Review scope: editable goal, audience, tone and constraints without an AI API or saved personal inputs, a test user improving on the static example, and browser/network checks of the stated data behavior. The email builder now offers one-level Undo after building, loading the fictional example or starting again. It restores the brief, prompt and export availability. Recovery stays in page memory and becomes unavailable after new edits or leaving the page; this limit is explained beside the control. A disclosure explains field limits and checking long pasted text for missing facts.

Desktop checks restored all six fields, edited prompt text, visibility and export state after loading an example, rebuilding and restarting. Enter and Space activated Undo. Repeated empty restarts and invalid builds preserved recovery; new edits cleared it. Manual selection covered all 858 characters of the fictional prompt. The Largest reading setting retained readable controls without horizontal overflow. Independent test-user comparison and network-level privacy verification remain open.

#19 — dependable copying and recovery

Review scope: full instructions and fictional source text, selectable fallback, keyboard activation, clear feedback, and phone and remaining empty-chat checks. The builder ignores delayed copy results after the underlying brief or prompt changes, avoiding obsolete success messages or selecting newer text after an older copy failure. Local mock-DOM checks cover delayed success, delayed failure, reset/undo and repeated attempts; these are source-level tests, not claims of a real delayed browser clipboard test. Physical-phone fallback and remaining fresh-start checks remain open.

#50 — understandable controls and status messages

Review scope: useful structure, labels, errors, image alternatives and media alternatives verified through screen-reader checks. Undo has visible guidance and restores meaningful focus. Identical status messages are no longer rewritten repeatedly. The length-limit disclosure is keyboard-operable. At Largest, the builder Undo control was about 84px high and the guidance disclosure about 76px high at the tested 1363px viewport. Screen-reader and physical-device verification remain open.

#61 — easy, deliberate guide sharing

Review scope: accurate share and print previews, correct destinations, a reused lesson excerpt and evidence of useful engagement. A new “Keep or share this guide” section offers a labelled public link, Copy guide link and Select guide link. It removes query strings and fragments, includes no practice text, and sends nothing for the reader. White, navy and deep-green styling uses inherited readable typography, square controls and clear focus.

The shared component applies to 24 guides with a finishing panel. Three were exercised individually: the email, prompt notebook and product manual guides. Actual clipboard contents matched each public URL exactly; full manual selection, keyboard operation, one component per page and Largest-text layout passed. The other 21 were not individually live-tested. A stale manual-guide render resolved after one normal reload. Broader preview, excerpt and engagement evidence remain open; no outreach was sent.

Ten individual edits

  1. Undo control and page-memory recovery guidance.
  2. One-level restoration of fields, prompt, error and export state.
  3. Recovery before loading the fictional example.
  4. Recovery before a valid rebuild.
  5. Recovery before restart, preserving useful history on repeated empty restarts.
  6. Delayed-copy result guards.
  7. Deduplication of identical status messages.
  8. Field-length and complete-paste guidance.
  9. Shared guide-link styles.
  10. Shared clean-link, copy and selection behavior.

Native readback matched both reviewed source bodies exactly. Titles, URLs and publication status were preserved. Eighteen local builder checks and shared-link unit tests passed. No physical phone/tablet, screen-reader, independent human-outcome or network-capture test is claimed. No owner identity was added.

Clearer copying controls and larger practice screens

Clearer copying controls and larger practice screens

#19 — complete exercise copying

Review scope: complete exercise inputs, selectable fallback text, keyboard activation and useful feedback, plus phone fallback and remaining fresh-start checks. The trip guide now provides a Close control and explanatory help for each of its two text-selection boxes. A shared handler supports Escape, updates the expanded state and returns focus to the correct Select control. Selecting text is explicitly distinguished from copying it.

Both panels were opened and reopened by keyboard. Their full 743- and 405-character texts matched the displayed source, and the entire value was selected. Escape and the Close controls hid the right panel, restored focus and set its expanded state to false. Physical-phone fallback and automatic clipboard transfer were not tested this round.

#50 — useful control names and accessible operation

Review scope: meaningful structure, image alternatives, labels, errors and media alternatives verified with screen-reader checks. Five notebook sources now identify the email-reply template, task-list template, options-comparison template, first exercise and second exercise. Their Copy, Select, Save and Close names distinguish the associated source instead of repeating a generic label. The exercise text itself is unchanged.

All five manual selections matched their full source text: 436, 344, 420, 517 and 578 characters. A sample task-list download contained the correct label, guide link and exact 344-character exercise. These are browser and source checks, not screen-reader certification. The previously recorded shared-header search-label defect remains unresolved.

#51 — readable practice components

Review scope: measured contrast, usable targets, non-color cues, enlarged text, actual browser zoom and narrow-screen reflow without lost essential content. The fictional screen rehearsal now inherits the article reading size. Its 17px base and 13–15px secondary text were replaced with 21.25px text at the checked normal setting. Buttons measured 57.875px high. With the on-page Largest setting, the component and buttons measured 31.875px and the buttons measured 81.78px high.

The manual source comparison now uses white ruled columns, green numbered highlights and 21.25px labels. Its four source phrases and four matching checks remain intact. At Largest, labels and checks measured 31.875px and the two columns became one 760px column. The rehearsal’s cream format menu and both components’ rounded panels were replaced with the established white, navy and deep-green treatment.

Text and selected-button color pairs measured 7.99:1–12.57:1 contrast; the window border measured 4.22:1 against white. Keyboard Export, PDF selection and reset worked, and the focused Export control had a visible 3px outline. All four edited guides had no horizontal overflow at the available 1,363px desktop width. The on-page text-size control is not an actual browser-zoom or physical-device test; those and screen-reader checks remain outstanding.

Verification and individual-change count

10 site edits: two trip text-box controls and their instructions, one shared close/Escape/focus handler, five notebook source labels, one complete manual source-component restyle and one complete rehearsal-component restyle. CSS properties are not counted as separate improvements. The public update and private release/reversal record are two additional evidence records.

Fresh native read-back matched all ten prepared changes exactly. Published status, titles and URLs were preserved; no exercise facts were rewritten. No new imagery, decorative emojis, form submission or personal publisher identity was added. No new backup, restore, independent reader trial or physical-device test was performed.

Commercial disclosure confirmed and four image cleanups

Commercial disclosure confirmed and four image cleanups

#8 — commercial independence and disclosure

Review scope: commercial-independence rules are published; relevant commissions, sponsorships, paid placements and supplied review products are disclosed beside recommendations; testing and access claims remain accurate. The publisher explicitly confirmed that AI Over 40 has no affiliate commissions, sponsorships, paid placements, or free or discounted products or services supplied for reviews.

The commercial-independence policy now records that dated declaration and requires an update and a nearby disclosure if a relationship changes. Its existing rules continue to prohibit buying favorable conclusions or rankings and distinguish purchased, publicly free, supplied and untested access. A short contextual disclosure was added near the beginning of Use AI for Honest Affiliate Content. Both statements and their link were verified live. The earlier 517-article link scan found no recognizable affiliate merchant tracking; the publisher’s declaration resolves the private-relationship gap that a link scan could not answer. This is current publisher disclosure, not an independent financial audit.

#43 — purposeful images

Review scope: an image-by-image purpose and duplicate-decoration review across article, page and template uses. Four unrelated stock photographs were removed: a gaming desk from the domains/NFT claims guide; a blank notebook and coffee from the consultant-selection guide; a camera and notebook from the affiliate-content guide; and a blank sketchbook with art supplies from the licensing guide. Their four obsolete credits were also removed. The attachments remain available for reversal.

Six additional photographs were retained after visual review because they support their topics: budgeting materials, lending figures, a business-data display, old family photographs, a virtual-reality headset and an event speaker. Their observed descriptions were accurate. The native inventory now contains 517 published articles, 460 featured-image assignments and 459 distinct assigned media IDs; 57 articles have no featured image. Inventory counts do not imply that every remaining image has passed visual review.

#54 — current sharing metadata

Review scope: accurate titles, descriptions, sharing previews and structured data matching visible content. Three removed images remained in cached sharing metadata and were refreshed through the SEO plugin using their exact existing descriptions. The affiliate guide’s disclosure edit refreshed its metadata automatically. All four live pages now omit the old sharing-image references and preserve their descriptions. Existing shares may still use an external platform’s older cache. The separate brand-author schema correction remains undeployed.

Blocked: #50 — complete accessibility review

Review scope: useful headings, landmarks, image alternatives, form labels/errors and screen-reader verification. A new shared-header issue was confirmed: the desktop and mobile search inputs reuse the same ID, and both labels point to it. In the inspected DOM, the first input has two associated labels and the second has none. The theme’s search template contains the hard-coded ID and label target. A supported theme-level correction is still needed; no live theme file was changed, and no physical-phone or screen-reader result is claimed.

Verification and individual-change count

Native read-back confirmed all four featured assignments are zero. Three article bodies differ only by the removed credit; the affiliate guide additionally contains its intended disclosure. All four live articles have the expected title, no main article photo or obsolete credit, no old sharing-image reference and no horizontal overflow at the available desktop width. The policy anchor occurs once, and the contextual disclosure points to it.

Separately, three stale SEO records were refreshed without rewriting their descriptions, and three evidence records were updated. No new imagery, decorative emojis or owner identity was added; the white, navy and deep-green identity remains consistent.

Clearer record labels and less decorative imagery

Clearer record labels and less decorative imagery

October 24, 2026.

#43 — give every image a useful role

Review scope: an image-by-image purpose review across articles, pages and templates, with unnecessary duplication and decoration resolved. Six featured photographs were removed after visual inspection: abstract lights from the teleportation explainer; the same circuit-board photograph from both the brain-implant and U.S.–China AI explainers; a blank laptop from the charity-research guide; a moon photograph from dream reflection; and a mug with a blank phone from symptom tracking. These images did not explain the specific activity, evidence or result discussed. Their six obsolete photo-credit paragraphs were also removed. The media files remain available for reversal; the guide instructions and source notes are unchanged.

The current article inventory has 517 published articles, 464 featured-image assignments and 463 distinct assigned media IDs; 53 articles have no featured image. These counts describe inventory, not completed visual review. The full collection and non-article image uses remain open. No generic replacement imagery was added.

#50 — communicate page structure clearly

Review scope: useful headings, landmarks, image alternatives, form labels and errors, and screen-reader verification. The expandable review groups on this page were using the first subsection as their label: for example, “Advanced: #7” appeared instead of the record title. The grouping script now uses the record’s first heading, and the shared heading style supports both existing heading levels. Subsections remain inside their proper record.

Keyboard activation collapsed the corrected group while retaining focus on its summary. Searching for the source-review/readability record returned its descriptive title; opening that result expanded the right group and moved focus to its summary. The original #7 subsection remained inside the record. No horizontal overflow was found at the available desktop width. This is a verified structure and keyboard correction; a complete screen-reader and form-error review remains outstanding.

#54 — keep sharing metadata aligned

Review scope: accurate titles, descriptions, sharing previews and structured data matching visible content. The six newly removed photos initially remained in cached sharing metadata. A seventh stale image was found for the earlier Lumosity photo removal. Those seven SEO records were refreshed through the SEO plugin while preserving their exact saved descriptions. The old image references are absent from fresh metadata responses. Existing shares may retain a platform’s own older cache.

The separate brand-author schema issue remains blocked: an Organization publisher does not correct an author still typed as Person. The prepared code correction is not deployed. This round does not close #54 or claim that an external sharing service has refreshed its cache.

Verification and individual-change count

Native read-back confirmed all six featured assignments are zero and all six article bodies differ only by the exact removed credit paragraph. The review-group behavior was checked live with keyboard activation and search. The seven refreshed descriptions match their previous saved values, and the removed images no longer appear in fresh sharing metadata.

Separately: seven stale SEO records refreshed without rewriting their descriptions, plus two evidence-record updates. The white, navy and deep-green identity and anonymous publishing identity are preserved.

Source review and follow-through on a readability report

Source review and follow-through on a readability report

October 24, 2026. Two live guide edits and one scheduled review were made, with four record changes counted separately.

#7 — track the exact reported problem

Review scope: a reader can report the exact article and issue, delivery is verified, and the responsible publisher follows the issue through resolution with a brief public change note. The original October 19 screenshot was matched to the four homepage lesson previews. Their current desktop photographs measure 360 × 240 pixels; main descriptions are about 22.3 pixels, supporting text 19.1 pixels and captions 18.1 pixels. All four practice links point to the intended lessons, with no horizontal overflow at the available 1363-pixel viewport. This verifies the existing implementation against the original report; it is not a new homepage edit or a second report. Form delivery, reporter acceptance and phone/zoom checks remain unverified.

#59 — turn reader reports into tracked fixes

Review scope: a weekly review turns specific, privacy-conscious feedback into tracked fixes; a click is not counted as successful task completion. The existing private issue record now connects the original screenshot, affected previews, current measurements and remaining checks. A weekly maintenance review is configured for Tuesdays at 9 a.m. Eastern, starting October 27. Its first run and delivery remain unverified; no Contact submission or independent reader acceptance is claimed.

#67 — maintain a ranked source-review queue

Review scope: official product changes and reader reports reach a real reviewer through a ranked queue, with monthly checks of volatile claims, quarterly checks of stable lessons and few low-value alerts. The private queue now ranks the manual clarification, the original readability report and the outstanding device/accessibility evidence. The scheduled review uses those records, limits briefings to three actionable items and preserves the call reserve.

The manual lesson now distinguishes a PDF attached directly to a chat from one stored as GPT knowledge or project files, following the linked official documentation. A scoped source-check note preserves the October 19 text-only demonstration date and explicitly states that this was not a new upload test. Official-source checks of Start Here, the email lesson and the alternatives guide required no wording changes. Screen and sharing guidance remain due for the next monthly source pass. The first scheduled delivery and recurring publisher review have not yet been demonstrated.

#72 — keep one honest results dashboard

Review scope: one small dashboard tracks task completion, findability, reported usefulness, return use, corrections and maintenance hours, producing a monthly keep, improve or stop decision against a real baseline. The existing dashboard now records the known correction evidence, the decision to clarify manual guidance, the maintenance queue. Independent human outcomes, return-use results and actual publisher hours remain uncollected. The update does not invent a baseline or establish a completed monthly decision cycle.

Verification and individual-change count

The saved manual body was compared with its prior version and differs only by the two intended replacements. The live paragraph, linked source, scoped source-check note and retained demonstration date were checked. Body text measures 21.25 pixels; no horizontal overflow was found at the available desktop width. The original screenshot match is evidence for an existing fix, not a new user study or physical-device test.

Larger search and project previews

Larger search and project previews

#43 — give images a useful role

Review scope: an image-by-image purpose and duplicate-decoration review. Removed the repeated blank-laptop photograph from “Does Running AI Locally Keep Client Data Private?” and the identical headset still lifes from “Would a Virtual Showroom Help Your Business?” and “Test a Virtual-World Business Idea Before Investing.” Those images did not explain the data-handling or business-validation tasks. The wider collection review remains open.

#44 — consistent, uncropped image treatment

Review scope: consistent aspect ratios and focal-point checks at desktop and phone widths. Search thumbnails now use a 360 × 240 pixel frame at the tested desktop width, up from 220 × 164. All six Projects photographs now use contained 3:2 frames, approximately 524 × 349 pixels instead of 524 × 235. The complete source photographs remain visible. Search and project layouts stack below 900 pixels in the CSS; actual narrow-width visual checks remain outstanding.

#50 — useful accessible descriptions and structure

Review scope: headings, landmarks, image descriptions, form labels/errors, media alternatives and screen-reader verification across the maintained routes. Existing project image descriptions were retained; unnecessary article images were removed. Project credits increased from 12 to 19.125 pixels and retain underlined links and visible keyboard focus. This is a limited image/credit improvement, not a completed screen-reader review.

#51 — readable size, contrast and controls

Review scope: measured text/control contrast, enlarged text, actual zoom, narrow reflow, usable tap spacing and non-color cues. Search and project descriptions increased from 17 to 21.25 pixels; their headings are 29.75 pixels. Search dates increased from 14 to 17 pixels; project number labels increased from 12 to 17. Search Read More targets measured 44 pixels high, and all 18 project lesson links measured at least 44.125 pixels. Measured white-background contrast: body/caption text 10.36:1, deep-green links 8.01:1, search dates 7.63:1. Actual zoom, physical-device and screen-reader checks remain open.

Blocked: #54 — correct structured-data meaning

Review scope: accurate search/share previews and structured data matching the visible publisher and authorship. The previously identified brand author typed as Person is not corrected by the Organization publisher setting. The prepared author-schema correction remains undeployed; this release makes no schema change.

Verification and individual-change count

Verified all six project photographs and all ten rows of the email search results at the available desktop width. No horizontal overflow was found. Keyboard activation opened the intended email guide from search and the journal practice from Projects. Native read-back confirmed the exact CSS addition, all three featured-media removals, and article bodies unchanged except for their three credit removals.

Automated mobile checks reported 100/100 for accessibility and best practices on search and Projects; Projects also scored 100/100 for SEO. Search scored 58/100 for SEO because its existing internal-results page is marked noindex and has no meta description; those settings were not changed. Automated mobile checks do not substitute for visual phone-width or physical-phone testing.

The white, navy and deep-green identity is preserved. No owner identity, invented testimonial or new generic AI imagery was added.

Publication evidence and clearer topic photographs

Publication evidence and clearer topic photographs

October 24, 2026.

#68 — evidence review before publication

The original requirement is to use AI for drafting and preliminary checks, require evidence review for factual changes, and retain actual walkthroughs before publishing verified instructions or testing labels. The publication checklist now explicitly requires a linked record before a Tested or recorded-demonstration label is published, preserves failed attempts and access barriers, and identifies AI-operated checks and missing independent human review.

This completion is supported by the existing source-review and consequential-claim work, retained flagship input/output records, and a current review of positive testing claims. The review covered 517 article bodies: five articles make positive site-trial claims, and Start Here contains the first-chat walkthrough. Their evidence links point to retained records. Fresh saved-content fingerprints confirmed that the article evidence labels were unchanged from the reviewed corpus; the nine changed bodies contained only the documented reading-control or image changes. All seven relevant record anchors were present once in the live page.

The first task’s keyboard selection covered all 394 prompt characters; closing the text box restored focus. The next lesson and contextual Help route still worked. This is an editorial review process, not an automated publishing lock. It does not complete the separate phone, screen-reader, three-assistant comparison or independent-reader requirements. No new outside AI-product test was performed in this round.

#43 — images with a clear purpose

Removed the abstract decorative illustration from the Lumosity evidence guide and four repeated blank-screen laptop photographs from the learning-platform, website-template, AI-service and small-app guides. Four captions referring to the removed photos were also removed. These articles now lead with their actual subject and guidance. The image files remain recoverable; no media file was deleted.

Full image-by-image purpose and duplicate-decoration review remains open. Removing five decorative placements does not establish that every retained image has been reviewed.

#44 — consistent image framing

All seven topic hubs now use the same square-cornered 3:2 frame with full-image containment. Everyday Life and Privacy & Safety were brought into the shared treatment and given neutral illustrative captions. Desktop checks confirmed the seven frames at approximately 523 × 349px, with the supplied image visible inside each frame and no horizontal page overflow. Remaining card/template consistency and phone-width checks are still required.

#50 and #51 — readable media text and checked contrast

The five existing topic-photo credits increased from 12px to approximately 19px; the two new captions use the same size. All seven hub images retain descriptive alternative text. Removing decorative article images reduces unnecessary image information. Credit links remain underlined and show a visible keyboard focus outline.

Against the actual light-blue hero background, measured caption contrast was 9.53:1 and link/focus contrast was 7.37:1. The changed articles and seven hubs showed no horizontal overflow at the tested desktop width. Separate automated mobile checks on Home and the manual lesson reported 100/100 for accessibility, best practices and basic SEO. These scores do not replace screen-reader, actual browser-zoom, narrow-layout or physical-phone testing; those requirements remain open.

Verification scope: native read-back matched the five changed article bodies and the two changed hub layouts. All seven hubs and five changed articles were checked in a desktop browser. These are AI-operated checks, not independent human testing.

Larger reading text and purposeful photographs

Larger reading text and purposeful photographs

October 24, 2026. Individual content and style changes are counted separately.

#47 — comfortable article reading

The original requirement was to tune shared article width, body size and line spacing, preserve browser zoom, and test long paragraphs, lists and prompts. Standard, Larger and Largest text choices are now available on 40 guides. The choices apply to the current page only and do not store a reading preference. The shared article template retains its reading-width limit and generous line spacing.

In desktop browser checks, ordinary body text measured 21.25px, 26.56px and 31.88px at the three settings. Long paragraphs, lists, a comparison table, recorded-result text and exercise prompts remained readable without horizontal page overflow in the sampled lessons. Keyboard activation worked. The complete 624-character task-list prompt and 725-character manual exercise copied correctly at Largest. Returning to Standard restored the original measured sizes. The viewport permits browser zoom; no zoom shortcuts are intercepted.

Checked lessons: turn notes into a checked task list, compare two options and find answers in a product manual. Control placement was also checked on the career and four-week learning guides. This is sampled verification, not a claim that every guide has been individually tested.

#43 and #44 — image purpose and consistent framing

Three generic illustrations were replaced with credited photographs of a handwritten checklist, notebook planning and laptop practice. Each image now has descriptive alternative text and a clear illustrative role. They appear on the task-list guide, career comparison and four-week skill ladder. All three use a consistent 3:2 frame that keeps the full photograph visible; desktop checks showed the relevant hands and activity intact.

The full image-by-image purpose review, remaining layout consistency and phone-width focal-point checks are still outstanding. A media inventory is not a substitute for visual review.

#50 and #51 — accessible controls and enlarged text

The new controls have a named group, labelled choices, a selected-state indication, a live status message, visible keyboard focus and an underline as a non-color cue. Controls measured at least 52px high with 12px spacing. Three new image descriptions were verified in the live page markup. Screen-reader testing, broader form and landmark review, actual browser-zoom testing, narrow-screen reflow and remaining contrast checks are still open.

Change count: 18 individual changes: three featured-photo replacements, three alt-text entries, three photo-purpose/credit additions, and nine shared control, placement and style changes.

Verification limits: these checks were performed by an AI assistant in a desktop browser. They are not independent human or physical-phone tests. No outside AI product was tested in this round. Print sizing remains separate from the screen text-size choices.

Sources beside the claim, and a verified first step

Sources beside the claim, and a verified first step

October 24, 2026.

  • #4 — Nearby official sources: completed. Readers can follow significant access, privacy and feature claims to the relevant documentation. Links were placed beside the claims in ten articles, the Lumosity study link was corrected, and Gemini and Start Here access guidance was refreshed. Review covered the structure of all 517 published article bodies, a semantic review of 28 core articles and targeted investigation of flagged source locations. The changed sources and published passages were checked. This is source-placement verification, not a fresh factual review of every statement or every external link across the entire archive.
  • #10 — First-time starting route: completed. The optional beginner branch includes one starting route, current access limits, a real first-message-box image and an on-page alternative. An AI-operated desktop walkthrough followed that route and transferred all 394 characters of the fictional prompt through Select prompt text, Copy and Paste into the signed-out ChatGPT message box. Send message became available. The no-account alternative also opened correctly. The existing October 21 screens and actual first-answer/revision record remain dated October 21; no new message was submitted in this October 24 route check.
  • #19 — Complete exercise copying: advanced through verification. The manual desktop route worked and closing its text box returned focus to the control. The automatic Copy button announced success, but this browser’s virtual clipboard did not expose its result, so automatic transfer is not newly verified. Complete phone fallback and remaining cross-lesson checks are still required.
  • #18 — Screenshots and device differences: remains partial. The first-chat images and controls are documented. Completion requires relevant device differences and fresh-start checks of difficult controls across the flagship lessons; it does not require a screenshot for every ordinary text step.
  • #25 — Five flagship demonstrations: remains partial. Actual text demonstrations for first task, email, screen help, sharing data and manual answers are already recorded. Remaining gates include items 18–19 and evidence that another person can follow each lesson. An AI-operated check is not independent human testing.

Separate work count: 22 targeted content edits across 12 articles and Start Here, plus two documentation records. The white, navy and deep-green identity is retained, and no personal publisher identity was added.

Privacy, correction search and discovery checks

Privacy, correction search and discovery checks

Round 44 — October 23, 2026. Five site edits were made: WordPress avatar display was disabled, three passages in the Privacy notice were updated, and the check-record search was corrected to include recent updates.

  • #6 — Privacy notice: advanced. The notice now explains the disabled avatar setting and that remote Lite Connect entry backups are not enabled in the saved WPForms configuration. Saved content and the live rendered notice were verified. Completion still requires confirmed retention and the remaining provider inventory.
  • #7 — Corrections: advanced. The search had omitted recent updates using a second-level heading. It now returns the photo and guide-preview records directly. Mouse and keyboard checks reached the correct headings and moved focus to them. Delivery and owner follow-through remain required.
  • #50 — Accessibility: advanced through verification. Automated mobile checks on Home, Contact, Help and the manual guide reported 100/100 with no flagged issues. Five priority routes each had one main heading. These checks do not replace screen-reader, form-error or complete media-alternative testing.
  • #51 — Contrast and reflow: advanced through verification. The same four mobile accessibility checks found no automated failures. Actual enlarged-text, narrow-screen interaction and touch-spacing checks remain; no successful browser-zoom test is claimed.
  • #53 — Search discovery: advanced through verification. The public crawler file points to the sitemap. Its 517 article URLs and 21 page URLs match the published inventory, without missing, unexpected or duplicate URLs. Home, Start Here, Library, Help and the manual guide each showed index/follow instructions and a matching canonical URL. Search Console checks remain; sitemap inclusion is not proof of indexing or rankings.

Separate counts: five site edits and two documentation records. No images, fonts or palette were changed, and no personal owner information was added.

The browser connection was interrupted during the final privacy-page visual check. The updated public HTML and saved settings were verified through WordPress. No form was submitted, and no real-device or screen-reader test was performed.

Purposeful photographs and clearer form fields

Purposeful photographs and clearer form fields

Round 43 — verified October 22, 2026. Five guides now use photographs connected to the activity: reviewing work together, making skills notes, practising a pottery activity, preparing appointment notes and reviewing news. Each photograph has a descriptive text alternative and a visible credit identifying it as illustrative.

The white, navy and deep-green design is retained. Contact and comment fields now have clearer boundaries. Sampled article text was already comfortably sized, so this round recorded its measurements without increasing it again.

Review scopes: progress and completion requirements
FeatureVerified this roundStill required for completion
#43 — Image purposeAdvanced. Five featured images replaced with activity photographs; each has a recorded purpose.Finish the sitewide image inventory and remove remaining imagery without a useful role.
#44 — Image cropsAdvanced. The five article images retain their natural proportions. Topic-preview geometry was checked at 320 × 210 pixels; Skills Map and Health Advice crops were visually confirmed on desktop.Finish the full image inventory and check important subjects at phone and tablet widths.
#47 — Comfortable readingAdvanced through verification. Sampled article text measured 21.25px with 37.19px line spacing. A long manual guide retained its prompts without horizontal overflow in normal and reading views. No font edit was needed.Verify actual enlarged text and browser zoom across representative paragraphs, lists and prompts. Reading view is not a zoom test.
#50 — Accessible structure and alternativesAdvanced. Five new image descriptions and visible photo credits verified in saved content and live articles.Complete screen-reader checks, landmarks, form labels and errors, and relevant media alternatives.
#51 — Contrast and reflowAdvanced. Two style edits strengthen seven visible fields. Border contrast increased from 2.04:1 to 4.45:1 on Contact and from 1.30:1 to 4.26:1 on comments. Visible focus was checked on both forms.Complete remaining color/control coverage, narrow-screen and enlarged-text reflow, and touch-spacing checks.

Separate edit count: 12 site edits: five featured-image replacements, five photo-credit insertions and two style edits. Five new media assets and two documentation records are tracked separately.

These are scoped AI-assisted checks, not a claim of full accessibility conformance or testing with human participants. No phone, tablet, screen-reader or successful browser-zoom test was completed this round. No form was submitted and no owner identity was added.

Audit update: clearer guide previews

Audit update: clearer guide previews

Verified October 22, 2026. Article previews now state the result, time, level, device and tools before a reader opens a guide. Reading estimates are clearly distinguished from task time.

Round 42 acceptance requirements and verified results
Review scope Requirement and evidence Status and remaining work
#14 — Guide previews Each of 517 published guides has a five-label excerpt: Result, Time, Level, Device and Tools. Verified coverage is 517/517, up from 173/517. Local-model installation and automation previews explicitly identify later-stage setup. Reading times are editorial estimates; task time varies. Completed. All 344 final excerpts match the saved fields. Live checks covered all seven topic pages (84 rows), Latest Articles (12 rows), homepage cues, archive, search and related-guide previews. The checked desktop layouts retained full labels without horizontal overflow.
#50 — Accessibility Source inventory covered all 517 article bodies, including the 28 core guides: 3,668 headings, no body H1s or skipped heading levels, and 88 native controls with markup names or labels. No inline images were present in these bodies. Advanced. Theme markup, featured-image alternatives, rendered behavior and screen-reader testing remain. Source checks alone do not establish accessibility conformance.
#54 — Search and sharing All 517 article bodies were compared with their supplied search and social-sharing metadata; no substantive mismatch was identified. Checks of 538 structured-data graphs found no structural failures. Advanced. The brand author is still typed as a Person. A correction has been prepared but has not been deployed or validated on the site. Passing structural checks does not resolve that semantic issue.

The metadata review checked alignment with the current article bodies. It did not repeat every external-source check, independently test products, run Google Rich Results testing, or perform human reader, physical-device or screen-reader tests. The structural checks used the native metadata output, not an external rich-results certification. The existing white, navy and green styling and 18px topic-preview text were retained; this was not a new photography or mobile-layout review.

Clearer requirements on 45 more guide previews — October 21, 2026

Clearer requirements on 45 more guide previews — October 21, 2026

Both remain partial. Earlier records retain their historical counts.

Original requirements and this round’s result
Audit numberCompletion requirementResult and remaining work
clear guide requirementsShow result, time, level, device and tool needs before a guide is opened; distinguish setup-heavy work.Advanced; still partial. Added all five labels to 45 guides covering privacy, scams, applications, writing, household planning, family support and costs. Complete label coverage increased from 128 to 173 of 517 published articles. 344 article excerpts still lack the complete set. Reading estimates are labelled as reading time, with task time allowed to vary. Tool requirements distinguish reading without an account, optional chat use, source documents and separate setup.
accurate search and share previewsMatch titles, descriptions, headings and social previews to the actual task; validate relevant structured data without invented ratings, credentials or outcomes.Advanced through verification; still partial. Screened all 538 published page/article metadata records; all have search titles and descriptions. There are 31 custom Open Graph overrides. Targeted body checks found no definite mismatch needing an edit. Live checks of the Privacy and Work hubs and the account-privacy guide confirmed aligned metadata; the sampled structured-data blocks parsed as JSON. This is not complete semantic or search-engine structured-data validation. Broader body alignment and structured-data checks remain.

Verification: all 45 saved excerpts matched the intended replacements exactly. A fresh full inventory confirmed 173 complete and 344 incomplete five-label excerpts. Live desktop checks showed complete new labels on the Privacy list and four updated Work entries. A stale Work list was resolved by clearing the site cache and reloading. Preview text remained 18px with 30.6px line spacing. The account-privacy preview opened the intended article; its heading, search description and inherited share description agreed with the visible guide.

Evidence limits: reading estimates use article word counts at roughly 200 words per minute; they are editorial estimates, not measured reader results. Metadata presence is not proof that every claim is current. Valid JSON is not full schema validation. No new human, AI-product, phone, zoom or screen-reader test is claimed. No messages were sent or forms submitted. No owner identity was added.

See the updated Privacy previews · See the updated Work previews

Clearer guide previews and a health-and-money scope review — October 21, 2026

Clearer guide previews and a health-and-money scope review — October 21, 2026

Earlier dated records retain their historical counts.

Original requirements and this round’s result
Audit numberCompletion requirementResult
boundaries for health and money guidanceKeep AI in support tasks; consequential recommendations require qualified review or narrowing/removal.Completed through the narrowing route. Screened the titles, excerpts and categories of all 517 published articles, then reviewed 119 relevant or adjacent article bodies, titles and excerpts plus the Health and Money hubs. Existing guidance keeps AI to records, questions, research explanation and checked arithmetic. No unsupported treatment or investment prescription was identified in that reviewed collection. Clarified one emergency-number instruction and corrected one research publication date.
optional guides and overlapping contentReview fit, available demand evidence and related pages before consolidating; record a reason for each decision, preserve useful URLs and keep valuable niche guides findable.Completed. Retained Fabulous, golf, local AI and automation as four distinct optional guides. Recorded 24 related-page comparisons, clarified each guide’s purpose, added links to broader practice or permission guidance and gave automation a direct route from AI Basics. All four URLs remain unchanged. These are content-fit decisions; no popularity or reader-demand claim is made.
consistent card requirementsShow result, time, level, device and tool needs before opening a guide; distinguish later-stage setup.Advanced; still partial. Home and seven topic lists now allow 700 characters instead of 180. Related previews allow 100 words instead of 60. Both limits fit all 128 currently labelled article excerpts. Topic-list text is now 18px with a comfortable maximum line length. Of 517 published articles, 389 still lack the complete five-label set.
purposeful imagesEvery image has an identifiable reader purpose; remove duplicate or irrelevant decoration.Advanced; still partial. Generic missing-image placeholders are hidden in topic-style lists, leaving clean text entries while retaining real photographs. The full image-by-image purpose and duplication review remains unfinished.

Why the four optional guides remain separate

  • Fabulous: compare a chosen habit with a simple baseline and inspect an app offer and billing. It now links directly to the general tool-value guide.
  • Golf: check completed-round records and prepare instructor questions. It links to the broader skill-practice lesson and does not claim tested equipment rankings.
  • Local AI: check compatibility and plan a fictional trial before installation. Client-data permissions remain a separate linked topic.
  • Automation: plan, inspect and stop one Gmail newsletter-label rule. Readers can first sort five fictional emails without changing account settings.

Verification: six edited article bodies and the final shared CSS matched the intended saved content exactly. All eight changed preview settings read back as 700; the related-preview setting read back as 100. The new Basics link and its original surrounding paragraph were verified. Live desktop checks confirmed complete Home and topic labels, the automation route and its five-message prerequisite, 18px preview text and removal of three placeholder images while a real photograph remained visible.

Material corrections: the health-advice guide now distinguishes the United States emergency number from local numbers elsewhere. The aging-clock guide identifies the cited Perspective’s publication as December 19, 2025, consistent with the publisher and PubMed record. These corrections do not represent clinical review.

Evidence limits: the health-and-money review assessed the scope of advice; it is not qualified clinical or financial sign-off, a complete refresh of every factual claim, or a review of every embedded video. Real patient use still requires appropriate approvals. No demand analytics, human study, product test, real-phone test or zoom test is claimed. No messages, account connections, purchases, redirects or deletions were performed. Owner identity remains unpublished.

See the clearer guide previews · Find the optional automation route · Health guidance · Money guidance

Worked examples across the lesson collection — October 21, 2026

Worked examples across the lesson collection — October 21, 2026

Earlier dated records retain their historical counts.

Original requirements and this round’s result
Audit numberCompletion requirementResult
examples and usable-result checksPair practice prompts with checked results or clearly labelled illustrative examples, give two or three practical checks, and test whether a reader can explain the key check rather than merely reproduce the answer.Completed under the disclosed AI-reader pilot method. Reviewed all 28 current Practice AI guides, four optional guides and Start Here: 33 lesson pages. Added missing worked answers and repaired mismatches. Retained 32 changed-detail AI-reader exercises from this review plus the earlier Start Here trial and a current content check.
consistent card labelsBefore opening a guide, readers can distinguish its result, time, level, device and tool needs, including later-stage setup work.Advanced; still partial. The related-card excerpt allowance rose from 22 to 60 words so the five labels are no longer cut short on the cards checked. The Fabulous excerpt was shortened to 40 words. The 32 curated article excerpts contain all five labels; older archive coverage remains unverified.

What the lesson changes improve

  • Complete examples: added notebook reply and comparison examples, a screen clarification, a meeting agenda, a meal table, a document-naming table and a reminder-writing example.
  • Faithful revisions: added worked shorter, warmer, comparison and conflict variants so readers can check what changed and which facts must stay intact.
  • Visible uncertainty: the trip plan no longer assumes an immediate bag drop at arrival; missing transport and hotel details remain questions. Notes retain a supporting quote and distinguish actions, ideas, dependencies and questions.
  • Specific teaching checks: language practice now includes a labelled dialogue, explained corrections and official reference links. Captions retain family uncertainty, and second-hand comparisons distinguish known subtotals from unknown final costs.
  • Reachable fallback: the answer-verification guide now links to the available “Reveal the four checks” disclosure, which opens without completing the interactive exercise.

Verification: all 26 lesson patches matched exactly once. All 16 saved post bodies matched the intended content in full, with block boundaries preserved. Live desktop checks covered the language exchange, the notes answer, travel constraints and totals, the fallback link and disclosure, and complete card labels. The excerpt setting was read back as 60.

Evidence limits: these are AI-reader exercises, not human usability studies or independent product tests. Reviewers could see the lesson examples; the exercises were not blinded. Start Here uses retained reader evidence with a current content check. No pronunciation, live booking, paid-app, real-phone or browser-zoom result is claimed. No messages, purchases or subscriptions were made. Owner identity remains unpublished.

Read the new illustrative language exchange · Check the worked travel plan · Browse the maintained practice guides

A reusable kit and a checklist you can return to — October 21, 2026

A reusable kit and a checklist you can return to — October 21, 2026

Earlier dated records below retain their historical counts.

Review scope requirements and this round’s result
Audit numberCompletion requirementResult
reusable project kitA completed example, editable worksheet, verification checklist and adjacent preview; independent use and actual reuse.Completed. An AI reader used the downloaded file for Riley’s room reply, then reused the blank worksheet for Alex’s agenda request without website assistance. The checked drafts contained 30 and 37 words. The worksheet now allows “none found” after checking and names the earlier task used for comparison.
browser progress checklistSix tasks, browser-specific saving, reset and a print route; return and continue; lessons remain accessible after clearing or storage failure.Completed. Two saved ticks returned after leaving for a lesson; a third task was continued. Download, text selection, reset and a fresh visit worked. Damaged saved JSON now follows the unreadable-data recovery path. Fourteen focused unit checks passed against the deployed component.
email-prompt builderEditable local output, demonstrable improvement over a static prompt, and browser/network checks of the stated data behavior.Advanced; still partial. New guidance explains that entries are temporary and should be saved before leaving. Edited text was preserved in Copy and the actual download; fields were empty after reload. This comparison found no correctness improvement over an already-correct static prompt. Production network behavior remains unverified.

What changed and what was verified

  1. Kit error check: “Error I removed, or ‘none found’ after checking” no longer implies that every draft must contain a mistake.
  2. Kit reuse note: readers name the earlier task, then record what changed in the new task.
  3. Checklist recovery: invalid JSON no longer gets mistaken for unavailable browser storage. Genuine storage failures still leave task links, visit-only ticks and selectable text available.
  4. Builder save guidance: readers are told that leaving or reloading can lose temporary entries, and to save their edited prompt first.

The additional version-stamp edit dates the kit October 21. The final downloaded kit matched the original file with exactly the three intended text replacements. The actual checklist export contained three checked and three unchecked tasks, all six lesson links and a return link. Reset left no ticks and remembering was off on the next visit. The print route is the documented text-file or copied-text workflow; no physical printer test is claimed.

Evidence limits: these were disclosed AI-reader pilots, not human usability studies or evidence of real-world email outcomes. The kit’s worked answer had been seen, so its pilot was not blinded. Checklist failure cases ran against the actual component in a local unit-test fixture; return, saving, downloading and reset were also checked separately in the live desktop browser. No real-phone, browser-zoom or production network-capture result is claimed. No messages were sent.

Preview or save the complete reply kit · Open the six-task progress checklist · Use the editable email-prompt builder

Three completed practice pilots — October 21, 2026

Three completed practice pilots — October 21, 2026

This round completes #33, #34 and #37 through an AI reader pilot and live browser checks, using the owner’s requested reader-assistance method. No human participant study is claimed.

RecommendationOriginal completion requirementVerified result
#33 — Five-skill learning pathShow the five skills and intended results; complete a different second task without a new account or course enrollment.All five path links and section targets worked. The AI reader completed and checked the first travel checklist and a different room-confirmation reply. A new link after the path now leads directly to that second exercise; the live link and explanatory reveal worked. Both actual results and their checks were retained.
#34 — Three mistake exercisesIdentify an invented detail, wrong total and unsupported source claim; explain the check; operate the reveals by keyboard.A separate AI reader received only the unrevealed exercises and correctly identified all three problems. Its reasoning matched the explanations. Enter, Tab, Space, Tab and Enter reached and opened all three reveals. Clearer instructions now ask for the error, source evidence and next check; Practice links directly to the full set.
#37 — Three-question chooser: completed pilotOffer one suitable lesson and one alternative from task, experience and time; verify every route and compare the choice with the ordinary Library.All 12 combinations produced the expected visible result and focus change. All nine distinct lesson destinations rendered. In three matched AI-reader scenarios, the chooser made the fit clearer for weekly planning and a short source-check exercise. The Library shortcut was easier for quick email practice. A direct Library bypass is now beside the chooser.

What the AI reader actually completed

First task: a checklist retaining the raincoat, charged phone, train ticket, Saturday 9:30 a.m. departure and still-undecided return train. The reader improved “Departure” to “Train departure” without changing facts.

Different second task — actual AI reader draft, not a sent message:

Hi Riley! You asked us to confirm a room for Friday at 6 p.m. by Wednesday. Room approval is still pending. I can provide 12 chairs. Would 12 chairs be enough?

The reply has 31 whitespace-separated words, below the 70-word limit. It keeps Friday at 6 p.m. and Wednesday as Riley’s requests, leaves approval pending, offers only 12 chairs and asks whether that is enough. The reader changed “Would that be enough?” to “Would 12 chairs be enough?” for clarity. The source prompt, first drafts, improved results and checks were retained in the working evidence record.

Mistake-check evidence

  • Invented detail: an appointment was offered as the delay reason even though no reason was supplied. Remove the invented reason.
  • Wrong total: $18 + $12 + $6 is $36, not $42. Recalculate the supplied figures; do not invent a fee.
  • Unsupported claim: pending room approval does not establish a confirmed booking or website evidence. Keep the status pending and check an actual confirmation before claiming otherwise.

Where the chooser helps—and where the Library is simpler

The Library already provides a direct two-minute email check, so the chooser adds unnecessary decisions for someone who wants that exact activity. Its visible shortcuts do not include a weekly plan or a citation check. For those sampled needs, the chooser exposes a suitable activity and its time estimate without requiring a search phrase or a search inside a longer guide. This is a conditional AI-reader assessment, not a measured speed gain or evidence that all readers prefer the chooser.

Five site edits: a direct heading destination for all three mistake types; clearer before-reveal reasoning instructions; a Practice-page link to that exercise set; an explicit second-task link after the five-skill path; and a Library bypass beside the chooser.

Limits: these checks used AI readers and a desktop browser. They do not complete the separate human-reader observation, real-phone, zoom or narrow-reflow requirements. No new account, external prompt submission, contact message, backup or restore was involved. Earlier round records below retain their historical counts.

Larger photos and clearer lesson text — October 21, 2026

Larger photos and clearer lesson text — October 21, 2026

This round advances #44, #47 and #51; none is newly complete.

Review scopeCompletion requirementThis round and remaining work
#44 — Image proportions and focal pointsConsistent image treatment, with important subjects intact at desktop and phone widths.Advanced: all four homepage task frames now measure 360 × 240px on desktop and preserve the full source image. Photo credits are larger; rows stack below 900px in the stylesheet. Other image families and phone-width focal-point checks remain open.
#47 — Comfortable article readingMeasured reading width and spacing, with long paragraphs, lists and prompts readable and copyable at normal and enlarged sizes.Advanced: fact labels, fact values, preparation/cost notes and manual-copy text are larger. The email article measured 760px wide with 21.25px body text and 37.19px line spacing. Long text fit; all 306 characters of the exercise copied, and the manual selection and Escape return worked. Enlargement remains unverified.
#51 — Contrast, enlargement and reflowRecord and fix contrast failures; verify enlarged text, zoom, narrow reflow, control spacing and non-color cues.Advanced: lesson links are larger and underlined. Sampled text contrast was at least 6.53:1; the sampled input boundary measured 3.59:1 against white. Navigation controls measured about 51px high. The email page scored 100/100 in the automated mobile accessibility check. Manual zoom and narrow-screen checks remain open.

Ten visible edits: larger fact labels; larger fact values; larger and better-separated lesson notes; larger, underlined desktop lesson navigation; larger responsive lesson navigation; wider homepage photo columns; larger homepage photos; larger photo credits; earlier stacking of photo/text rows; and article-sized manual-copy text.

Verification limits: these were AI-assisted reader and technical checks at a 1363 × 936px desktop viewport. Browser zoom shortcuts did not change the viewport, so no enlargement or phone-width pass is claimed. The automated score is not a complete accessibility certification. An explicit reload showed the latest email content and styles after an initially stale page. No independent human-reader study, real-device test or new product trial was performed.

AI reader walkthrough and clearer practice routes — October 21, 2026

AI reader walkthrough and clearer practice routes — October 21, 2026

This round advanced #10, #20, #25, #60 and #72.

An AI assistant reviewed the five flagship lessons and made eight small lesson changes. This was an informed AI review, not a blind beginner test or evidence from five human readers. A neutral reader task sheet, observer protocol and results dashboard remain private drafts for a later trial.

What changed

  • First task: the beginner setup now links directly to step 1 and the complete fictional prompt.
  • Email: the opening navigation now includes the exercise that needs no AI account.
  • Screen help: the complete fictional prompt has a direct link, distinct from the blank template for another screen.
  • Manual answers: the complete fictional exercise comes first in the lesson links; the template for a real manual is clearly labeled.
  • Sharing information: the section link describes the check accurately. The finish check now keeps date, location, budget and helpers blank, avoids assuming a currency, and asks why private guest records are unnecessary.

Original requirements and status

FeatureProgress in this roundStill required for completion
#10 · Beginner starting routeAdvanced: direct setup-to-prompt link; the earlier browser checks in this session reached the chat message box and copied all 394 prompt characters.A real beginner reaches the first prompt without assistance; verify the alternative when access is unavailable.
#20 · Checked examplesAdvanced: compared the five flagship source/example checks and clarified the sharing completion check.Reader explanations of the decisive checks, plus remaining maintained-set coverage.
#25 · Five flagship guidesAdvanced: clearer entry points across first task, email, screen, sharing and manual lessons; existing dated demonstrations preserved.Another person follows each from a fresh start, with retained results and all requirements in #17–21 satisfied.
#60 · Observe five readersPreparation advanced: neutral tasks and an observation protocol are drafted. AI observations are labeled separately.Five willing readers of varied experience, including adults in their 40s and 50s; record help and repeated obstacles, fix them and retest. AI review does not replace this requirement.
#72 · Results dashboardAdvanced: a private dashboard separates human results from AI checks, with an initial editorial improve decision.A real baseline and monthly keep, improve or stop decisions using completion, findability, usefulness, return use, corrections and measured maintenance time.

Verification and limits

The homepage first-task link, beginner disclosure, complete-prompt copy and opening chat route were observed in desktop Chrome. The homepage email link also reached the correct lesson. The browser connection then failed; the email answer-reveal action and the three other interactive exercises were not verified in that browser session. No new external AI prompt was submitted.

After editing, all five saved pages matched the intended changes. Their public rendered pages contained the revised wording and the required link destinations exactly once. The rest of the saved content was preserved. This verifies content and destinations; it does not establish final click behavior, phone layout, comprehension, usefulness or return use.

Complete source material for practical exercises — October 21, 2026

Complete source material for practical exercises — October 21, 2026

Original requirements and remaining work
FeatureCompletion requirementStatus
Starting requirementsConsistent result, time, level, device and tool needs distinguish short practice from setup-heavy work before readers commit.Advanced: five overviews and matching excerpts identify a beginner exercise using fictional text, with optional AI access and no new account or upload. Remaining archive coverage and entry-route verification are incomplete. Ten minutes is an estimate, not a reader-test result.
Checkable resultsPair prompts with checked results or clearly labeled illustrations, decisive checks and evidence that readers understand the key check.Advanced: all five exercises now include their source material, illustrative answer, three checks, repair wording and an explanation question. No actual AI-product response or independent reader-comprehension result is claimed. Wider coverage remains open.
Accurate previewsTitles, descriptions, headings and share previews reflect the task; relevant structured data validates without invented outcomes.Advanced: five descriptions now match the exercise. Public search descriptions, Open Graph descriptions and canonical URLs were verified. Wider collection coverage, actual social-platform display and formal schema validation remain open.

Five exercises changed

Checks against the source material

  • Customer answers: $120 includes one revision requested within seven calendar days after delivery. Extra revisions, their cost and next-day delivery remain unapproved.
  • Customer feedback: two size-chart messages, one tracking-link message and one positive comfort message total four messages. These do not establish a unique-customer count or a population percentage.
  • Finding notes: the April 10 location change supersedes April 3; gloves remain required. No start time, parking policy or calendar year is invented.
  • Voice-note outline: twelve refers to packets, the year is uncertain and attendance is unknown. No income or success claim is added.
  • Tutoring notes: 20% of $80 is $16, leaving $64; 20% of $50 is $10, leaving $40; 10% of $90 is $9, leaving $81. The answer key distinguishes a discount from the final price.

Publication verification

Each exact patch matched once. Saved article bodies and excerpts matched the prepared versions. All five public exercise sections contain the full prompt, illustrative result, checks, repair and finish condition. All five public heads display the intended search and Open Graph descriptions and unchanged canonical URLs. New sections use white, navy and deep-green rules without added imagery, decorative emoji or rounded-card layouts. Public attribution remains AI Over 40.

Separate automated checks

Mobile Lighthouse accessibility and basic SEO scans returned 100/100 for Home, Start Here and Content Library, with no flagged issues in those categories. No changes were made solely to improve these scores. These automated results do not establish real-device usability, screen-reader compatibility, search indexing or complete accessibility.

No new product-interface, external-source, browser-interaction, mobile visual-layout or independent learner test was performed for the five article edits. The examples are editorial illustrations, not reported AI responses. Next evidence needed: wider requirement/preview coverage and readers explaining the checks without assistance.

Private health worksheets and fictional practice — October 21, 2026

Private health worksheets and fictional practice — October 21, 2026

Original completion requirements and remaining work
FeatureCompletion requirementProgress and limits
Starting requirementsConsistent result, time, level, device and tool requirements distinguish practice from setup before readers commit.Advanced: five overviews and five excerpts now state the task and optional access. These exercises need no new account, upload or purchase. Full archive coverage remains incomplete; times are estimates.
Checkable examplesPrompts have checked results or labeled illustrations and decisive checks that readers can explain.Advanced: three blank-sheet examples and two fictional text exercises now include complete prompts, illustrative answers, three checks, repair requests and explanation questions. Independent reader comprehension and remaining collection coverage are still unverified.
Safe health and finance scopeKeep tasks within organization and questions; consequential recommendations require appropriate review or narrowing/removal.Advanced: sleep-diary practice now uses fictional entries, the guide no longer asks AI to reformat a clinician’s plan, and medication practice clearly limits AI to an empty worksheet. Clinical decision boundaries and the stroke emergency section remain. Wider collection review remains open; no clinical review is claimed.
Purposeful mediaEvery retained image or instructional visual has a clear explanatory purpose; remove unsupported or redundant decoration.Advanced: removed the medication article’s embedded video, for which no retained content review or transcript was available in the guide. This was an embed removal, not deletion of a media asset. Complete image-by-image review remains open.
Accurate previewsTitles, descriptions, headings and share previews reflect the actual task; relevant structured data validates without invented outcomes.Advanced: five descriptions and the medication/sleep-diary titles now describe the limited exercises. All five public heads show the intended descriptions, inherited Open Graph descriptions and unchanged canonical URLs. Social-platform display, formal schema validation and wider coverage remain open.

Guides changed

What was verified

  • All 33 bounded changes matched once. Saved article content, titles and excerpts exactly matched the prepared versions.
  • All five public practice modules contain the complete prompt, explicitly illustrative result, repair request and completion check.
  • Blank appointment, rehabilitation and medication sheets contain five, three and six requested labels, respectively, with no supplied medical values.
  • The class answer key matches Tuesday, 6 p.m., Room B, an apron and Sunday, without inventing a calendar date or saved reminder.
  • The fictional sleep intervals each equal eight hours in bed, not known time asleep. The first tracker estimate stays separate; the second is missing, not zero. No two-night tracker average is calculated.
  • The medication video is absent from the public article body. Search and Open Graph descriptions and canonical URLs were checked on all five live pages.

Evidence and boundaries

Official NHLBI sleep-apnea and insomnia diagnosis pages, FDA interaction guidance, MedlinePlus rehabilitation and CDC stroke warning signs were consulted for the specific boundaries recorded in the guides. This was not a comprehensive recheck of every clinical statement. NIA’s cognitive-health page returned a robot-verification screen when opened; the memory guide explicitly makes no fresh clinical source-review claim. No attempt to bypass that screen was made.

No AI-product trial, independent reader, visual-layout, keyboard, mobile or screen-reader test was performed this round. The examples are editorial illustrations. New sections retain white backgrounds, navy text and deep-green rules; no new imagery, decorative emoji or rounded-card layout was introduced. Publisher attribution remains AI Over 40. Next checks: reader comprehension, the remaining collection and the original device/accessibility requirements.

Optional guides: choose a task and check the result — October 21, 2026

Optional guides: choose a task and check the result — October 21, 2026

Completion requirements and remaining work
Original itemRequirementResult and status
Sources at claimsReaders can trace important access, privacy and feature claims to relevant official documentation.Advanced: clarified the golf rules source and added dated LM Studio setup documentation beside the steps. Full maintained-collection coverage remains open.
Starting requirementsConsistent result, time, level, device and tool needs distinguish short practice from setup-heavy work before a reader commits.Advanced: golf now has an opening overview and matching archive excerpt. The 10-minute estimate covers fictional practice, not equipment setup. Wider archive coverage remains open.
Checkable examplesEach prompt has an actual checked result or labeled illustration, decisive checks, and evidence that readers understand the key check.Advanced: golf now includes all four source notes, an illustrative result, three checks, repair wording and two explanation questions. No independent reader-comprehension test was performed.
Guide fit and overlapInventory weak-fit and overlapping guides; record reasons, consider demand and links, and preserve useful URLs or redirects.Advanced: reviewed Fabulous, golf and both local-AI guides and their internal references. Retain the optional niches and separate installation from data-permission guidance. No page was merged or removed. Reader-demand evidence and the wider overlap inventory remain outstanding.
Accurate previewsTitles, descriptions, headings and sharing previews match the task, with valid relevant markup and no invented outcomes.Advanced: revised the golf description; the public head shows the new search and Open Graph description and preserves the canonical URL. Wider metadata review, actual social-platform display and formal schema validation remain open.

Keep decisions and why

  • Fabulous: retain as optional habit-app buying guidance. Its existing fictional exercise compares a new offer with a working reminder; it does not establish an AI feature or medical benefit. No further edit to this guide was needed.
  • Golf records: retain as an optional application of checking source information. The new exercise works without equipment or a purchase and belongs after a round, not in live playing advice.
  • Local-AI installation: retain for compatibility and a small fictional setup trial. The source note distinguishes a model download from loading it into memory; no installation was performed.
  • Local AI and client data: retain for data handling and permissions. Added a direct route to the installation guide while keeping its separate privacy purpose. Existing URLs and next-lesson routes remain.

Internal-reference checks covered published post/page bodies and Elementor data for these four URLs. References were found in Tools & Reviews, Help, editorial records and related local/client-data articles. This is not an external backlink audit or evidence of reader demand; no traffic performance is inferred.

Verification and limits

Exact content patches matched once. Public rendered HTML includes the golf overview, complete N1–N4 prompt, labeled answer and checks, the local setup source note and the client-data routing paragraph. The golf head has the intended description, Open Graph text and canonical URL. The answer was compared with the supplied notes: N2 is six including the penalty, N3 remains uncertain and N4 supplies no distances or video. Official Arccos sensor guidance, R&A Rule 4.3 and LM Studio getting-started documentation were reviewed on October 21. No AI-product, keyboard, mobile, visual-layout, screen-reader or independent learner test is claimed in this round. New content uses the established white, navy and deep-green styling; no new imagery was added. Public attribution remains AI Over 40.

First chat in five screens — October 21, 2026

First chat in five screens — October 21, 2026

The 84-second silent captioned walkthrough is now published with reading pauses, a complete checked result and a full written transcript. It uses actual still captures, not continuous screen recording. Desktop playback, pause, seeking to the final result and keyboard access to the transcript were verified. This follow-up adds one lesson edit (the video player, caption and transcript/download links); record updates are bookkeeping. Items #10, #18, #25 and #50 remain advanced, with their previously recorded checks outstanding.

Open the new five-screen walkthrough. It uses actual product captures, large instructional images, 20px body text and 18px captions, with a white, navy and deep-green presentation. Previous, Next and Show all controls support a reader-paced sequence. A complete written version includes the prompt, both observed answers and the checks.

Review scopes, completion requirements and remaining evidence
FeatureCompletion requirementWork in this roundStatus and remaining work
First-use branchA beginner reaches the first prompt unassisted using a verified route; an alternative is available if access fails.Updated the signed-out desktop route, dated first-message image and access/history guidance. The existing no-account lesson remains available. Manual copying transferred the full prompt.Advanced. An unassisted beginner and fresh-start device check are still required.
Clear steps and real screenshotsA fresh-start walkthrough leaves no unexplained click; cropped, annotated screenshots match the controls and relevant device differences are verified.Added five numbered screens with explicit instructions for the message box, Send control, first answer, same-chat follow-up and final check. Each image has a “What to notice” caption and full-size link.Advanced. Desktop controls match this observed session. Phone/touch layouts and an independent fresh-start walkthrough remain unverified.
Five flagship demonstrationsAll five flagship guides meet recommendations 17–21, retain actual inputs, outputs, corrections and conditions, and can be followed by another person from a fresh start.Added visual evidence and a new complete first-task trial. The earlier five actual text trials remain available; the other four were not rerun in this round.Advanced. Remaining screenshots, device coverage and independent reader checks across the five guides are not yet complete.
Captioned screen demonstrationA short captioned video has pauses and a transcript, shows a complete checked result, and leaves the written lesson sufficient.Published the 84-second silent video using five actual captures, readable captions and pauses. The image sequence and complete written walkthrough remain available.Completed in the October 21 publication follow-up. Desktop playback, pause, final-result seeking and keyboard transcript access verified. The video is explicitly labeled as a sequence of actual still captures, not a continuous screen recording.
Accessible structure and alternativesScreen-reader checks communicate page structure, controls and important image information, including form labels/errors and media alternatives.Added descriptive image alternatives, explanatory captions, a complete text version, keyboard controls, a status announcement and stable focus at both sequence boundaries.Advanced. Desktop keyboard and accessibility-tree inspection passed the scoped checks below; this is not a screen-reader or full-site accessibility pass.

What was changed and checked

  • Opened the walkthrough with Enter; used Next and Previous with the keyboard; checked all five screen states, both boundary states and retained focus.
  • Used Show all, checked that all five sections became visible, and returned to one-screen mode. Opened the complete written walkthrough with the keyboard.
  • All five hosted images loaded. The narrow follow-up image now includes the complete message and Send control. Visible captions, text size and focus indicators were inspected in desktop Chrome.
  • Source markup retains all five screens and the text alternative without requiring the control script. JavaScript-disabled browsing itself was not tested.
  • The MP4 was checked for 84-second duration, H.264 playback format, readable first/follow-up/final frames, and a complete checked final result. It contains no audio or personal account information.

Actual first-task trial

Date and conditions
October 21, 2026, approximately 11:30–11:39 UTC. An AI assistant operated desktop Chrome and signed-out ChatGPT. The model/version was not displayed. Only the fictional text below was submitted; no account, payment, attachment or provider-setting change was used.
Prompt
Turn these fictional notes into a short checklist. Group it into “Before leaving” and “Still to decide.” Keep every detail. Do not invent missing information.

Notes: Pack a raincoat. Charge my phone. Bring my train ticket. My train leaves Saturday at 9:30 a.m. I have not chosen a return train.

Put the departure time above the checklist. Keep choosing a return train under “Still to decide.”
Observed first answer
Departure: Saturday at 9:30 a.m.

Before leaving
Pack a raincoat.
Charge my phone.
Bring my train ticket.

Still to decide
Choose a return train.
Follow-up
Make the checklist easier to scan without changing any facts.
Observed revised answer
Train Travel Checklist

DEPARTURE
Saturday · 9:30 a.m.

Before leaving
Pack a raincoat
Charge my phone
Bring my train ticket

Still to decide
Choose a return train
Checks and corrections
Both answers kept Saturday at 9:30 a.m., the raincoat, phone and ticket, and the undecided return train. No station, journey length or return booking was invented. No correction was necessary in this attempt. The transcript simplifies the visual checkbox presentation; another response may differ.
Copying check
The lesson’s Copy prompt button reported success, but the automation’s virtual clipboard was empty. The manual route—Select prompt text, Copy, then Paste into ChatGPT—transferred all 394 characters. This is a scoped automation result, not proof that readers’ clipboard buttons fail. Touch-device copying was not tested.

Sources checked October 21: OpenAI’s signed-out access guidance (new tab) for regional access and session-only history; Free-tier FAQ (new tab) for tool limits and account conditions. Documentation review does not establish access in every region or on every device.

Limits: this was one AI-operated fictional desktop attempt. No independent beginner study, phone or tablet test, screen-reader session, account-creation test or service-reliability claim is recorded. Public attribution remains AI Over 40.

Clearer health and money practice — October 20, 2026

Clearer health and money practice — October 20, 2026

Deferred items remain 58, 62, 63 and 64.

Original requirements and current status

AuditCompletion requirementThis round and remaining work
Traceable sourcesConnect significant access, privacy and feature claims to relevant official documentation near the claim.Advanced. Added dated, claim-specific checks or limitations in 13 guides. Corrected the aging-clock paper to a 2026 Perspective and distinguished that analysis from consumer-product validation. The EPA source could not be retrieved and is explicitly not labelled freshly verified. FitXR membership and Fabulous product/billing documentation were also checked within their limited product-description scope. Full source coverage remains incomplete.
Clear starting requirementsConsistently show result, time, level, device and tools before readers commit to a guide, distinguishing practice from substantial setup.Advanced. Added structured entry panels and matching native excerpts to 13 guides. Device purchases, clinical setup, document review and professional assessment are separated from short practice estimates. Uniform coverage across all maintained entry routes remains to be verified.
Appropriate health and financial scopeKeep AI tasks to preparation, organization, assumptions and arithmetic; obtain qualified review or narrow/remove consequential recommendations.Advanced. Thirteen fictional exercises stay within those support tasks and state what the answer must not infer. The health-organizer introduction now explicitly keeps personal information in a privately filled structure. This is not qualified clinical or financial review of the full collection; remaining consequential claims still need that review or further narrowing.
Purposeful imagesGive each image an explanatory role and remove duplicate decoration or off-brand imagery.Advanced. Removed five decorative featured photos and two video embeds without a retained verification record or transcript. Kept the four existing photographs showing monitoring equipment, a clinical conversation, research-note arithmetic and VR hardware. Media files were retained; the whole-site image-purpose review is unfinished.
Accurate previews and markupAlign titles, descriptions, headings and sharing previews with the actual task; validate relevant structured data without invented outcomes or credentials.Advanced. Aligned 39 description fields across 13 guides. Verified their live canonical URLs, preview images, anonymous author name and parseable JSON-LD, with no rating fields added. This is not full schema conformance, crawler validation or proof of rankings; wider collection coverage remains open.

Affected guides

Verification and limits

All 13 stored bodies and excerpts matched the intended changes exactly. Titles, addresses, original publication dates and the intended featured-image state were preserved or changed as recorded. All 13 public heads showed the intended search, Open Graph and X descriptions, correct canonical address, AI Over 40 as author, and the expected image presence or absence. Their JSON-LD parsed, and no aggregate-rating or review-rating fields were present.

In desktop Chrome at a 1363 × 936 viewport, the health-organizer and aging-clock answer disclosures opened with the Enter key. The health-organizer article used a roughly 633-pixel reading column, 21.25-pixel body text and 37.19-pixel line height. Both checked layouts had no horizontal overflow. The corrected 2026 research reference was visible and the aging-clock video embed was absent. One automated mobile accessibility scan of the updated health organizer reported 100/100 with no flagged issues.

The white, navy and deep-green treatment uses open rules, readable type and unrounded entry sections. No decorative emojis or generated imagery were added. The new worked answers are explicitly illustrative. No product trial, independent reader success, real-phone/tablet check, browser-zoom test, complete screen-reader assessment or qualified clinical/financial review is claimed. Existing source-review dates remain historical; this round’s dated notes identify its narrower source scope.

Assistant trial, source dates and access limits — October 20, 2026

Assistant trial, source dates and access limits — October 20, 2026

Deferred items remain 58, 62, 63 and 64.

Original requirements and current status

AuditCompletion requirementResult and remaining work
Traceable official sourcesPlace relevant official documentation beside significant access, privacy and feature claims.Advanced. The guide now distinguishes October 20 source checks from Gemini’s earlier October 1 review, which could not be refreshed. Microsoft and Perplexity documentation is separated from the access actually observed. Full maintained-guide coverage and unverified claims remain open.
Original assistant comparisonTest three assistants on identical everyday tasks; retain plans, prompts, outputs and limitations, with a dated conclusion and reasons to skip an option.Advanced, incomplete. One invitation and one revision were completed in signed-out ChatGPT. The draft’s unsupported promise was identified and removed in the revision. Perplexity returned a sign-up message; Copilot reached sign-in. Three comparable completed outputs, more than one everyday task and comparable checking effort are still missing. No winner is claimed.
Comfortable article readingVerify shared reading width, body size, spacing and copyable long content at normal and enlarged size while preserving browser zoom.Still partial. No new shared-template change was needed for this content update. The enlarged-text and browser-zoom checks could not be completed, so this recommendation remains open.
Structure and alternativesVerify headings, landmarks, image alternatives, labels, errors and media alternatives with screen-reader checks.Checked, still partial. Automated mobile accessibility scans of the Content Library, Help, product-manual lesson and updated comparison guide each reported 100/100 with no flagged issues. These scans do not establish a complete screen-reader or form-error pass. The final interactive disclosure/copy check was interrupted and is not claimed as passed.
Contrast and reflowMeasure contrast and verify enlarged text, browser zoom, narrow reflow, tap spacing and non-color cues; fix recorded failures.Still open. Automated scans provide limited evidence only. Manual enlarged-text, zoom, reflow and interaction checks remain incomplete; no full accessibility conformance is claimed.

What changed and what was verified

The guide now separates documented options, invented teaching examples and the observed trial. It includes the revision request and actual revised answer, access limitations, account and device conditions, reasons to skip an option, and updated evidence labels. New content uses the existing white, navy and deep-green article treatment; no decorative imagery or emojis were added.

The stored body and excerpt matched the intended edits exactly. The existing title, address and original publication date were preserved. The live trial section and source labels were read back, and the page’s canonical address and public author identity, AI Over 40, were verified. The updated guide’s automated mobile accessibility scan reported 100/100. No independent human verification, paid-plan assessment, timed productivity comparison or completed keyboard/screen-reader/zoom test is claimed.

Evidence limits: the trial used fictional information in a signed-out cloud Chrome session. No account was created, no subscription was purchased and no private records were uploaded. Provider documentation and this one session do not establish universal access conditions or typical performance. The three-assistant recommendation remains unfinished.

Purposeful photographs and clearer guide entry — October 20, 2026

Purposeful photographs and clearer guide entry — October 20, 2026

Deferred items remain 58, 62, 63 and 64.

Fifteen guides now explain what readers will prepare, how long the initial task may take, their starting level and the device and tools needed. Times are planning estimates rather than measured reader results. Sixteen decorative featured illustrations were removed; hiking and reading now use photographs of the actual activity. The layout uses white, navy and deep green, open rules, larger type and unrounded image frames.

Original requirements and current status

AuditCompletion requirementThis round and remaining work
Guide entry requirementsShow a consistent result, time, level, device and tool requirement before opening a guide; distinguish short planning tasks from setup or production.Advanced. Added structured overviews and native excerpts to 15 guides. Avatar video and digital-product guides clearly separate preparation from production. Existing overviews on the other three guides were retained. Full archive coverage and verification of every entry/listing remain open.
Purposeful imagesGive every retained image a clear activity, instruction or result role; remove decorative, duplicate or off-brand imagery.Advanced. Removed 16 decorative featured illustrations and replaced two others with visually inspected real photographs of navigation and book selection. All 18 obsolete captions were corrected. The inventory covered 517 published articles, but this was a focused review of 18 explicitly identified illustrations, not a complete visual review of every site image.
Image proportions and reading layoutUse consistent image proportions, preserve important subjects and check images with adjacent text on desktop and phone.Advanced. The two photographs use contained 3:2 frames, no rounded border and no text overlay. Both rendered at approximately 633 × 422 pixels in a 1,363 × 936 desktop viewport, with no horizontal overflow. The new summary text rendered at 21.25 pixels. Wider template, phone, tablet and zoom checks remain open.
Accessible structure and alternativesProvide meaningful image alternatives, logical structure, usable labels/errors and accessible media; verify the complete route with assistive technology.Advanced. Added factual alternatives for the two photographs and labelled H2 summary sections with result paragraphs and definition lists on 15 guides. Checked the rendered summaries and sampled browser accessibility-tree reading order. A browser accessibility tree is not a screen-reader test; full keyboard, form, media and assistive-technology coverage remains open.
Accurate previews and markupAlign titles, descriptions, headings and social previews with the page and validate relevant structured data.Advanced. Detected and corrected stale share-image data through Yoast for all 18 guides. Public output now shows the two new photographs or omits a removed article image. Verified all 18 canonical URLs, anonymous author labels and parseable JSON-LD. Text metadata was preserved. Complete sitewide coverage and formal external validation remain open.

What was verified

  • Read back all 18 native articles. The 35 bounded body changes, 15 excerpts and 18 featured-image settings matched the prepared result. Titles, public addresses and original publication dates were preserved.
  • Checked the rendered article output and all 15 new summary sections. Original media files were retained, and no article was deleted.
  • Checked the public head of every affected guide after the Yoast refresh. All 18 old illustration addresses were absent, both new photo addresses were correct, and no removed article image remained in Open Graph output.
  • All 18 JSON-LD graphs parsed; author metadata remained AI Over 40 and canonical addresses matched the existing guides. This is not formal schema validation or a guarantee of how an external social service caches a preview.
  • Visually checked both photographs on desktop: the compass/hand and the hand/book remain in frame, with a factual alternative and linked photographer credit. The avatar guide’s browser reading order clearly separated script planning from video production.

Photo sources: navigation photograph by Jamie Street; book-selection photograph by Christin Hume, both on Unsplash. These are real photographs, not generated instructional demonstrations. The avatar-account wording was checked against Synthesia’s generation instructions.

Limits: this was an AI-assisted presentation and metadata check. It was not a new complete factual review of the existing lessons, independent human usability testing, a phone/tablet/zoom test, a screen-reader test or external schema validation. No personal identity, private prompt or user account information was published.

See the 18 affected guides
Clearer feedback, privacy and correction records — October 20, 2026

Clearer feedback, privacy and correction records — October 20, 2026

Deferred items remain 58, 62, 63 and 64.

Guide feedback now includes Text or photos that are hard to read, carries the guide and selections to an editable Contact message, and opens directly at the relevant section. Feedback and Contact controls use larger type with white, navy and deep-green styling and ruled sections. The privacy notice names the verified form/email configuration and explains what enters the feedback page address.

Original requirements and current status

AuditCompletion requirementThis round and remaining work
PrivacyInventory collection and services; publish matching handling and retention information at collection points and in the footer.Advanced. Verified WPForms Lite, WP Mail SMTP and the Gmail mailer; added feedback-address disclosure and a direct privacy link beside the choices. Provider/hosting retention, any additional form storage and any needed specialist review remain unconfirmed.
CorrectionsCarry the exact guide and issue into a report, publish material changes, and verify delivery and publisher follow-through.Advanced. Added the readability route, direct arrival and public review responsibility. Recorded a real earlier size report with a desktop recheck below. Contact-message delivery and complete publisher follow-through remain unverified.
Search discoveryCheck indexing eligibility, robots, canonical URLs and intended sitemap coverage using public checks and Search Console; resolve demonstrated conflicts.Blocked. Archive page 2 shows different articles but identifies page 1 as its canonical. A narrow server-side correction is prepared privately; it is not deployed. Changing the WordPress posts-page setting would replace the existing Elementor archive layout. Search Console evidence remains unavailable.
Accurate previews and markupAlign titles, descriptions, headings and social previews with the actual page, and validate relevant structured data.Advanced; external validation blocked. Updated Contact descriptions and Privacy descriptions/social titles through Yoast, then checked the live head and schema types. The external validator presented a robot check. This is not formal validation or complete archive-wide metadata coverage.
Feedback and reviewOffer low-friction reasons without private prompt contents; conduct weekly reviews that turn specific reports into checked fixes.Advanced. Added a readability reason, preserved editable messages, and began a private issue review with a real report. Recurring weekly review and closure history remain outstanding. A selection, click or Contact visit is not counted as completed learning.

A reported size problem: implementation and recheck

The October 19 report said that the homepage lesson photos and accompanying text were too small. The earlier change record documents the enlargement. The October 20 recheck measured all four photographs at 300 × 200 pixels, with the complete photograph contained, and the main descriptions at approximately 22.3 pixels. Supporting details measured approximately 19.1 pixels. The desktop viewport was 1,363 × 936 CSS pixels.

Status: implemented and desktop-verified. Phone, tablet, zoom and reporter-acceptance checks are still open. The report arrived through direct site feedback, not through the Contact form; this record is not evidence of form-email delivery. No reporter identity is published.

What was verified

  • Native content matched the prepared changes, and both revised scripts passed a syntax check.
  • Empty feedback choices produced field errors and focused the first missing answer.
  • The readability choice and task-progress selection arrived with the correct guide address in an editable message; the final route opened at the Contact feedback section.
  • Changing the issue preserved appended fictional text. An unrelated existing message was kept, with a separate button to add guide details.
  • An external guide-address parameter was rejected. No Contact message was submitted during these checks.
  • Lesson selectors rendered with 21.25-pixel text and 59-pixel height; the button was about 63 pixels high. Contact selectors used 20-pixel text and 51-pixel height. The privacy body rendered at 21.25 pixels.
  • Contact and Privacy descriptions and social fields matched the published content. Their existing ContactPage/WebPage schema types were retained. JSON parsing does not substitute for external schema validation.

Limits: these are AI-assisted desktop checks, not independent human testing, email-delivery confirmation, a legal review, Search Console verification or a complete mobile/accessibility review.

Relevant technical reference: Google’s pagination guidance calls for a separate canonical URL for each page in a sequence. A prepared fix is not counted as a completed change.

Five real ChatGPT trials, including the corrections — October 19, 2026

Five real ChatGPT trials, including the corrections — October 19, 2026

The published checklist, Morgan email, EventDesk screen-help, sharing-data and manual exercises were submitted to ChatGPT. Twelve prompts were completed across five fresh conversations: five initial prompts and seven follow-ups, including one explicit repair.

Test conditions: October 19, 2026, approximately 9:41–9:49 p.m. Eastern time (October 20 UTC). An AI assistant operated ChatGPT’s signed-out website in desktop Chrome at a 1,363 × 936 CSS-pixel viewport. Each initial prompt began in an empty conversation. The exact model/version was not displayed. Ordinary text chat was used, without an account, paid subscription, uploaded file or changed provider settings. Inputs were the published fictional exercises; no personal records were submitted. Follow-ups remained in their original conversations.

Evidence limits: These are five observed text demonstrations, not independent human trials or a reliability study. They do not verify phone/tablet use, screen readers, screenshot recognition, file uploads, provider privacy settings or a real appliance. Earlier illustrative examples remain illustrative. Response wording can vary.

How the record is presented: the input text is exact. The response text was captured from the rendered answer; layout is simplified, and email-panel labels and checklist counters are retained. These transcripts contain fictional practice material.

First checklist

Passed the stated fact checks. Both answers preserved Saturday at 9:30 a.m., the raincoat, phone and ticket, and the undecided return train. No station or booking was invented.

Initial answer observed: (UTC).

Read the actual input, response and follow-ups
Initial input
Turn these fictional notes into a short checklist. Group it into “Before leaving” and “Still to decide.” Keep every detail. Do not invent missing information.

Notes: Pack a raincoat. Charge my phone. Bring my train ticket. My train leaves Saturday at 9:30 a.m. I have not chosen a return train.

Put the departure time above the checklist. Keep choosing a return train under “Still to decide.”
Observed response
Departure: Saturday at 9:30 a.m.

Before leaving
Pack a raincoat.
Charge my phone.
Bring my train ticket.
Still to decide
Choose a return train.
Follow-up 1
Make the checklist easier to scan without changing any facts.

Observed (UTC).

Train Travel Checklist

DEPARTURE

 Saturday · 9:30 a.m.
Before leaving
Pack a raincoat
Charge my phone
Bring my train ticket
Still to decide
Choose a return train

Email

Passed the stated fact and length checks. Both replies preserved Friday and Tuesday, asked whether Tuesday worked, and invented no reason, attachment or additional commitment. The captured text contained 43 words initially and 51 after revision, even including the email-panel labels; both were below 90.

Initial answer observed: (UTC).

Read the actual input, response and follow-ups
Initial input
Write a friendly email reply to Morgan in under 90 words. Say I cannot complete the report by Friday and can send it on Tuesday. Ask whether Tuesday will work. Do not invent a reason for the delay, mention an attachment, or promise any additional work. Use the facts provided and keep the language natural.
Observed response
Email
Subject
Report Timeline

Hi Morgan,

I wanted to let you know that I won’t be able to complete the report by Friday, but I can send it on Tuesday instead. Would Tuesday work for you?

Thanks for your understanding!

Best,
[Your Name]
Follow-up 1
Make the reply a little warmer without adding a reason, an apology I did not request, or any new commitment.

Observed (UTC).

Email
Subject
Report Timeline

Hi Morgan,

I hope you’re doing well! I wanted to let you know that I won’t be able to complete the report by Friday, but I can send it over on Tuesday instead. Would Tuesday work for you?

Thanks so much for your understanding!

Best,
[Your Name]

Screen help

Passed the typed-label checks. The first response identified Export as an inference and waited. After the supplied PNG/JPG/PDF menu, it suggested PDF and asked what appeared next. It did not claim that a file had been created. This tested typed labels, not image recognition.

Initial answer observed: (UTC).

Read the actual input, response and follow-ups
Initial input
This is a fictional practice exercise, not a real app. In an imaginary flyer editor called EventDesk, my flyer is already saved. I want a separate PDF copy to print. The toolbar shows Preview, Export, and Share. I have not clicked anything yet. Using only those visible labels, suggest one next step and explain what I should look for. Say when you are making an inference. Do not invent controls, download a file, publish anything, or assume a step has succeeded. Wait for me to report what appears.
Observed response
Next step: Click Export.

My inference: Since your flyer is already saved and you want a separate PDF copy for printing, Export is the most promising option to investigate. The label alone does not guarantee that it supports PDF.

What to look for: Check whether the next screen offers a PDF format or a print-ready PDF option. Don’t confirm or finalize anything yet.

Tell me what appears after you click Export, and I’ll guide you through the next step based only on what you can see.
Follow-up 1
The menu shows PNG, JPG, and PDF.

Observed (UTC).

Next step: Select PDF.

My inference: PDF is the most suitable choice for a separate copy intended for printing, since it can preserve page dimensions and layout.

What to look for next: A preview, export settings, or a confirmation screen. Don’t assume the PDF has been created yet.

Tell me what appears after you select PDF.

Sharing data

Revised after a budget-unit assumption. The first answer kept the event at two hours, indoors, for 12 adults, but added a dollar sign to the blank budget. The published shortening follow-up removed that assumption and kept all four requested fields blank. Planning suggestions were suggestions, not confirmed event facts. No private guest records were supplied or requested for the chat.

Initial answer observed: (UTC).

Read the actual input, response and follow-ups
Initial input
Create a preparation checklist for a two-hour indoor neighborhood gathering for 12 adults. Group it into before, during, and after the event. Leave the date, location, budget, and assigned helpers as blanks for me to fill in.
Observed response
Neighborhood Gathering Preparation Checklist

For a two-hour indoor gathering with 12 adults. Use this checklist to plan ahead, welcome everyone comfortably, and make cleanup easy.

Event details

Date

Location

Budget

$ ________________________________

Assigned helpers

1. Before the event

Recommended timeline: 1–2 weeks before through 30 minutes before

Planning and invitations
Confirm date, venue, and two-hour schedule
Invite 12 adults and request RSVPs
Ask about dietary restrictions and accessibility needs
Set the budget and decide how costs will be covered
Assign helpers for food, setup, welcoming, and cleanup
Food and supplies
Plan light snacks, finger foods, and drinks for 12 adults
Purchase or prepare food, including dietary-friendly options
Provide water, coffee, tea, or other beverages as appropriate
Gather plates, cups, napkins, utensils, and serving trays
Prepare name tags, markers, and conversation prompts if desired
Indoor space and setup
Clean the gathering area, kitchen, and restroom
Arrange seating for 12 with room to move around
Set up a separate food and beverage station
Check lighting, temperature, and ventilation
Make sure entrances and walkways are clear and accessible
Set out trash and recycling bins
Check restroom supplies, soap, and hand towels
Final checks — 30 minutes before
Pick up or finish preparing refreshments
Chill drinks and set out serving utensils
Put away valuables and secure pets if needed
Set out name tags and any welcome materials
Review helper assignments and the event schedule

0 of 22 tasks completed

Reset
2. During the event

Event duration: 2 hours

Welcome and settle in · First 15 minutes
Greet guests at the entrance
Introduce neighbors who may not know one another
Point out refreshments, restroom, and seating
Offer drinks and help guests feel comfortable
Connect and socialize · Middle 90 minutes
Give a brief welcome and explain the purpose of the gathering
Use a simple introduction round or conversation starter
Encourage everyone to participate without pressure
Refresh snacks, drinks, napkins, and serving areas
Monitor room temperature, comfort, and accessibility
Keep an eye on time and invite quieter guests into conversations
Wrap up · Final 15 minutes
Thank everyone for attending
Share any neighborhood announcements or next steps
Mention plans for future gatherings, if applicable
Help guests collect personal belongings
Invite volunteers to assist with light tidying

0 of 15 tasks completed

Reset
3. After the event

Immediately afterward and within the next 1–2 days

Cleanup and reset
Collect dishes, cups, and serving utensils
Store leftovers safely and discard spoiled food
Wipe tables, kitchen counters, and other used surfaces
Empty trash and recycling bins as needed
Sweep or vacuum floors and clean spills
Clean the restroom and replenish supplies
Return furniture and shared spaces to their original condition
Check for forgotten items and secure the venue
Follow-up and administration
Thank guests and helpers
Return borrowed supplies and equipment
Record actual spending and compare it with the budget
Reimburse helpers or settle shared expenses, if applicable
Share agreed-upon neighborhood updates or contact details
Note what worked well and what to improve next time

0 of 14 tasks completed

Reset
Quick supply list for 12 adults

Use this as a final shopping and packing reminder.

Food and drink

3–4 snack or appetizer options
12–18 drink servings, plus water
Ice, if needed
Dietary-friendly food options

Serving and dining

15–18 plates, cups, and sets of utensils
20–30 napkins
Serving bowls, trays, and utensils
Food labels, if needed

Room and cleanup

12 seats, plus a few spare if available
Trash bags and recycling access
Paper towels, cleaning cloths, and surface cleaner
Hand soap and restroom supplies

Welcome and activities

Name tags and markers
A short welcome or introduction plan
Optional neighborhood information or sign-up sheet

0 of 15 supplies checked

Reset

Planning tip: Keep the gathering simple: comfortable seating, easy-to-eat refreshments, a short welcome, and plenty of time for conversation. For a group of 12, one host and two or three helpers can usually cover the main responsibilities without making the event feel overly formal.
Follow-up 1
Make it shorter without filling any blank or changing the known facts.

Observed (UTC).

Neighborhood Gathering Checklist

Event: Indoor neighborhood gathering
Guests: 12 adults
Duration: 2 hours

Event details

Date: ____________________

Location: ____________________

Budget: ____________________

Assigned helpers: ____________________

1. Before the event
Confirm date, location, and schedule.
Invite guests and confirm attendance.
Check dietary needs and accessibility.
Plan snacks, drinks, and refreshments.
Purchase food and gathering supplies.
Clean the gathering space and restroom.
Arrange seating for 12 adults.
Set up refreshments, trash bins, and recycling.
Assign setup, hosting, and cleanup tasks.
Prepare name tags or conversation starters, if desired.
2. During the event
Welcome guests and introduce neighbors.
Point out refreshments and restroom facilities.
Give a brief welcome and introduce the gathering's purpose.
Encourage friendly, inclusive conversation.
Replenish snacks, drinks, and napkins.
Keep the space comfortable and accessible.
Share neighborhood announcements, if applicable.
Thank guests and prepare to wrap up on time.
3. After the event
Store leftovers safely and discard waste.
Wash dishes and clean surfaces.
Empty trash and recycling.
Clean the restroom and tidy the space.
Return furniture and borrowed items.
Check for forgotten belongings.
Thank guests and helpers.
Record expenses and settle shared costs.
Note improvements for the next gathering.
Quick supply checklist
Snacks and refreshments
Water and other drinks
Plates, cups, utensils, and napkins
Serving trays and utensils
Seating for 12 adults
Trash bags and cleaning supplies
Name tags and markers, if desired

Manual answers

Required a repair to restore omitted precautions. The timer answer matched Timer, page 4. The first damp-cloth answer was incomplete: it said not found, then mentioned dry cloth and no liquid cleaner, but omitted switching off, unplugging and not opening the housing. An explicit source-recheck prompt restored all precautions and answered the damp-cloth question directly. Bulb replacement was correctly marked not found in the supplied material.

Initial answer observed: (UTC).

Read the actual input, response and follow-ups
Initial input
Use only the following fictional practice excerpt to answer: How can I tell that the timer is canceled? Cite the section and page shown. Do not add instructions that the source does not contain. If an answer is missing, say “not found in the provided material.” This is teaching material, not a real product manual.

Example Luma Desk Lamp L1 — practice excerpt.
Page 4: Timer. Press TIMER once to select 30 minutes. Press it a second time to select 60 minutes. Press it a third time to cancel the timer. The timer indicator is off when the timer is canceled.
Page 5: Cleaning. Switch off and unplug the lamp before cleaning. Wipe the outside with a soft, dry cloth. Do not use liquid cleaner. Do not open the lamp housing.
Observed response
The timer indicator is off when the timer is canceled. (Timer section, page 4.)
Follow-up 1
Can I use a damp cloth?

Observed (UTC).

Not found in the provided material. Page 5 says to wipe the outside with a soft, dry cloth and not to use liquid cleaner.
Repair prompt
Recheck the Cleaning section on page 5. Does a damp cloth match the stated soft, dry cloth instruction? Give a direct answer, cite the page, and preserve every cleaning precaution from the supplied excerpt. Do not add new instructions.

Observed (UTC).

No. A damp cloth does not match the stated instruction to use a soft, dry cloth.

Cleaning section, page 5: Switch off and unplug the lamp before cleaning. Wipe the outside with a soft, dry cloth. Do not use liquid cleaner. Do not open the lamp housing.
Follow-up 3
When should I replace the bulb?

Observed (UTC).

Not found in the provided material.

Further checks needed

The earlier terms-approval blocker for #25 is resolved.

Requirements still needed for completion
FeatureProgress and remaining requirement
Brand-new routeAdvanced through a submitted prompt and checked answer in the signed-out desktop route. A beginner must still follow the route independently.
Complete copyingAdvanced: all five exact published exercise texts worked in empty chats. Phone fallback checks remain.
Checkable resultsAdvanced: actual results, decisive checks and a failed answer followed by repair are retained. Full maintained-scope coverage and reader comprehension remain.
Five demonstrationsAdvanced: all five actual inputs, outputs, follow-ups, corrections and conditions are retained. Closure still requires the full items 17–21 standard, including outstanding device/screenshots coverage and another person completing each from a fresh start.

Release verification — October 19, 2026: All five live guides display their recorded-result sections and evidence links. All five public transcripts match the captured input and response text exactly, including both imperfect initial answers. Each transcript disclosure opened with the Enter key. Native content matched all prepared bounded replacements; unrelated Elementor content, article titles, URLs and original publication dates were preserved. The representative result section measured 20-pixel text with 34-pixel line height; all five guides had no horizontal overflow at the tested desktop viewport. Physical-device and independent-reader checks remain outstanding.

Recheck trigger: repeat the relevant trial if its source material, prompt, access route or provider behavior changes, or a reader reports a mismatch. A successful trial is not a promise of the next response.

A real starting screen and a reply kit to reuse — October 19, 2026

A real starting screen and a reply kit to reuse — October 19, 2026

None is newly completed. Deferred items remain 58, 62, 63 and 64.

Start Here now shows an actual signed-out desktop message box, gives one action per starting step and links access claims to current official documentation. The email guide adds a second fictional task, a checked illustrative reply, three kinds of mistake to catch and a complete editable kit. Its prompt builder now keeps unanswered questions separate from confirmed facts.

Scope and remaining checks
FeatureProgress this roundStill needed to complete
Official sourcesAdvanced: OpenAI’s access/history and Free-tier documentation were checked and linked beside the first-tool instructions.Complete significant-claim source mapping throughout the maintained scope.
Brand-new user routeAdvanced: current message-box image, signed-out desktop composer check, access limits and a no-account alternative.Finish the fresh-start route through a submitted prompt and verify that a beginner can follow it without assistance.
Clear steps and screenshotsAdvanced: five single-action starting steps and a real interface image with explanations of the typing and Send controls.Complete difficult-control screenshots and verified device differences for the flagship walkthroughs.
Complete exercise copyingAdvanced: the second reply copies all instructions and fictional facts; automatic and manual copying, keyboard activation and Escape recovery passed.Finish fresh-chat execution and phone fallback checks across the maintained exercises.
Results readers can checkAdvanced: another honestly illustrative reply has five decisive checks that distinguish a requested booking from approval.Complete the remaining example/check coverage and establish reader comprehension.
Five real demonstrationsBlocked at execution: five exact published fictional inputs, follow-ups and acceptance checks are retained in a private test record. The first prompt was placed in a signed-out composer; no message was sent.Approval to accept the service’s first-use terms, then actual outputs, corrections and conditions for all five demonstrations and fresh-start walkthroughs. No product result or independent reader test is claimed.
A second useful taskAdvanced: readers can apply the same checking method to a new room-and-chairs reply using the page or their own notes, without a new account.Verify independently that readers finish the second task through the five-skill path.
Three mistake typesAdvanced: invented explanation, incorrect total and unsupported booking claim each show the source and draft before a separate keyboard-operated explanatory reveal.Run the learner pilot and record whether readers identify the issue and explain the verification step.
Reusable project kitAdvanced: the preview and editable text download include a completed example, second task, blank worksheet, verification checks and reuse note. The downloaded file exactly matches the on-page source.Obtain evidence that a reader actually reuses the kit independently.
Email prompt builderAdvanced: optional unknowns field, preserved open questions, example/reset support and disabled exports after a changed brief until rebuilding or editing.Compare usefulness with the static prompt using a test reader and complete an independent network/data-behavior check.

Verified on the live desktop website: first-task clipboard contents; full second-exercise clipboard and manual-selection contents; five keyboard reveals; Escape closing and focus return; three missing-field errors; unknowns in the built prompt; export controls disabled after changing the brief; updated unknowns after rebuilding; edited-prompt copying and download; reset clearing the new field. The standalone kit download exactly matched its full 3,879-character text. The new screenshot loads with its descriptive alternative. Saved content matched the prepared changes and the email’s title, URL and original publication date were preserved.

Reading and design: the tested 1,363-pixel desktop viewport had no horizontal page overflow. The new exercise text measured 21.25 pixels and its copy control about 63 pixels high. White reading surfaces, navy text, deep-green rules and the existing typography replace the builder’s large rounded background. The instructional screenshot is genuine, not generated. Publisher anonymity is preserved.

Verification limits: the requested mobile accessibility report returned a cached 100/100 result from an earlier check; it is not fresh evidence for these edits. Physical phones/tablets, screen readers, browser zoom, reader comprehension/reuse and independent network inspection were not established. No outside AI response, email delivery, paid service or account creation was tested.

These cover the starting component, its evidence update, the second-task/kit component, nine builder changes, the email evidence update, two copies each of the clarified prompt restriction and worked reply, sentence-case controls and two final reader-facing wording fixes. Private drafts, test preparation, cache clearing and this audit record are excluded.

See the first-tool route · Try the second reply and get the kit · Use the updated email builder.

Private health worksheets and clearer previews — October 19, 2026

Private health worksheets and clearer previews — October 19, 2026

None is newly completed or newly blocked. Deferred items remain 58, 62, 63 and 64.

Four existing guides now offer empty worksheets that readers can copy to paper or a private document without an AI account. The optional prompts ask only for blank layouts: a telehealth follow-up sheet, an appointment note, a caregiver handover log and a pharmacist question sheet. Real health information is added privately, and care instructions remain with the original source or qualified professional.

Review scope requirements and this round’s status
FeatureProgressStill needed to complete
Traceable official sourcesAdvanced: HHS privacy guidance and WHO guidance on AI health-response limitations are linked beside the four revised activities.Connect significant access, privacy and feature claims across the remaining guides to the relevant official documentation.
Entry requirementsAdvanced: four opening overviews and four saved excerpts state result, estimated time, beginner level, device and tool needs; no account, upload or paid feature is required for these worksheets.Finish consistent guide and archive labels, including clear distinctions for local-model installation and advanced automation.
Health and finance support tasksAdvanced: four exercises now request empty layouts instead of inviting personal health notes or care instructions into an AI response. Existing professional-care and crisis guidance is retained.Narrow or remove the remaining consequential claims across the health and finance collection, or obtain appropriate qualified review.
Purposeful imagesAdvanced: the mental-health guide’s generated illustration is replaced with a real photograph of a blank notebook, with a descriptive alternative and photographer credit explaining its purpose.Complete the image-by-image purpose and duplicate-decoration review across the site.
Accurate search and share previewsAdvanced: four search descriptions, two branded SEO titles and four excerpts now match the limited task; the mental-health social image was refreshed to the new photograph.Finish remaining metadata and social-preview coverage, and validate relevant structured data against visible content.

Verified: saved article bodies and excerpts exactly match the prepared changes. All eight worksheet/check links worked through keyboard activation on the public pages. Public titles, descriptions and Open Graph titles/descriptions match the intended values. The new photograph loads with the intended alternative text and credit; the public Open Graph image now uses it. No new form fields or uploads were introduced. All four pages had no horizontal overflow at the tested 1,363-pixel desktop viewport. The mental-health guide’s automated mobile Lighthouse accessibility check scored 100/100 with no issues flagged.

Limits of this check: this was an AI-assisted editorial and implementation review, not a clinical review, product trial, reader study or complete collection review. The five-minute worksheet estimate was not measured with readers. Actual social-platform rendering, formal structured-data validation, screen-reader use, browser zoom and physical-phone checks were not established by this round.

Design and privacy: navy text, deep-green rules, white backgrounds and the existing Montserrat/Noto Sans typography continue the standing design brief. The notebook photograph illustrates preparation, not a patient story or testimonial. Publisher anonymity is preserved.

Clearer lesson choices and larger controls — October 19, 2026

Clearer lesson choices and larger controls — October 19, 2026

None is newly completed. Eleven bounded source edits were made: ten functional changes and one review-note update.

The existing three-question chooser now starts from the Library and Practice search, opens from its direct link, and keeps beginner suggestions relevant to the selected task. Suggestions explain the practice result, estimated time, beginner level, device and access requirements. A selected-answer summary and Change my answers control help readers revise their choice. Each result points to the five-skill learning path.

Review scope requirements and this round’s status
FeatureProgressStill needed to complete
Entry requirementsAdvanced: chooser results now show outcomes, time, level, devices and access; the full email lesson opens at its overview.Complete the remaining guide and archive preview coverage, especially advanced setup requirements.
Learning pathAdvanced: each recommendation explains the skill practised and links to the existing five-skill path.Verify that readers can complete a second useful task without opening another account or learning another platform.
Lesson chooserAdvanced: relevant primary and alternative routes for all 12 answer combinations; direct Library access and answer revision.Reader comparison showing that the chooser makes lesson selection easier than ordinary Library browsing; continue maintaining destination suitability.
Accessible structure and controlsAdvanced: labelled result region, visible focus on the result heading, larger controls, and preserved field-specific validation.Complete screen-reader checks across headings, landmarks, images, controls and error recovery.
Readability and input targetsAdvanced: 20px chooser text and controls, measured 53px lesson links, 56px buttons and 57px selects; square corners and green rules on white.Real enlarged-text, zoom, narrow-device and touch checks across the site, with any failures repaired and recorded.

Verified: all 12 answer combinations produced the intended main and alternative lessons with matching summaries and outcomes. All nine destination anchors exist on the public site. Both full-email routes were rechecked after changing their destination to the entry overview. Library-to-chooser navigation, fragment opening, keyboard submission, field-specific missing-answer messages, preserved answers, reset, stale-result hiding and visible result focus passed. Native saved content exactly matched the intended changes. The Practice page’s mobile Lighthouse accessibility audit scored 100/100. This automated result does not establish full accessibility or replace reader testing.

Design and privacy: navy text, deep-green rules, white backgrounds, Montserrat headings and Noto Sans text preserve the standing brief. No new imagery or decorative emoji was added. The chooser does not save answers or call an AI service. Publisher anonymity is preserved.

Try the updated lesson chooser.

Larger lesson previews and reading checks — October 19, 2026

Larger lesson previews and reading checks — October 19, 2026

Each of the four homepage lesson previews received six changes: photograph sizing and treatment, heading size, description size, access-detail size, action-link size, and semantic grouping with heading-first text order (24). Four shared changes cover the single-lesson row layout, photograph credits, article typography and reading width including inherited legacy/mobile text, and practice-prompt typography.

44 — Image proportions and focal points: advanced
The review scope requires consistent image treatment and desktop/phone checks that preserve faces, hands and relevant activity. The four lesson photographs now occupy 300 × 200-pixel frames at the tested desktop width, use a consistent 3:2 frame with the entire photograph contained, and retain descriptive alternatives and credits. They previously rendered 136 pixels wide. The phone rules place the image above the text. Full hero/subpage consistency and visual phone focal-point checks remain open.
47 — Comfortable reading and copying: advanced
The review scope requires measured shared-template width, body size and line spacing, plus long paragraphs, lists and prompts at normal and enlarged browser settings. The shared template now uses 1.25rem text, 1.75 line spacing and a 65ch maximum measure; older wrappers inherit the reading size. In the tested browser, the 17-pixel root size produced 21.25-pixel article text. The email lesson measured 760 pixels wide, with matching list and prompt sizes. Browser zoom checks remain open.
50 — Accessible structure and alternatives: advanced
The review scope requires meaningful headings, landmarks, image descriptions, labels, errors and media alternatives confirmed with screen-reader checks. Each homepage lesson is now an article with its heading before the access details and description. The main landmark, one page-level heading, meaningful photograph alternatives and visible reading order were inspected. Both tested pages returned 100/100 in the automated mobile accessibility category. This does not establish a complete screen-reader or form-error review.
51 — Contrast, enlarged text and reflow: advanced
The review scope requires measured contrast, usable controls, non-color cues, browser zoom and narrow reflow without lost content. Measured lesson text/link contrast against white ranged from 7.63:1 to 12.57:1; action links remained underlined and about 49 pixels high. The homepage and sampled article routes had no horizontal page overflow at the tested 1,363-pixel viewport. Mobile automation passed on the homepage and email lesson. Manual narrow-screen, text-enlargement and browser-zoom checks remain open.
69 — Draft, reversal notes and release checks: completed
A private draft of the changed homepage was saved and read back before publication. Exact bounded style changes and the four original lesson structures were recorded with reversal instructions before release. Live native readback matched the prepared changes. The homepage, first lesson, contextual Help return, next lesson, email, trip and purchase destinations were checked after publication. The first task remained usable through its manual-copy path. This completes the release process requirement; it does not close the separate device, accessibility or reader-testing recommendations.

Interaction checks: the first lesson’s setup disclosure and the email answer reveal opened by keyboard. Manual selection covered the complete 394-character first prompt and 306-character email prompt. A keyboard copy-and-paste test preserved all first-prompt text; the single-line test field removed only its four newline characters and was cleared without submission. Contextual Help opened the copying answer and returned to the first exercise. The next-step link reached the context lesson.

Limits: automatic copy controls reported success, but this cloud browser’s clipboard readings and paste results were inconsistent, so automatic clipboard behavior is not claimed as fully verified. No external AI service was used and no form, email or personal data was submitted. No real-phone, browser-zoom, independent screen-reader or reader-comprehension test is claimed. Automated accessibility scores do not establish complete accessibility.

Design: navy text, deep green links, white surfaces, Montserrat/Noto Sans typography, open ruled rows and purposeful existing photographs continue the site’s visual direction. No personal publisher identity was added. Recheck after template, stylesheet, image, copy-control or route changes.

Fictional health and money practice — October 19, 2026

Fictional health and money practice — October 19, 2026

Requirements and progress in this round
AuditCompletion requirementThis roundStill needed
SourcesSignificant access, privacy and feature claims must be traceable to relevant official documentation nearby.Advanced. Eight exercises now place checked official source links beside the reason for limiting the task. Source checks cover the statements identified there.Claim-to-source review across the remaining collection; this was not a full re-review of every statement in these articles.
Lesson requirementsBefore opening a guide, readers can distinguish its result, time, level, device and tool needs, including later-stage setup.Advanced. Eight overview blocks and eight native archive/search excerpts identify the result, estimated exercise time, device and optional tool access. These join the 77 previously labeled guides.Requirements for the other 432 archive articles still need review. Exercise times are estimates, not measured reader-test results.
Checkable examplesPrompts have checked results or explicitly illustrative examples, two or three usability checks, and readers can explain the key check.Advanced. Eight complete practice modules include source material or a blank-template request, an explicitly illustrative answer, three checks, a repair step and a request to explain one check.Reader comprehension checks and remaining maintained-set verification. No named AI product was tested in this round.
Health and finance boundariesKeep material within organizing, questions, assumptions and arithmetic; obtain qualified review or narrow/remove consequential recommendations.Advanced. Practice uses seven fictional datasets and one empty worksheet request. The exercises require no personal health records, bank statements or connected accounts, and reject diagnosis, treatment or investment decisions.Continue reviewing consequential claims across the relevant collection; obtain qualified review where narrowing does not resolve a claim.
Accurate previewsTitles, descriptions, headings and share previews reflect the visible task; relevant structured data validates without invented credentials or outcomes.Advanced. Eight SEO titles and eight descriptions now match the revised tasks. Yoast readbacks show matching inherited Open Graph values; a public retirement-page head spot-check confirms the title, description and Open Graph text.Wider metadata coverage, actual social-platform rendering and formal structured-data validation. Valid JSON syntax is not formal schema validation.

What changed

Verification and limits

The native readback of all 517 published articles matches the expected bodies and excerpts. Existing article titles, URLs and publication dates were preserved. The eight public pages show complete practice modules, requirements, official source links and three checks. Five example calculations were checked independently of the displayed AI-style answer. The retirement-page JSON-LD parses as JSON; no formal schema validation is claimed.

The new sections use the existing white, ink and deep-teal identity, open layouts and text that wraps. No decorative imagery, emoji or rounded-card grid was added. No personal identity or real reader records were added. An unpublished draft pilot and exact private change record support review and reversal.

No interactive browser, phone/tablet, zoom, screen-reader, fresh-reader comprehension or external AI-product test was completed in this round. No message or form was sent. These limits are why the

October 18, 2026: clearer archive entry requirements and accurate previews

October 18, 2026: clearer archive entry requirements and accurate previews

Deferred items remain #58, #62, #63 and #64.

Clear requirements before opening a guide

Review scope: consistently show the result, time, level, device and tool needs before a reader opens a guide, and identify later-stage installation or account setup.

Changed: all 77 maintained article excerpts now use the same five labels, drawn from their existing instructions. Times distinguish an on-page exercise from real production or setup where needed: the creative project separates planning from making; the course guide separates planning time from its sample course duration; local AI and Gmail filters identify extra installation or setup time. The local-privacy practice explicitly needs no local installation.

The Latest AI guides archive now explains the labels, allows the full excerpt to display and preserves three readable lines for level/time, result, and device/tools. The existing open list layout, typography and photographs were retained; no new card grid, decorative icon or imagery was added.

Verified: native readback matched all 77 revised guide excerpts. Six archive pages (1, 2, 3, 13, 16 and 43) showed 29 updated excerpts exactly, including the three separate health-preview changes below. The local-installation and Gmail-setup search results displayed their later-stage labels, device and tool requirements. The saved archive layout matched the prepared three component changes.

Still required: the other 440 published articles need their entry requirements reviewed. They have not been assigned invented difficulty levels or task durations.

Search and sharing previews match the task

Review scope: titles, descriptions, headings and sharing previews describe the actual task; relevant structured data validates against visible content without invented ratings, credentials or outcomes.

Changed: the hospital, stroke and sleep-apnea excerpts and search descriptions now reflect their blank-organizer exercises for private completion. Their SEO titles use the AI Over 40 name consistently. The Privacy Notice now has a deliberate title and description instead of a missing search description and a long automatic sharing excerpt.

Verified: all four public page heads contain the intended titles, search descriptions, Open Graph titles and Open Graph descriptions. Each page’s JSON-LD parses as JSON, and publisher/author names in those records remain AI Over 40. These are source-output checks; JSON parsing is not formal structured-data validation or a rendered preview on a social platform.

Still required: the broader semantic preview review, actual sharing-platform rendering checks and formal structured-data validation. These selected metadata edits do not close #54.

Integrity, release record and limitations

All 517 published article bodies and original publication dates matched the pre-change state. All article excerpts matched the intended final values: 80 changed and 437 unchanged. Exact before/after excerpts, SEO values and archive settings were saved in a private draft before the live rollout. The native batch method was first tested on an unpublished draft; the first live excerpt was checked in the public archive before the wider rollout.

One batch timed out after its changes had saved. A fresh read identified 51 completed excerpts and 29 still pending, with no unexpected values; only the pending changes were then applied in smaller batches. No completed excerpt was blindly repeated.

No fresh browser interaction, phone, tablet, zoom or screen-reader pass is claimed. No form or message was submitted, no backup was created and no restore was run. The publication dates and clinical instructions were not rewritten to imply a new review.

October 18, 2026: privacy notice and narrower health exercises

October 18, 2026: privacy notice and narrower health exercises

The deferred items remain #58, #62, #63 and #64.

Scope and remaining checks

#4 — Official sources near changeable claims: advanced
The review scope requires traceable documentation for significant access, privacy and feature claims across the maintained scope. This round checked LM Studio’s system requirements and offline documentation, specified device requirements and connected app activity, and recorded the exact source-check scope. Three health guides now identify the official sources checked and the material exercise update. These selected checks do not establish complete source coverage.
#6 — Accurate site privacy notice: advanced
The review scope requires an inventory of contact, comments, analytics, cookies and email tools; matching handling and retention statements; links at collection points and in the footer; and specialist review where needed. A new Privacy Notice covers confirmed behavior and explicitly identifies unknown retention/provider details. It is assigned as the WordPress privacy page and linked before the Contact form, from the disclaimer and in the footer. A footer link is available on article pages; a notice beside every collection control, a complete provider inventory, confirmed retention and any needed specialist review remain outstanding.
#31 — Keep health and finance material within support tasks: advanced
The review scope requires consequential recommendations to receive appropriate qualified review or be narrowed/removed. The hospital, stroke and sleep-apnea exercises now ask for empty organizers that readers fill privately, instead of supplying health records or clinical instructions for AI transcription. The hospital guide also keeps the original discharge plan authoritative. This is a selected health-content improvement, not a qualified clinical review or a clearance of all health and finance material.
#32 — Evidence-based content-fit decisions: blocked
The inventory and prior Fabulous, golf and advanced-local-AI scope decisions remain recorded in the October 17 content-fit record. Completion still needs usable page-level reader-demand evidence alongside links and strategic fit. No suitable traffic, search-query or demand report was available in this round. No new consolidation or URL change was made.
#69 — Draft, rollback and live-route verification: blocked at interaction checks
The review scope requires reversible release preparation and passing checks on affected routes and the live first-task route. Exact changes and rollback instructions were saved in a private draft before publication, and the Privacy Notice was read back in draft before publishing. Native content and public HTML checks passed. The browser could not be reached, so clicks, keyboard behavior, zoom, mobile reflow and the live first-task interaction were not newly tested.

What changed, and what was verified

  • Using AI After a Hospital Stay: original care-plan authority, blank organizer workflow and a dated source/update note — 3 edits.
  • AI After a Stroke: blank questions page and a dated source/update note — 2 edits. Existing emergency guidance was preserved.
  • Sleep Apnea: blank appointment note and a dated source/update note — 2 edits.
  • Try Local AI on Your Laptop: hardware requirements, offline/connected distinction, material-update note and precise source-check note — 4 edits.
  • Privacy Notice publication, Contact notice link, disclaimer link, disclaimer update date, core privacy-page assignment and footer-menu item — 6 edits.

Read-back checks matched the intended content and original publication dates for all 517 published articles. The four changed guides displayed the new passages in public HTML, with the replaced exercise text absent. The saved Contact layout matched the intended change, and the public notice, Contact disclosure, disclaimer and footer linked to the correct notice. The core privacy-page setting and footer menu item were confirmed. There are now 517 published articles and 21 published pages.

Limits: public HTML is not an interactive browser test. No contact or comment form was submitted; email delivery, deletion handling and consent behavior were not tested. No retention period, product test, independent human review or clinical approval is implied. No backup was created and no restore was run.

The notice uses a plain white reading surface, ink text and the existing typography. No personal publisher name, decorative emoji or generated imagery was added.

October 18, 2026: finished core lesson components and a printable first task

October 18, 2026: finished core lesson components and a printable first task

Earlier records kept 22 and 23 open until the entire legacy archive had been expanded. That added archive-wide closing requirement is corrected here. The five core lessons are the first task, email, screen help, sharing data and manual answers. This record does not claim that every archived article is complete, that the five external AI-product walkthroughs under 25 are finished, or that real-reader and device testing has occurred.

The 16 comprise five concise-introduction/local-help edits under 22; eight finish-panel, duplicate-removal and continuation edits under 23; and three print-resource edits under 24 (one PDF, its download section and its Start Here link). The same edits also support 20 and are not counted again.

Local recovery and first-use terms: completed for the core component scope

Review scope: explain missing controls, unavailable features and poor answers where they occur, define unfamiliar terms at first use and remove repeated introductory material.

Four core introductions are shorter and retain the necessary explanations. Start Here now has three local recovery answers for unavailable or paid chat access, copying difficulties and invented travel details. Its manual/on-paper routes preserve the same fictional task. The other four core lessons retain their task-specific help and existing explanations. Native content and the public first-task HTML were checked; earlier keyboard checks on the existing help panels remain recorded separately. New reader success is not claimed.

A finish and one next lesson: completed for the core component scope

Review scope: a task-specific completion condition, one recommended next lesson or project and a secondary browse link at the lesson ending.

The four article endings now use the established open white, ink and deep-teal component. The privacy and screen checks name the actual fictional exercise; email and manual endings use the same structure. Two repeated completion statements were removed. Start Here now puts its explicit finish check beside its existing next-lesson button and a new practice-browse link. All five core endings and every recommended destination anchor were verified in rendered content.

Readable print resource and task record: completed for the core component scope

Review scope: a useful print resource and one-page record containing the goal, full prompt, essential checks, a checked-result/next-attempt record and guide/version information.

The first-task companion completes print coverage of the five core lessons. Its one-page PDF contains the full travel prompt, a clearly labeled illustrative answer, three checks and two note fields. The public page serves the exact verified PDF bytes. The blank and filled layouts were inspected, both note values survived saving and reopening, and the file identifies its guide and October 18 version. The existing five article companions were also confirmed as one-page PDFs with guide/version information and two record fields. No physical-printer test or new browser Print-dialog test is claimed.

Get the first-task practice sheet · Try the first exercise

Examples and decisive checks: advanced; reader explanation evidence remains

Review scope: pair the prompt with an actual checked result or a clearly labeled illustrative example, show the decisive checks, and establish that readers can explain the key check.

The new printable exercise keeps the full prompt, example and three checks together; the first-task ending asks why an invented return booking must be removed. The material supports a reader check but is not a substitute for observing a reader explain it. Independent comprehension evidence remains missing, so 20 stays open.

Draft changes and recoverable release: advanced; live interaction verification blocked

Review scope: prepare template/feature changes in a draft or staging workflow, retain a usable reversal note, and verify the live first-task route after publication.

After publication, all 517 article bodies matched the expected WordPress hashes; original publication dates were unchanged. The Start Here Elementor tree and practice-sheet content matched the intended result. The homepage entry link, first exercise, finish links and destination anchors were verified through rendered HTML. Browser control timed out, so fresh clicking, keyboard interaction and download-click checks could not be completed. This keeps 69 open.

Verification limits: rendered HTML and native data checks are not a browser-interaction test. No new phone/tablet, zoom, screen-reader, independent-reader or external AI-product test is claimed. No forms or messages were sent; no backup or restore was run. Public attribution and PDF authorship remain AI Over 40. Existing purposeful photography and site typography were preserved.

October 18, 2026: clearer entries, practical help and guide endings

October 18, 2026: clearer entries, practical help and guide endings

48 remain open: 44 active and four deferred. This round advanced 14, 22, 23 and 43; none is newly closed. The selected changes were completed and verified, but the broader collection requirements below remain unfinished.

The 70 comprise 39 entry and presentation edits under 14, 11 terminology and help edits under 22, 18 endings under 23, and two image/caption edits under 43.

Entry requirements

Review scope: show the result, time, level, device and tool needs before opening a guide, and distinguish short practice from later-stage setup.

Changed and verified: all 21 featured guide entries on the seven topic pages now use the same five labels. Three Start Here routes, six project introductions and the five-skill path also state their requirements. Local-AI installation and Gmail-rule previews identify later-stage setup; both warnings were visible in public search results. Times distinguish reading, planning and optional setup rather than promising a completed real-world task.

Design check: a shared presentation update replaces the topic entries’ rounded shells with open white columns, ink text, deep-teal rules and underlined links. Nine decorative icons are hidden. All 21 entries were checked in the public desktop browser: no horizontal overflow, no rounded shell, no visible decorative icon and a minimum measured text/link contrast of 8.01:1 on white. Existing purposeful photographs remain.

Still required: consistent entry requirements across the wider article-archive listings. This round does not label every legacy article.

Local recovery and first-use terms

Review scope: explain missing controls, unavailable features and poor answers where they occur; define unfamiliar terms at first use and remove repeated introductory filler.

Changed and verified: seven first-use explanations clarify prompts, the prompt builder and screenshots across Start Here and five core articles. Four new local help panels cover screenshot-upload/payment limits, manual-upload or omitted-warning problems, unavailable answer-check tools and invented private details. Each opened using Enter with a visible focus outline; each provides an on-page route using fictional material.

Still required: review first-use terminology, recovery guidance and repeated introductory material across the remaining collection. These are website checks, not proof of success with an external AI product or a new learner.

A finish and one next lesson

Review scope: end each guide with a specific completion condition, one recommended next lesson or project and a secondary browse link.

Changed and verified: 18 older “Keep going” endings now place the exercise’s actual finish check beside one primary continuation and one browse link. All 18 rendered panels have that structure, and all continuation targets are published guides. The decluttering-to-home-inventory continuation was followed successfully in the browser. Four other endings initially flagged by a text search already met the requirement and were preserved.

Still required: the wider article archive needs the same end-of-guide consistency. Earlier or mid-lesson completion statements alone have not been counted as finishing this recommendation.

Purposeful imagery

Review scope: retain imagery only when it has a clear explanatory or contextual role; remove decorative or misleading repetition.

Changed and verified: the answer-checking guide now opens without its generic AI-generated featured illustration. Its obsolete caption was removed and its search/social preview data refreshed. The public page has no main-content image or Open Graph image, and its schema author remains “AI Over 40.” The original media file was retained.

Still required: review the remaining featured-image collection for purpose, duplication, provenance and suitability. Removing this one illustration does not complete the site-wide review.

Verification and limits: all 517 article bodies matched the expected WordPress hashes; all original article publication dates stayed unchanged. Both edited hub pages and all eight saved Elementor page trees matched the intended changes. The new entries, help and endings were checked in desktop Chrome. No real phone/tablet, browser-zoom, screen-reader, independent-reader or external AI-product test is claimed. No form submissions, account connections, site backup or restore operation were performed. Public attribution remains the AI Over 40 brand.

Purposeful images, readable controls and accurate previews — 18 October 2026

Purposeful images, readable controls and accurate previews — 18 October 2026

48 remain open: 44 active and four deferred. Newly completed: 45. Advanced and still open: 43 and 54. External structured-data validation under 54 was blocked by a human-verification challenge. Deferred: 58, 62, 63, 64.

Sixteen additional Yoast metadata refreshes cleared obsolete share-preview artwork without changing the descriptions. Record keeping, cache maintenance and verification calls are separate.

advanced

Review scope: Every image should have an explainable role: an activity photo, an instructional screenshot or a preview of a finished result. Remove duplicate decoration and off-brand imagery.

Changed: Removed 37 inline illustrations across 19 articles and 17 decorative header illustrations, including imaginary medical dashboards, glowing devices and abstract graphics. The affected guides now lead with their task and readable text. Original media files remain available; they were not deleted. Two existing featured images were retained.

Verified: All 19 public articles show the intended image changes, one main heading and no horizontal overflow at the checked desktop width. A complete native-content check found no inline image tags or photo backgrounds in the 517 published article bodies. The wider featured-image purpose, provenance and duplicate-decoration review is still required, so 43 remains open.

completed

Review scope: Keep body copy off photographs, give instructions solid backgrounds, distinguish primary buttons from secondary actions, and keep links and labels readable across image crops and interaction states.

Changed: On all seven topic hubs, “Back to Content Library” is now an outlined secondary action beside the filled primary task button. The shared button styles define readable hover and keyboard-focus states. Guided-practice callouts use solid ink backgrounds with white underlined links and a teal accent. The editorial-policy eyebrow uses deeper teal text.

Verified: The default-state check covered all 20 published core pages and 2,800 visible text and control elements, plus the 19 changed article pages. After correcting the practice links and editorial label, the checked text meets the applicable 4.5:1 normal-text or 3:1 large-text contrast threshold against its rendered backgrounds. Body copy is separated from photographs, so its readability does not depend on the photo crop. Category-button hover and visible keyboard focus, contact-field text and placeholders, and expanded header search were checked. No forms were submitted. The broader real-device, zoom and assistive-technology recommendations remain separately open; this is not an accessibility certification.

advanced; external validation blocked

Review scope: Titles, descriptions, headings and share previews should describe the real task, and relevant structured data should validate without invented ratings, credentials or outcomes.

Changed: Six descriptions now explain the privacy-preserving blank worksheets in the chronic-care, health-organizer, lab-question, menopause-diary, blood-pressure-log and medication-review guides. Sixteen stale Yoast preview records were refreshed after header artwork was removed.

Verified: All 517 published articles and 20 published pages have search titles and descriptions. On all six changed guides, the rendered search and Open Graph descriptions match the new wording, the article-schema headline matches the visible main heading, JSON-LD parses, and no rating values appear in the checked graph nodes. The checked public author name remains AI Over 40. All 17 removed featured images are absent from the affected live share previews after refresh. JSON parsing and field comparisons are narrower checks than formal schema validation.

Still required: The broader preview review, paginated archive metadata work and external structured-data validation. Schema.org’s validator redirected this browser to a Google human-verification challenge; no challenge was attempted and no validation pass is claimed.

Integrity and limits: Full content hashes matched the expected result for all 517 published articles, including every unchanged article. Original publication dates are unchanged. The shared stylesheet matched the intended text exactly. These were AI-assisted technical and visual checks, not an independent human review, clinical review, user study or product test. Public publisher identity remains AI Over 40. The ink, white and teal palette, established typography and existing purposeful core-page photography are retained.

Recheck: when adding an image or guide, changing a topic-hub layout, altering interaction styles, or changing a title, description, featured image or schema setting.

Clear disclosures, private worksheets and lesson endings — 18 October 2026

Clear disclosures, private worksheets and lesson endings — 18 October 2026

49 remain open: 45 active and four deferred. Newly completed: 5. Advanced and still open: 8, 22, 23 and 31. The full commercial-relationship check under 8 is blocked by information not available in site content. Deferred: 58, 62, 63 and 64.

These are 28 content edits plus 11 update notes, across 19 articles and two shared site areas.

completed

Review scope: Use a short, truthful AI-assistance disclosure consistently, distinguishing AI contributions from documented human verification.

Changed and checked: Nine previously missing article notices were added. All 517 published article bodies now contain an explicit AI-assistance disclosure in the stored-content inventory. The shared footer explains that source checks are AI-assisted unless a guide records human review, and points to the review methods and dated product-test evidence. Existing detailed evidence notes remain intact.

Scope and remaining work: This closes disclosure coverage and clarification. It does not certify the accuracy of every article, establish clinical review, or turn a website check into an AI-product test.

advanced; relationship confirmation blocked

Review scope: Publish commercial independence rules; disclose relevant sponsorships, commissions and provided products beside recommendations, with clear test and access status.

Changed and checked: The commercial rules now specify naming the company and benefit received, including commissions, sponsorship, free or discounted access and relevant business or personal connections. A direct footer link makes the policy easier to find. The wording links to FTC endorsement guidance.

Scope and remaining work: A scan of 517 article bodies found no recognizable affiliate-link patterns. That cannot establish whether private sponsorship, gifts or other material relationships exist. A complete relationship inventory and confirmation of recommendation-level disclosures are still required; no statement that every product was purchased or independently tested has been invented.

advanced

Review scope: Provide local recovery for missing controls, unavailable features and poor answers; define jargon at first use and remove repetitive introductory filler.

Changed and checked: Seven edits: privacy terms clarified in three health guides; trim, bleed, safe area and physical proof defined in the first card prompt; trigger defined in the automation guide; a local Gmail-filter recovery panel added; and one repetitive introductory sentence removed. The recovery panel opened correctly in the public desktop browser.

Scope and remaining work: These changes cover five guides. The first-use terminology and repeated-filler review is not yet complete across the wider collection.

advanced

Review scope: End each guide with a completion condition, one recommended next lesson or project, and a secondary browse link.

Changed and checked: Four guides—Houseplant Care Log, Small Automation, Career Options and Career Backup Plan—now end with a specific stopping condition, one primary lesson link and a separate topic-browse link. The new sections use a deep-teal rule, white background, dark text and underlined links without rounded card shells.

Scope and remaining work: The four updated endings were checked live. Broader consistency across every guide remains unfinished; existing mid-lesson checks were not counted as newly completed audit recommendations.

advanced

Review scope: Keep health and finance material within support tasks and obtain appropriate qualified review for consequential claims, or narrow or remove those claims.

Changed and checked: Six health prompts now ask for empty worksheets: a chronic-care agenda, personal health organizer, menopause diary, lab-question sheet, blood-pressure log and medication-review checklist. Readers fill real details privately from original records. The prompts supply no diagnosis, medicine, dose, target reading or treatment decision.

Scope and remaining work: This is a practical narrowing of six tasks, not a qualified review of the health or finance collection. A structured screening of 95 relevant guides informed the choice; it was not a full clinical, legal or investment fact-check. Consequential claims elsewhere remain to be reviewed or narrowed.

Verification: the native saved-content fingerprints matched the expected content for all 517 published articles; original publication dates were unchanged. All 19 affected articles were opened in the public desktop browser: the new copy and shared footer were present, and there was no horizontal page overflow at the tested width. The new automation-help panel opened, its primary next-lesson link reached the intended guide, and the footer link reached the commercial policy. The WordPress footer value and commercial policy were read back after saving.

Public crawl limit: a separate automated HTTP crawl received a Mod_Security 406 response and was stopped. It is not counted as a successful full-site crawl. The source inventory and the 19-page browser check are distinct evidence. No new phone, browser-zoom, screen-reader, clinical or independent human test is claimed.

Dates and design: eleven notes describe the actual editorial changes without claiming fresh product tests or whole-article source reviews. Original publication dates remain unchanged. Public publisher identity remains AI Over 40. The new endings retain the site’s deliberate ink, white and deep-teal styling with readable text and no decorative imagery or emojis.

Recheck: when a guide is added, an editorial or commercial arrangement changes, the footer template changes, a linked lesson moves, or a reader reports a mismatch. New article disclosures must describe the actual AI contribution and any documented human review.

Guide choices, completion steps and link crawl — 17 October 2026

Guide choices, completion steps and link crawl — 17 October 2026

50 remain open: 46 active and four deferred. Newly completed: 55. Advanced and still open: 14, 23, 32 and 54. The reader-demand part of 32 is blocked by unavailable page-level traffic evidence.

Each topic layout change covers its three columns.

Weak-fit and overlapping content: advanced; demand check blocked

Review scope: inventory weak-fit or overlapping pages before consolidation; review Fabulous, golf and advanced local AI against reader demand, links and strategic fit; explain each decision and preserve useful URLs and discovery routes.

Checked: an inventory of all 537 published articles and pages, a title-overlap screen with six candidate pairs, the contents of those candidates, and the three named niche guides. This was an inventory and targeted review, not a complete editorial assessment of every article. Current incoming links and the public Tools collection were inspected. Link presence is not evidence of reader demand.

Provisional decisions: keep Fabulous as optional habit-app evaluation, with a reminder or paper routine as a comparison; keep golf technology as an optional reference for an existing interest, not a beginner AI lesson or tested ranking; keep local AI as a later-stage computer setup guide, distinct from the separate explanation of client-data permissions. Each now states its fit and a concrete finish. All three remain available at their existing URLs and in AI Tools & Reviews.

Overlap decisions: retain Garden Refresh and Room Refresh for their different spaces and checks; Society and Movies for their different subjects; Small DIY Project and Small Project for physical planning versus a fictional work plan; Movie and Book recommendations for different catalogs and availability checks; Small AI App and Local Deals App for general task testing versus merchant terms and offer expiry; Small Party and Small Project for event logistics versus a bounded work plan. Similar titles alone did not justify merging these pages.

Still open: page-level traffic, search-query and reader-demand evidence was not available. These are reasoned provisional retention decisions, not evidence that the pages are popular.

Internal links, redirects, downloads and recovery

Review scope: crawl internal links; test redirects, downloads and the 404 page; provide a useful Library or Help recovery route; fix confirmed failures without loops or changed intent.

Crawl: the 537 published articles and pages seeded a public HTTP crawl on 17 October 2026. It followed 34,928 internal link occurrences, including repeated navigation and archive links, to 724 distinct public page URLs. All 724 returned HTTP 200. No discovered internal page target was left unchecked, and the linked section IDs resolved in the retrieved HTML. Admin, feed, preview and action routes were excluded. This is a page-link check, not a claim that every image, script, external website or possible search query was tested.

Redirects and recovery: the older DIY and Home Maintenance addresses each returned one permanent redirect to the intended current guide, followed by HTTP 200; neither looped. The two direct internal-link corrections from the previous check remain in place. An intentionally missing address returned HTTP 404, displayed search and Library/Help navigation, and its Library link opened the correct page using the keyboard.

Downloads: nine files were obtained through their public browser controls: the general practice PDF and Word file, the notebook Word and text files, and all five lesson-companion PDFs. The general PDF has two readable pages; each companion has one readable page and two note fields. Both Word files have valid document packages and the text notebook is nonempty. This verifies those downloads in the desktop browser; it is not a new physical-printing or every-device test.

Result: no confirmed internal page-link or section-link failure remained in this crawl. The missing full-crawl evidence is now present, completing this recommendation together with the earlier link corrections and the verified recovery route. The separate archive canonical issue remains open under 53.

Clear guide-entry requirements

Review scope: consistently show the result, time, level, device and tool needs before a reader opens a guide, and distinguish short chat exercises from later-stage setup work.

Changed: 12 featured entries across AI Basics, Everyday Life, Work & Productivity and AI Tools & Reviews now state those requirements. They distinguish reading a fictional example from optionally using an AI service, and clarify account, installation, payment or connection needs. The weekly-planning entry distinguishes the roughly ten-minute practice from its fictional 90-minute plan; the skills-map duration is an editorial allowance for a first draft.

Design and verification: the four featured sections now use unboxed columns, fine rules, dark blue backgrounds and clear text links. Twelve rounded card shells and their decorative icon tiles were removed. Montserrat headings, Noto Sans text and purposeful existing photography were retained. All labels passed native readback and public HTML checks; the four published desktop layouts were inspected without horizontal page overflow. Mobile stacking settings were retained, but real phones were not newly tested.

Still open: other topic entries, archive listings and setup-heavy guides need the same consistent requirement labels. This is progress on the original item, not completion of the whole site.

A finish and one next step

Review scope: end each guide with a recognizable completion condition, one recommended next lesson or project, and a secondary browse link.

Changed and verified: Fabulous, Golf Technology and Local AI now each have a concrete decision that finishes the task, one primary next guide, and a secondary topic link. Their next steps respectively cover testing a tool before paying, practising a skill with AI, and distinguishing local processing from client-data protection. All three endings were verified in the published HTML; the local-AI ending was also checked in the refreshed browser. Publication dates were preserved. Update notes identify this editorial change without claiming a new product test or external source review.

Still open: the wider guide collection needs the same ending pattern and a consistent single primary next step.

Accurate previews and structured data

Review scope: align titles, descriptions, headings and share previews with the actual task; validate relevant structured data against visible content without invented ratings, credentials or outcomes.

Changed and verified: the Practice-page excerpt now says 28 short lessons instead of 23, matching the current collection. The saved excerpt was read back. The metadata inventory covers all 537 published articles and pages, which have custom SEO titles and descriptions; the existing 28-lesson SEO title was already correct and was retained.

Still open: full semantic checks of previews across the collection, social preview rendering, and structured-data validation against visible content. Metadata presence alone does not complete those requirements.

Font delivery, internal links and archive checks — 17 October 2026
Lesson recovery, next steps and reading view — 17 October 2026

Lesson recovery, next steps and reading view — 17 October 2026

52 remain open: 48 active and four deferred. Newly completed: 21. Advanced and still open: 14, 22, 23 and 24.

Consistent entry requirements

Review scope: show result, time, level, device and tool needs before opening a lesson; distinguish local installation and advanced automation from beginner practice.

Changed: the four Home lessons now each show beginner level, approximate time, device choice and the no-account on-page route beside their existing result descriptions. Four Library tools now show time, level, device and access requirements with task-specific input notes. Local-installation and automation warnings already present in Practice were retained.

Verified: all eight saved labels and their live Home/Library rendering; no horizontal overflow at the inspected desktop width. Still open: consistent coverage of other topic and archive entry points. Phone and tablet interaction was not newly tested.

Realistic mistake and repair: completed for the maintained guide set

Review scope: each maintained guide shows an honestly labelled realistic mistake, a repair that preserves the source or constraint, and when to stop and verify elsewhere.

Coverage: the same maintained set identified before this round: 77 guided articles plus Start Here, 78 lessons in total. Existing examples were reviewed; 26 article sections and the first-task section needed changes. The child-learning exercise now includes a deliberately invented “2 + 6 = 9” error, a concrete correction by counting to eight, and a route to an adult or trusted learning resource. Other changes add task-specific stopping points and the appropriate original record, responsible person, official instructions or actual saved file to check.

Verified: edited text read back exactly; the child example opens on the live page and the first-task repair is visible. Remaining examples already supply a faulty answer or assumption, a correction and a relevant independent check in the exercise or surrounding lesson. These are teaching examples, not recorded tests of an AI product. This closure does not claim every legacy archive article has been revised.

Troubleshooting and first-use jargon

Review scope: concise help for missing controls, unavailable features and poor answers, with jargon explained at first use and repeated introduction filler removed.

Changed: six first-use explanations clarify local models/prompts, automation, transcription, downloadable products and agent permissions. One repeated introductory paragraph was shortened. The repair guidance above also makes recovery more explicit without adding another panel.

Verified: all seven exact text edits read back; the local-AI definition appears on the live page. Existing troubleshooting and prompts were preserved. Still open: the complete first-use terminology and filler review across maintained and broader guide content.

Completion and one next lesson

Review scope: end each guide with a completion condition, one recommended next lesson or project and a secondary browse link.

Changed: eight guides now end with that structure: digital products, social posts, family storybooks, family-history trips, learning projects, practice logs, learning circles and helping a child check AI. Each condition refers to the actual exercise. Older generic next-step blocks were replaced where present.

Verified: all eight sections have exactly one primary next-lesson link and one browse link. Every primary destination is a currently published maintained lesson; the child-to-storybook route was followed in the browser. Still open: broader archive endings and competing continuation routes elsewhere.

Common print view and task record

Review scope: a useful readable print view with essential checks and guide/version information, plus a one-page task record where appropriate.

Changed: a shared reading/print control is included by 35 maintained lessons. It opens answer and evidence details, removes site navigation and copy controls, shows the guide URL and update date, and offers a return to normal view. Escape restores the previous details state. Common print styles use white paper, dark text and readable type and omit navigation, comments, feedback and related-story sections. The existing one-page practice PDFs remain available.

Verified: keyboard entry and Escape exit on the email and local-AI guides; all answer details open; prior open/closed states return; an entered fictional email brief survives the switch; guide URL/date appears; no desktop horizontal overflow. Source read-back confirms the shared script and stylesheet.

Still open or blocked from verification: extend the control beyond those 35 maintained lessons. The available browser did not expose a print dialog, so actual print pagination, physical output and printer behavior were not verified. Opening all checks on screen is not a completed paper-print test. No new phone, tablet or screen-reader test is claimed.

Release checks: all 87 edits matched their intended saved content across 38 fields; original publication dates remained unchanged on all 77 articles. The 34 affected article/page update labels distinguish editorial work from product tests and source reviews. Elementor pages were refreshed through the page-save process and checked on the public site. The anonymous AI Over 40 attribution and white, ink and deep-teal design were retained. No site backup or restore test was run.

Evidence limits: automated/editorial website checks, not independent human verification or hands-on AI-product tests. Earlier external-source check dates remain unchanged. Exact old/new snippets and release results are retained in a private editorial draft.

Guide dates, disclosures and release checks — October 17, 2026

Guide dates, disclosures and release checks — October 17, 2026

Coverage: 77 published articles with maintained lesson panels, plus Start Here: 78 lessons. The 440 other published archive articles are outside this round’s content and metadata verification. AI performed the editing, preliminary checks and browser interactions described here; no independent human review or hands-on AI-product test is claimed.

distinguish publication, change and checking dates

Review scope: consistently identify original publication, substantive updates, source checks and actual testing where applicable, with a material-change note and no cosmetic redating.

All 78 maintained lessons now have a native expandable “Publication, updates and evidence” record. Original publication dates are retained. Each record states the latest recorded material lesson change, separates a source check from a general review, and explicitly identifies missing AI-product testing and human-verification records. Forty lessons, including Start Here, have an earlier dated source or practice-evidence check; 38 do not have a separately recorded source-check date. Missing evidence is labelled, not invented. October 17 is identified as the date-label clarification, not a fresh source review or product test.

Verification: all 77 saved article replacements matched the intended snippets and rendered one review component with the required fields. Their original WordPress publication dates and Article datePublished values were unchanged. Start Here’s rendered component and original October 4 publication date were also checked. Native review details opened and closed by keyboard on the meal-planning and Start Here pages; a dated source record opened correctly in the context lesson. No horizontal overflow appeared in these desktop checks.

consistent, evidence-limited AI disclosures

Review scope: a short, truthful disclosure across the full set, describing AI assistance and what a person actually verified.

All 78 maintained lessons now describe AI drafting, revision and preliminary checks consistently. Their review details state that no separate independent human-verification record is available. This does not imply an endorsement or pretend that a fictional example was a live product result. Still open: disclosure coverage and evidence in the 440 other archive articles, and accurate attribution of any future human verification.

task-specific search and share previews

Review scope: titles, descriptions, headings and share previews match the actual task; structured data matches visible content without invented ratings, credentials or outcomes.

Updated 25 search titles and eight descriptions across 29 articles. Titles now use the consistent AI Over 40 name; selected descriptions explain the fictional exercise and its actual limits. Existing Open Graph inheritance was retained where it already worked. Examples include deciding which AI updates are worth trying and planning a local-AI trial before installation.

Verification: all 29 native SEO save responses matched the requested fields. Across all 77 article records, search and Open Graph text agreed, canonical URLs matched the lesson URLs, Article publication dates remained original, and all 770 internal structured-data references resolved. No rating markup or personal author identity appeared in these graphs. The two example pages were also checked in the live browser: title, description, Open Graph text and canonical address agreed. These are structural and content checks, not a search-engine rich-result guarantee. Still open: the wider archive and complete social-image and structured-data validation coverage.

a retained evidence and walkthrough checklist

Review scope: factual changes have appropriate evidence review; new operational instructions have actual walkthrough records; a “Tested” label has a corresponding test record and consequential claims receive suitable review.

The public publishing checklist now includes draft preparation, a rollback method and post-release learner-route checks. The current update changes editorial labels and metadata; it does not claim newly tested AI instructions. Still open: the appropriate independent human review for consequential claims and actual product walkthroughs before any future Tested claim. This is an editorial gate, not automated enforcement.

Release and route checks

The review-note pilot was first saved as an unpublished draft and its saved and rendered HTML checked. One live pilot was then checked before the wider component rollout. The retained draft contains exact before/after component snippets, the bounded stylesheet addition and previous SEO values; the pilot’s WordPress revisions are also present.

Live route check: Home → first five-minute task → answer-checking Help → return to the checklist → next context lesson → Practice AI. The complete 394-character first-task prompt was selected, and Escape closed the text box and returned focus to its button. No outside AI prompt or contact/comment message was sent. Still open: a full template/feature staging rehearsal, real-device and assistive-technology coverage. Desktop checks are not substituted for those tests.

Design: white backgrounds, dark ink, deep-teal rules and visible keyboard focus. The new evidence panel is a compact disclosure with native details, not a repeated rounded card. Existing typography and purposeful photographs were retained; no decorative emojis or generated imagery were added.

53 remain open: 49 active and four deferred. Newly completed: 3. Advanced and still open: 5, 54, 68 and 69.

October 17, 2026: lesson access, next steps and keyboard journeys

October 17, 2026: lesson access, next steps and keyboard journeys

Three are completed in the core learning scope; two advanced and remain open. The core scope is Start Here and the 28 lessons listed on Practice AI. Broader article-archive labels and guide endings are still outstanding.

consistent labels before opening a lesson

Review scope: show result, time, level, device and tool needs consistently, and distinguish later-stage setup from a short practice exercise. All 28 Practice AI listings now state these needs. Creative practice separates planning from making the card; household practice separates planning time from doing the tasks; the trip estimate matches its lesson. Local installation and account-setting guides now carry later-stage labels, and the client-data guide separates its fictional exercise from software installation. Home, Library and broader archive listings still need the same coverage, so this recommendation remains open.

contextual Help with a route back

Review scope: link confusing lesson steps to the relevant Help answer, preserve a return route and explain links that open another tab. The 28 Practice AI overviews now offer lesson-specific copying Help; one remaining generic creative-lesson link was corrected. Twelve missing lesson-return routes were added. Each of those twelve journeys was followed in desktop Chrome: the copying answer opened and the return link reached the correct exercise anchor. Source checks confirmed return targets for all 28 lessons. Start Here’s answer-checking Help and return worked by keyboard. Every new-tab link in the 28 rendered lessons and Start Here was labeled; links that stay in the same tab retain normal browser Back behavior.

access requirements for core practice

Review scope: give the finished result, time, prerequisites, device, tool or plan and possible cost before the task. All 28 Practice AI lesson overviews now contain those fields. Start Here already supplies them. Each panel distinguishes the free on-page fictional exercise from optional use of an external AI service with its own limits and prices. Lesson-specific notes cover microphones, optional spreadsheet software, craft materials, setup and extra making time. All 28 rendered overviews were checked for the complete field set. The sample meal lesson was visually inspected for readable spacing and hierarchy.

a completion check and one next lesson

Review scope: end each guide with a completion condition, one recommended next lesson or project, and a secondary browse link. Twenty-two unfinished practice endings were revised; six already met the requirement and were retained. All 28 rendered practice lessons now have a completion condition and one primary continuation in their finish component. The five-skill sequence and existing practice-record links were preserved. Internal lesson anchors were checked against rendered destinations. Broader article-archive endings and competing routes outside this core collection still require review, so this recommendation remains open.

the core desktop keyboard route

Review scope: use menus, tabs, copy controls, reveals and forms by keyboard, with visible focus, a logical order, a working skip link and no keyboard trap in the checked route. Home’s skip link reached the main content and then its first action. The first lesson’s disclosures, copying controls, Help answer, return and next lesson were reached and operated without a mouse. A new Select prompt text control selected the complete 394-character first-task prompt; Escape closed it and returned focus. Header navigation, header search and Library search worked by keyboard. The Practice AI chooser identified all three missing answers, focused the first error and produced a suggestion after keyboard correction. Lesson search recovered from no results. The email reveal, selection fallback and prompt builder worked; the contact form tab order reached name, email, message, submit and the following link with visible focus. No contact or comment message was sent.

54 remain open: 50 active and four deferred. Newly completed: 15, 17 and 49. Advanced but still open: 14 and 23.

These were delivered in 66 successful content-save operations.

Evidence limits: these are AI-assisted desktop website checks, not independent reader testing or a test of an AI product. Automatic copy showed its success message, but the browser’s separate virtual clipboard did not expose the copied text; actual clipboard transfer to another application was not verified. The visible selection fallback was verified against the full prompt. No phone, tablet, screen-reader, enlarged-text or narrow-reflow trial was performed in this round; those separate accessibility items remain open. Contact-form navigation was checked without sending a message.

Design and identity: the white, dark-ink and deep-teal system, Montserrat headings, Noto Sans reading text, open ruled sections and existing purposeful photographs are retained. No decorative emojis, generic AI pictures or personal publisher identity were added.

Browse the updated practice lessons · Try the first lesson.

October 16, 2026: shared visual system and reading checks

October 16, 2026: shared visual system and reading checks

Five recommendations were targeted; two are completed and three advanced. The design brief calls for a distinctive, readable identity without default beige surfaces, repetitive rounded panels, decorative emojis or generic AI imagery. Existing purposeful photographs and anonymous AI Over 40 publishing identity are retained.

a documented reusable design system

Review scope: document type, spacing, buttons, links, callouts and navigation patterns, then check consistency across the five main templates. Shared WordPress styles now document and apply white reading surfaces, dark ink, deep teal actions, Montserrat headings, Noto Sans prose, underlined links, a visible focus ring and open ruled navigation. Pale teal callouts distinguish examples and instructions. Home, Start Here, Library, Help and the email flagship were checked in desktop Chrome. Shared selectors make the same component rules reusable.

clear reading surfaces and control hierarchy

Review scope: labels remain readable across relevant templates, image crops and interaction states. Instructions retain solid backgrounds; primary and secondary actions are distinct; ordinary links are underlined. Verification found and corrected a low-contrast PDF link and a secondary-button hover state. The checked copy fallback, example disclosures and Help finder remain usable. Collection-wide crops, phone layouts and the remaining interaction states still need review; this recommendation remains open.

shared article reading width and type

Review scope: long paragraphs, lists and prompts remain comfortable to read and copy at normal and enlarged sizes. The shared article template now uses an 18-pixel body, 1.75 line height and a maximum 70-character measure. The email article measured approximately 721 pixels wide with a 31.5-pixel line height. Prompt copying, manual selection and Escape focus return worked. Enlarged reading was not verified: the available browser zoom shortcut did not change the viewport. This recommendation remains open.

an unobtrusive core learning interface

Review scope: core learning content and controls work without unnecessary motion, autoplay or repeated overlays. Decorative CSS animations, transitions and smooth scrolling are disabled by default, including for readers who prefer reduced motion. All five templates reported no active content animation/transition or autoplay attributes. No intrusive promotional overlays appeared. Home and Start Here disclosures, the Help finder, and the email copy, selection and correction controls worked. Help’s inspected scripted focus movement uses immediate scrolling. This verifies the five-template desktop scope; it is not a real-device or assistive-technology certification.

measured contrast and remaining enlargement checks

Review scope: measure text/control contrast, enlarged text, zoom, narrow reflow, tap spacing and cues beyond color. After the corrections, sampled rendered text on opaque backgrounds passed the 4.5:1 normal-text and 3:1 large-text thresholds; the lowest sampled ratio was 5.00:1. The five templates had no horizontal overflow at 1,363 by 936 pixels. Checked email copy controls were approximately 49 pixels high, the PDF control 53 pixels, and keyboard focus had a visible 3-pixel ring. Links, disclosure arrows and text feedback provide non-color cues. Full control-state coverage, zoom, narrow reflow and real touch spacing remain unverified; this recommendation stays open.

57 remain open: 53 active and four deferred. Newly completed: 41 and 48. Advanced but still open: 45, 47 and 51. Their remaining checks have not been counted as completed.

Evidence limits: these are AI-assisted desktop website checks. Contrast sampling excludes text over images, gradients or translucent backgrounds and is not a complete accessibility audit. No phone, tablet, actual reduced-motion preference emulation, successful zoom test, screen-reader trial or independent reader test was performed. Default motion is disabled regardless of the preference setting. No personal publisher identity was added.

Review again when: templates, shared styles, controls, imagery or navigation change. Remove the bounded October 16 design-system block to reverse these stylesheet changes.

See the updated homepage · See the Content Library.

October 16, 2026: a weekly household plan to save and use

October 16, 2026: a weekly household plan to save and use

The household guide now contains a complete fictional planning project. An AI assistant prepared the example, checked it against the supplied notes and verified the desktop controls. The project prepares a plan; it does not perform the household tasks.

1. A clear project and finish point

The introduction explains what readers can keep, estimates 15 minutes for preparing and checking it, and distinguishes planning from doing the work. The on-page example and downloads need no AI account or uploads. Section links lead to the source notes, worked plan, editable copy and checks.

2. A complete six-note exercise

Fictional notes H1–H6 supply the week, available blocks, agreed owners, task estimates, parcel details, appointment and review. The full prompt preserves uncertainty, separates quote requests from purchases and excludes messages, bookings and calendar changes. Clipboard copying matched the full on-page exercise. Its text download contained all six notes, instructions, checking-note fields and a direct guide link.

3. A worked weekly plan with checked totals

The four agreed blocks total 110 minutes: 75 minutes of tasks and 35 minutes unused. The example preserves each duration and traces tasks to H-number sources. October 19, 20, 21, 22, 23 and 25 were checked against the 2026 calendar. Wednesday’s appointment and unknown transport time remain outside the four-block total; no transport owner or booking was invented. This is a labeled illustrative plan, not a recorded AI-product response or household trial.

4. An editable plan outside the chat

Copy weekly plan, Select plan text and Save plan (.txt) controls accompany the filled example. Clipboard copying and keyboard selection included all 2,707 characters. Escape closed the selection box and returned focus to its control. The 3,306-character plan download was read and checked for the complete example, source-check fields, correct file heading and direct guide anchor. The ordinary exercise download and the existing creative-brief download were also checked after extending the shared controls.

5. Real-world verification and clear entry routes

The plan asks readers to confirm owners, available time, travel, appointment details, current pickup hours and any real quote’s price, scope and expiry. Blank fields record the actual source, date checked, agreement and unresolved questions. The correction reveal now matches the expanded notes. Practice AI finds this project for “weekly household plan,” shows its 15-minute planning estimate and opens the starting section. Projects also links directly to it. The recommended next exercise covers an overloaded schedule.

Independent reader use and real-world household outcomes have not been established.

Checks and limits: Desktop Chrome at a 1,363 by 936 pixel viewport was used for copying, keyboard selection, Escape focus return, downloads, the correction disclosure and entry routes. The new section links resolved, new household IDs were unique and the page had no horizontal overflow at that viewport. The shared script passed a syntax check. No phone, tablet, screen-reader, live AI-product, calendar, booking or independent reader trial was performed.

Review again when: the source notes, dates, plan, export format, shared controls or entry routes change.

See the worked weekly plan · Open the editable-plan section.

October 16, 2026: five photographs matched to their lessons

October 16, 2026: five photographs matched to their lessons

Five guides now use activity-based featured photographs for thumbnails and sharing metadata. An AI assistant inspected the photographs, replaced the previous featured images and added descriptive alternative text and source credits. The photographs are illustrative; they are not screenshots, reader testimonials or evidence of product testing.

1. Creative projects: materials readers can recognise

Colored pencils replace the robot illustration. The photograph points toward making a small card and keeps the image tied to the lesson’s practical activity. Open the creative-project lesson. Photo by Lucas George Wendt on Unsplash.

2. Privacy practice: an everyday laptop setting

A woman using a laptop at home replaces the worried cartoon character. The scene gives the privacy exercise an ordinary, capable setting without displaying readable personal information. Open the privacy practice. Photo by LinkedIn Sales Solutions on Unsplash.

3. Learning a skill: hands at the computer

Hands using a laptop replace the generic group-meeting image. The photograph supports the guide’s individual digital-skills practice; it does not show or verify the spreadsheet example. Open the learning-a-skill lesson. Photo by Christin Hume on Unsplash.

4. Returning to a hobby: a real activity

An acoustic guitar being played replaces the imagined hobby-room illustration. The photograph makes the guide’s practical purpose easier to recognise. Open the old-hobby lesson. Photo by Jacek Dylag on Unsplash.

5. Comparing options: making useful notes

A person writing in a notebook beside a laptop replaces the floating-brain graphic. The image supports organising a comparison without implying that AI has made the decision. Open the comparing-options lesson. Photo by Luke Southern on Unsplash.

Checks and limits: All five featured-image assignments, alternative text and credits were checked. Each guide’s rendered Open Graph image and structured-data image URL point to its new photograph, and the canonical URL is unchanged. The creative-project thumbnail was also checked in the live site search results in desktop Chrome at a 1,363 by 936 pixel viewport. No phone, tablet or external social-platform preview test was performed; third-party services may retain older cached images.

It does not complete the site-wide image inventory, phone crop checks or the full metadata and structured-data review.

Review again when: a lesson’s task, featured image, thumbnail layout or sharing metadata changes.

October 16, 2026: a creative project readers can take away

October 16, 2026: a creative project readers can take away

An AI assistant expanded the existing creative-card lesson into a complete fictional planning exercise. The five improvements below provide a defined task, choices to compare and an editable brief that can be used outside the chat.

1. A clear brief and starting point

The lesson now supplies two fictional personal details, the intended recipient, the card format, available materials and a finish point. Planning and making time are separated: about ten minutes for the brief and twenty more to make and check a card. These are estimates. Readers can follow the examples without an AI account.

2. A complete exercise to copy or save

The prompt contains the full fictional brief, three design directions and explicit limits on invented memories and extra materials. Live clipboard copying and manual selection were checked. The exercise text download contained the complete prompt, checking-note fields and a direct link back to the exercise.

3. Three worked ideas and a reasoned choice

A decorative sprout, two pots and mostly lettering each have a front design, two inside sentences and a practical trade-off. The selected lettering example is explained. The personal statements were compared with the supplied fictional details. Decorative shapes are identified as design choices, not evidence of a real plant or event. These are illustrative teaching examples prepared with AI assistance.

4. An editable project brief

The filled fictional brief includes the chosen design, exact wording, materials, time estimates, source details, asset checks and unfinished review fields. Copy, Select brief text and Save brief (.txt) controls are available. Clipboard and selection checks both included all 1,529 characters. Escape closed the selection box and returned focus to its button.

The downloaded file was opened and checked for the whole brief, its guide anchor and its unfinished physical-review field. It identifies itself as a creative-project brief. An ordinary exercise download and the earlier product-manual reference-card download were also opened and checked after extending the shared controls.

5. Clear completion checks and a route into the project

The lesson distinguishes reopening a completed plan from inspecting a physical card. It asks readers to check wording, fold direction, reading size and any added material before sharing. The correction exercise now matches the expanded source brief. Practice AI describes the planning and making stages, finds the lesson for “creative brief,” and links directly to its starting section. The lesson also links onward to the family-memory project.

The deliverable is a usable text plan outside the chat; its fictional facts are traceable to the supplied brief and it includes no third-party images or downloaded artwork. Added assets require their own source and permission check. This does not establish that a reader has made or tested the physical card.

Checks and limits: Desktop Chrome checks on October 16, 2026 used a 1,363 by 936 pixel viewport. The lesson had no horizontal overflow and its internal section links resolved. Copying, manual selection, keyboard activation, Escape focus return, downloads, the correction disclosure and the Practice AI search route were checked. The shared script passed a syntax check. No real phone, tablet, screen-reader, independent reader or physical-card trial was performed. No new AI-product test is claimed.

Review again when: the fictional brief, worked wording, download format, shared controls or lesson entry points change.

Open the creative-project lesson · Go to the editable brief.

October 16, 2026: clearer guide cards and saving controls

October 16, 2026: clearer guide cards and saving controls

An AI assistant made five refinements to related-guide cards and the shared lesson controls. These changes help readers browse another guide or keep the right practice material.

1. Related-guide heading spacing

The small related-guide headings no longer inherit the main article title’s large line spacing. The manual guide’s three related headings now use 20-pixel text and a 28-pixel line height, with underlined title links. The main article heading retains its own styling. The same smaller heading treatment was also checked on the email guide.

2. Consistent photograph frames

Related-guide photographs use matching three-to-two frames. Each image fits inside its frame without an added crop, so differing source proportions do not move one title above the others. All three manual-guide frames measured about 242 by 161 pixels in the desktop check. Existing image descriptions and destinations were retained.

The theme supplied empty links labelled “Related post link” after the related summaries. Only these empty controls are now hidden. Keyboard checks moved from the first guide title through the next two photograph and title links to the comment field, with no invisible stop between them. Meaningful guide links remain available.

4. Matching visible and accessible download names

The shared save control’s accessible name now begins with the exact visible label, “Save exercise (.txt),” followed by its purpose. Previously it used different wording. The manual and email guides both displayed the matching name. An automated mobile check had flagged a visible-label/name mismatch; this change was checked in the live page structure, not in a screen-reader or voice-control user trial.

5. A clearly identified reference-card template

The manual guide now offers “Select card text” and “Save card (.txt)” beside its existing copy control. The text download identifies itself as a blank reference-card template and links directly back to that template. It retains the model, question, source, edition, section/page, answer, warnings, unresolved details and check-date fields.

Keyboard selection highlighted all 544 characters of the blank card. Escape closed the text box and returned focus to its selection button. A downloaded card file was opened and checked for the complete template and direct guide link. A separate ordinary exercise download still contained the complete supplied exercise and notes section.

Checks and limits: Desktop Chrome checks on October 16, 2026 used a 1,363 by 936 pixel viewport. The sampled manual and email pages had no horizontal overflow. The shared control script passed a syntax check before publication. These are website and generated-file checks; no phone, tablet, screen-reader, voice-control or independent reader trial was performed. No new AI-product test is claimed.

Review again when: the related-post template, thumbnail treatment, shared copy controls or reference-card fields change.

Open the reference-card template.

October 16, 2026: clearer homepage choices

October 16, 2026: clearer homepage choices

An AI assistant refined five parts of the homepage to make the available tasks easier to scan while keeping the site’s existing colours, photographs and useful examples.

1. Clearer lesson choices

The four task cards now share a short explanation of the beginner lessons and the option to read fictional examples without an AI account. The estimated times are about five minutes for the first task, five to ten minutes for email, ten to fifteen minutes for travel and about ten minutes for shopping. The shared access line explains that these exercises need no uploads.

2. Smaller, useful photographs

The four existing task photographs sit beside the text on wider screens, with their proportions and credits retained. Desktop checks confirmed that all four images loaded and that the cards were about 314 pixels tall, compared with 533 pixels before this change.

3. A quieter topic directory

The six topic routes appear as a ruled directory on a pale-blue background, with navy headings and underlined links. Decorative topic icons and the individual card backgrounds were removed. All six destination links were retained; the Everyday Life link was opened using the keyboard.

4. Simpler latest-article previews

The latest-articles area still selects the three newest posts automatically. It now shows linked titles and summaries without large preview photographs or repeated read-guide buttons. Desktop checks confirmed that the three title links remained, the extra images and buttons were absent, and the prompt-notebook title opened its article.

5. Consistent links and keyboard focus

Task and topic links use underlines and a minimum height of 44 pixels. A white inner ring and gold outer outline mark keyboard focus. The first-task and Everyday Life links showed both rings and opened with Enter. The homepage email example also opened and closed with Enter. Button transitions and task-card movement were removed; the four card wrappers reported no animation or transition.

Checks and limits: Desktop browser checks on October 16, 2026 used a 1,363 by 936 pixel viewport. The page had no horizontal overflow, task and topic headings displayed at 24 pixels, and its total height changed from 5,556 to 4,434 pixels with the default example state. This is a layout measurement, not a measured improvement in loading speed or reader success. Narrow-screen layout rules were added, but this round did not verify phones, tablets, screen readers or independent readers. No new AI-product test was performed.

Review again when: lesson durations, homepage content, image dimensions, theme styles or the latest-articles widget change.

View the updated homepage.

October 16, 2026: clearer lesson previews

October 16, 2026: clearer lesson previews

An AI assistant updated the search and sharing text for these five lessons to describe the practice available on each page.

1. First-task preview

Names the five-minute travel-checklist exercise and the option to follow along without an AI account.

Preview title: Your First AI Task: A 5-Minute Practice | AI Over 40
Description: Turn fictional travel notes into a checklist, check every detail, and improve it once. Try the five-minute exercise or follow along without an AI account.

Open the first-task lesson.

2. Email lesson preview

Describes the editable prompt builder, saving the prompt, and checking dates and promises. The article summary uses the same description.

Preview title: Write Clear Emails With AI | AI Over 40
Description: Practise a clear email reply, check dates and promises, then build and save your own editable prompt. Use fictional details; no email plugin is needed.

Open the email lesson.

3. Screen-help preview

Makes the fictional, no-upload practice route clear before a reader opens the guide. The article summary uses the same description.

Preview title: Use AI to Understand a Confusing Screen | AI Over 40
Description: Practise asking for one clear next step with a fictional screen. Check visible labels and what changed. No screenshot, upload or account is needed to practise.

Open the screen-help lesson.

4. Manual lesson preview

Names the source checks, warnings and reference-card result. The article summary uses the same description.

Preview title: Find Answers in Product Manuals With AI | AI Over 40
Description: Find an answer in a fictional manual, check the source and warnings, and keep a reference card. Try the ten-minute practice without an upload or AI account.

Open the manual lesson.

5. Privacy practice preview

Describes choosing the minimum information for the fictional event task. The title identifies this as practice, and the article summary matches the description.

Preview title: What to Share With AI: Privacy Practice | AI Over 40Description: Choose which details a fictional event checklist needs, then build a smaller prompt.

Open the privacy practice lesson.

Checks and limits: On October 16, 2026, desktop browser checks confirmed the page title, description, Open Graph and X fields, and the WebPage name and description in the structured data. Canonical addresses, indexing settings, existing preview images and original publication dates were retained. The four article summaries were also updated. This checks the metadata supplied by this website; it does not verify search rankings or how another platform displays a preview. No new AI-product test or independent reader test was performed.

Review again when: a lesson’s task, access requirements or available practice changes.

October 16, 2026: clearer evidence and review records

October 16, 2026: clearer evidence and review records

An AI assistant updated the first-task, email, screen-help, sharing-information and manual guides. These checks cover the published website and the cited documentation.

1. Identify what each exercise demonstrates

All five guides now use a consistent evidence label with a specific explanation of their fictional material and limits. The labels point to retained records and do not present illustrative answers as observed AI-product outputs.

2. Separate publication, lesson changes and review dates

Each guide now has an expandable Dates, sources and review details panel. Original publication dates were checked against WordPress records and retained. The panels distinguish the latest recorded substantive lesson change, source/example checks, website checks and this editorial-note update. Earlier October 12 notes are identified as earlier changes. All five panels opened using Enter with focus on their native summary control.

3. Explain AI assistance consistently

All five guides now describe AI assistance in drafting, revision and the checks recorded for that lesson. Each links to its own record and states that no independent human review or AI-product demonstration is recorded. No human reviewer or product test was invented.

4. Keep a complete record for the five starting guides

The core guide review register records each guide’s URL, purpose, evidence, source material, tool/account requirements, review conditions, reviewer, limits and next review trigger. The five guide-to-record routes and five returns to the intended exercise anchors were checked. The full prompts still copied correctly: their clipboard text matched the visible source text in the first-task, email, screen-help, privacy and manual exercises. The privacy prompt was checked after opening its answer reveal.

5. Keep the lesson available while reading a source

Nine official-source links across the first-task, screen-help, privacy and manual guides now visibly say they open in a new tab. Their destinations and link attributes were inspected. The existing manual-upload link was updated to OpenAI’s current redirected address. The nine cited pages were reviewed for the specific guidance used by these lessons. A keyboard test of the first lesson’s getting-started source opened the expected source in a separate tab while the lesson and its expanded review note remained available; the source tab was then closed.

Limits: checks used desktop Chrome at a 1,363 × 936 CSS-pixel viewport. No horizontal page overflow was observed on the inspected guides or register. Real phones, tablets, screen readers, zoom/reflow, fresh-chat outputs and independent reader outcomes were not verified in this round. Recheck after changes to a guide, its sources, review links or shared controls.

October 16, 2026: five improvements to the practice tools

October 16, 2026: five improvements to the practice tools

An AI assistant updated the existing tools and checked them in a desktop browser using fictional inputs. The changes make results easier to navigate, keep and return to.

1. Close a text box without losing the result

The time calculator, cost calculator and six-task checklist now have Close text box buttons and Escape support. Both methods hid the text box, reset its expanded state and returned focus to the matching Select button. Calculator results stayed available. The checklist’s six ticks and remembering choice stayed unchanged.

2. Keep a calculator result as a text file

Both calculators now offer a .txt download containing the current inputs, results, assumptions, saved date and guide link. Downloaded files were opened and inspected. In the time example, changing setup from 5 to 6 minutes changed the exported first-attempt total from 22 to 23 minutes. With a fictional $50 ceiling, $20 monthly charge and $120 annual charge, the cost file contained a $30 monthly average, $360 annual total and $140 first-month scenario. Changing an input hides the previous result until recalculation; blank or invalid inputs did not produce an exportable result.

3. Keep the email prompt you actually edited

The email prompt builder now offers Save prompt (.txt). A file downloaded from the fictional Morgan example contained an instruction added directly to the editable prompt, along with a saved date and guide link. Changing the source fields disabled saving until the prompt was rebuilt or edited; an empty prompt also disabled saving. The file is a copy on the reader’s device, not an AI reply, an email submission or a saved website account.

4. Keep unanswered lesson choices visible

The lesson chooser now explains each missing answer beside its field. Submitting three blank choices marked all three and focused the first. After one answer, the two remaining questions stayed marked and the status named them. Completing all three cleared the errors and produced the expected suggestion. Start again cleared the messages and returned focus to the first question.

5. Move between a calculation and its inputs by keyboard

A successful calculation now moves focus into a named results region with a visible outline. Edit these times and Edit these amounts return to the first input. Keyboard activation, the normal fictional results, updated results and invalid-entry focus were checked in both calculators. The changed calculator layouts showed no horizontal page overflow at the inspected desktop size.

Limits: the checks above do not constitute a complete accessibility audit, a screen-reader test, or tests on real phones and tablets. Before these changes, separate automated mobile-profile accessibility scans reported 100/100 for Home, Practice and Help; those scans did not identify the text-box dismissal or chooser-feedback gaps found during manual interaction. The downloads were tested in one desktop browser; device download labels and behavior can differ.

October 15, 2026: finding lessons and getting help

October 15, 2026: finding lessons and getting help

An AI assistant made these five connected updates and checked the live desktop-browser behavior described below.

1. Find five more practice lessons

Practice AI now lists 28 lessons, including family updates, retirement activities, AI announcements, total effort and evidence checks. Searches for caregiver, retirement, news, time saving and agreement each found the intended new lesson. The existing misspelling emial still found the email lesson. Clearing search restored the 28-lesson status and focus to the search box. The live page title and social title both say 28.

2. Get help and return to the same exercise

All five guides now pass their lesson context to Help. Their copying links opened the copying answer, displayed the correct lesson name and returned to the matching exercise anchor. Both mouse and keyboard activation were used during these checks. General Help links also preserve the lesson context.

3. Tell fillable companions from print worksheets

The worksheet answer now distinguishes the five one-page lesson companions, each with two fillable notes fields, from the two-page AI Practice Sheet and Home-project kit, which have no fillable fields. Page counts and form-field counts were checked in all seven PDF files. The answer explains saving, reopening and checking a typed note, and gives a plain-text fallback. This was a file-structure check, not a fresh test of every PDF app or a physical print test.

4. Find and use a saved text exercise

A new saved-text answer explains where to look for a .txt download, what it contains, what it leaves out and how to keep personal notes. Copying help now explains Close text box, Escape, returning focus and editing the pasted copy. Opening the saved-text answer from search preserved the current lesson’s return link.

5. Keep Help search matched to the right answers

Search keywords are now associated with permanent answer identifiers, so inserting an answer no longer shifts keywords onto neighboring questions. All 18 answers remain searchable. Clipboard, fillable, txt, made up, upload, cookies and calculator each returned the intended answer first. Enter moved focus to the first result; opening it expanded the correct answer. A no-match query showed a useful message, and Clear Help search restored the starting status and search-box focus.

Limits: these are checks of the website’s lesson navigation, text, search and PDF structure. They do not establish how an AI product performs. Real-phone, tablet, screen-reader and broader accessibility testing remain outstanding.

October 15, 2026: five complete practice exercises

October 15, 2026: five complete practice exercises

An AI assistant added or expanded these exercises, checked the illustrative answers against their fictional notes, and inspected the live pages in a desktop browser. The examples teach a checking method; they are not outputs from a new product comparison or evidence of real reader outcomes.

Family update: accepted tasks stay separate from completed work

Open the family-update exercise. Added a complete fictional prompt, an example update, unresolved-detail checks and a repair prompt. The notes and answer preserve Tuesday at 6 p.m., Thursday at 10 a.m., Alex’s 9:15 a.m. arrival and the Wednesday-evening update. Groceries remain unassigned and the clinic’s preparation answer remains unknown. Removed two decorative inline illustrations.

Retirement activities: keep the limits visible

Open the activity-planning exercise. Added fictional source notes and a checked trial plan. Tuesday and Thursday mornings remain free, the plan proposes at most one outing and $20 of new spending, and unverified fees, availability and bookings stay unknown. Removed two decorative inline illustrations. This is an activity-planning exercise, not a retirement-affordability assessment.

AI announcements: try, wait or keep what works

Open the announcement exercise. Added three fictional announcements, reasoned decisions and a correction example. The list-grouping feature merits a small checked test; the waitlisted calendar feature needs more facts; the stated $39 photo add-on does not fit the household-list task. Removed two decorative inline illustrations. No real subscription, account connection or feature was tested.

Total effort: count preparation, checking and saving

Open the whole-task exercise. Expanded the existing practice into a complete copyable prompt and added a careful worked conclusion. The fictional stages total 4 + 6 + 12 + 3 = 25 minutes, which is 3 minutes longer than 22. The example does not establish equal result quality or predict other tasks.

AI agreement: preferences are not commitments

Open the evidence-review exercise. Added a fictional book-group plan, a balanced review and a repair prompt. Four Tuesday preferences plus two Saturday preferences and two nonresponses account for all eight people. The answer does not invent acceptance of 6 p.m., a room booking, a fee or an inability to attend Tuesday.

Browser checks: on all five live pages, the manual-copy box selected the complete exercise and its text matched the displayed source. The answer reveals opened with the keyboard and the result links reached the intended anchors. Escape was exercised on all five; focus return was explicitly checked on the retirement and evidence-review pages. A family exercise was copied and pasted successfully, and the whole-task text download was opened and checked for its full exercise, guide link and notes fields. No form was submitted. No horizontal page overflow was observed at the inspected desktop width.

Limits: no real phone, tablet, screen-reader, fresh-chat output, physical print or reader-comprehension test was completed for this update. Existing source-review dates remain distinct from the exercise date.

Related October 14 changes: the shared manual-copy box gained a Close button and Escape handling with focus return. The family-update and retirement featured illustrations were replaced with illustrative photographs and attribution. Those changes were checked during the earlier update; the inline-image removals above were completed on October 15.

October 14, 2026: clearer article browsing and page descriptions

October 14, 2026: clearer article browsing and page descriptions

An AI assistant made thesechanges and checked their live output in a desktop browser. These are interface and technical checks, not a reader study or a complete accessibility audit.

Older-guide navigation

The article archive now explains its publication-date order, distinguishes publication dates from review records, and shows the current page and total. Its page links have descriptive accessible names in a named navigation region. The existing link destinations are retained, with controls measuring at least 48 by 48 CSS pixels in the tested desktop layout.

Keyboard checks followed Previous from page 2 to page 1, Next from page 1 to page 2, and the page-number link to page 44. The first, second and last page counters were correct; the last page had no Next link. Tab focus was visible, and no horizontal overflow was observed in this desktop viewport. Real phones, screen readers and JavaScript-disabled browsing were not tested.

About, Contact, Latest Guides and Disclaimer descriptions

These four pages now have descriptions specific to their purpose and matching titles for search and social sharing. Their structured data identifies About as an AboutPage, Contact as a ContactPage, the article archive as a CollectionPage and Disclaimer as a WebPage. Live titles, descriptions and structured data were inspected. Search engines and sharing services may choose other text or retain a cached preview.

Technical findings still open

The public robots file points to the Yoast sitemap index. All 20 published pages appeared in the page sitemap, and the privacy-notice draft did not. This was a bounded public check, not Search Console verification or a full broken-link crawl.

A mobile-profile Lighthouse lab run on the homepage returned performance 68/100, largest contentful paint 6.5 seconds, cumulative layout shift 0 and total blocking time 10 milliseconds. It had no CrUX field data. Its automated accessibility, best-practices and SEO categories each returned 100/100; those scores do not establish complete accessibility or technical correctness. Unused JavaScript and CSS were flagged for investigation; no speed improvement is claimed from this round.

The archive’s page 2 displays different articles but still emits a canonical URL pointing to page 1. That conflict remains open for a server-side fix. Google’s pagination guidance recommends each page in a sequence have its own canonical URL. These two technical follow-ups have been recorded separately.

October 14, 2026: analytics choice and clearer data explanations

October 14, 2026: analytics choice and clearer data explanations

An AI assistant checked these website changes in desktop Chrome. This is a limited implementation check, not a complete privacy, security or legal-compliance review.

Analytics opt-out and footer access

Help now provides a Google Analytics choice using the existing MonsterInsights opt-out function. The function and current opt-out cookie names were checked in the site’s public page source against the provider’s documentation, checked October 14, 2026. Enter activated the control; its cookie-checking interface reported a saved preference, retained after reload and after visiting Contact. A new Analytics choices footer link reached the open answer from Contact. The control states its browser/device scope, that clearing cookies removes the preference, and that it does not erase past data or block every connection.

Not established: an independent network-traffic audit, prevention of collection before the choice, other services’ behavior, advertising/consent settings in the analytics account, or compatibility with every cookie-blocking extension. This is not an all-cookie consent interface.

Contact guidance without losing an unsent message

The Contact page now places a concise Before you send notice above the form, asks for only the guide link and issue details needed, and points to the data explanations and analytics choice. Help links are labelled as opening in a new tab. Checked: the notice rendered above the single existing form; the Help link opened its answer in a new tab; a fictional unsent message remained in the original form and was then cleared. No message was submitted and delivery was not tested.

Explain the named helpers and progress storage

The new data explanation distinguishes the browser-based email builder and calculators, clipboard/file actions, optional six-task tick storage, outside AI services and Contact submissions. Help search now indexes 17 answers. “Analytics” ranked the opt-out answer first; Enter reached it. “Calculator” found the local-helper explanation. Direct links opened both native disclosure panels. The private privacy-notice draft was updated with the verified implementation facts and outstanding owner checks; it remains unpublished.

Limits: phone/tablet browsers, screen readers, JavaScript-disabled behavior, cookie-denial failures and independent reader trials were not tested. A full data inventory and confirmed provider, retention and deletion practices are still needed before publishing the site’s privacy notice. Recheck these controls after analytics, theme, form or helper-code changes.

October 14, 2026: lesson finding and evidence navigation

October 14, 2026: lesson finding and evidence navigation

These checks were performed by an AI assistant on the published website in desktop Chrome and against WordPress metadata. They describe website behavior, not results from independent readers or outside AI products.

Find a practice lesson by task

The lesson finder searches the existing 23 lesson cards and selected everyday words. Checked: “emial” found the email lesson; “food” found groceries and meals; “spreadsheet” found the small-skill lesson; and “manual” found the manual lesson. A no-match query provided recovery guidance. Enter focused the first result; Clear search emptied the results and returned focus to the input. All 23 cards, the three-question chooser and the progress checklist remained present.

Find the five complete printable exercises

The practice-sheet page now lists the email, screen-help, manual, sharing-information and answer-checking companions. All five links were followed to their intended guide section. Each section exposed its existing, ready PDF download with the correct filename. This update changed navigation only; the PDF files were not regenerated. Their file and note-field checks remain in the October 14 printable-companion record.

Search and open the check records

The record finder searches headings and recorded details, with dated groups that can be opened individually or together. Checked: manual, calculator, keyboard and a no-match query; Enter to the first result; opening a matching record; clearing and input focus; opening and closing all groups; and a direct manual-record link after reload. A keyboard-focus issue on a group heading was corrected, then retested: the summary retained focus and Enter closed it. Existing record IDs and text were preserved.

Clarify what the tool guides demonstrate

The tool-guide hub now explains the difference between explanations, fictional exercises and recorded product tests. Its evidence link opened this record finder by keyboard. Its rendered search title, description, Open Graph and X title/description were aligned with that scope; its structured page type identifies a collection. The revised desktop introduction was visually inspected. No search-ranking or social-platform preview result is claimed.

Remove outdated labels from three rewritten guides

The email guide no longer carries its old automation/plugin tags. The sharing-information guide retains AI privacy concerns and data privacy risks. The answer-checking guide retains AI reliability and misleading AI responses. The corresponding rendered article keywords were checked. Guide addresses, primary topic categories and lesson content were retained; the tag terms themselves were not deleted.

Limits: physical phones, tablets, screen readers, JavaScript-disabled sessions and reader trials were not tested. The full lesson and record content remains in the document as a fallback; that design is not a substitute for a separate no-JavaScript test. Recheck the finders after changing card markup or record headings, and the printable directory after changing download anchors.

October 14, 2026: five printable practice companions
October 14, 2026: finding help, saving exercises, and reader feedback

October 14, 2026: finding help, saving exercises, and reader feedback

These implementation and desktop-browser checks were performed by an AI assistant. They do not establish reader comprehension, task completion, or whole-site accessibility.

Search the 24 AI words

Open the word finder. Added on-page matching of terms, definitions, and selected everyday phrases; result counts; no-match guidance; a Show all words control; and keyboard opening of the first match. The definitions and their source-check date were not changed.

Checked: “made up” found hallucination, “remember” found memory, “scan” found OCR, and “promt” found prompt. A no-match query kept search and recovery available. Enter opened and focused the matching explanation. Show all words restored all 24 terms and focused search. A topic link cleared the filter, and a direct memory link opened its explanation.

Search the 15 Help answers

Open Help search. Added a compact result list with question and keyword matches prioritized, clear/no-match recovery, and links that open the selected answer. Existing lesson context and return links remain available.

Checked: clipboard, cannot upload, too long, made up, and a no-match query. The invented-details answer ranked first for “made up.” Enter moved to the first result; keyboard activation opened and focused the answer. Clearing removed results and restored input focus. The email lesson’s context and return link survived the search.

Save a shared exercise as a text file

Shared exercise-copy controls now include Save exercise (.txt). The file contains the complete selected exercise, the guide title and link, a save date, and blank fields for the reader’s goal, checks, corrections, saved result, and next attempt. It explicitly links back for the guide’s worked examples; it is not a download of the entire article.

Checked: actual downloads from the email guide, both exercises on the manual guide, and the work-instructions guide. Extracted exercise text matched the visible text exactly: 306, 725, 544, and 1,060 characters respectively. Guide links and notes fields were present, including the instructions prompt’s anchor. Existing email clipboard copying and complete manual selection still worked. The new script constructs the file in the browser and adds no upload or storage call.

Scope: this applies to guides using the shared exercise controls. Bespoke builders and other older controls are separate. Download-denied behavior and phone downloads were not tested.

Find the practical tools from the Library

Open the Library tools section. Added four clearly described routes: email prompt builder, whole-task time calculator, subscription-cost calculator, and six-task progress checklist. The tools sit in their own full-width section below the task shortcuts.

Checked: all four links reached their intended controls and anchors. The desktop section rendered once at 1,120 pixels wide with two card columns; the hero/search layout remained separate. The existing “emial” spelling-help suggestion still returned email and inbox guides. A one-column narrow-screen rule is included, but a real phone check remains outstanding.

Prepare feedback about a stuck or completed task

The feedback panel is piloted on five guides: email, confusing screen, product manual, sharing data, and checking an AI answer. Readers choose how far they got and a feedback topic, then review the guide address and selected details in the existing Contact form. The interface states that nothing is sent until they submit that message and asks them to leave out private prompts and records.

Checked: one working panel on each pilot guide; missing-choice errors and first-invalid focus; partway/different-screen and not-started/copying routes; keyboard activation; correct guide, topic, and progress in the Contact draft; changing to completed/helpful; preservation of appended reader text; and the explicit Add guide details action when a message had a different beginning. Test text was cleared afterward.

Not tested: form submission or delivery. No messages were sent and no real reader feedback was collected. There is no completion counter or automated report added by this feature; choosing an answer or opening Contact is not evidence of finishing a task. A future review of submitted feedback and actual reader trials remain separate work.

Shared limits: checks used desktop Chrome. Physical phones, tablets, screen readers, disabled JavaScript, and reader trials were not tested. Recheck these controls after changing their markup, the shared copy block, the Contact form fields, or the linked guide anchors.

October 14, 2026: work exercises and whole-task timing

October 14, 2026: work exercises and whole-task timing

These checks were performed by an AI assistant using the published text, source pages, and desktop Chrome. They are not independent human reviews, reader trials, or product benchmarks.

Work instructions: two routes from four notes

Open the complete instructions exercise. Added four fictional source notes, a complete request and a missing-quantity request, a copyable full prompt, and a worked procedure. The worked routes retain Riley’s review and do not invent purchasing permission, a deadline, or a submission system.

Checked: the 1,060-character prompt copied exactly; manual selection covered the complete prompt; Enter opened the answer; contextual Help opened the invented-details answer and returned to the exercise. The W3C accessible-writing guidance was rechecked for the existing advice about clear headings and instructions.

Not tested: a real workplace process, owner approval, or a colleague following the instructions. Recheck after any source-note or process change.

Service packages: scope, arithmetic, and a longer-job scenario

Open the service-package exercise. Added a complete fictional USD brief, two bounded package descriptions, a comparison table, and an incorrect-profit-claim repair. P2’s time and direct costs include P1; they are not added twice.

Checked: 4 × $40 + $20 = $180, leaving $60 from $240; 6 × $40 + $30 = $270, leaving $90 from $360; 9 × $40 + $30 = $390, leaving a $30 shortfall. All remainders are before other expenses. The 1,600-character full prompt copied exactly and selected in full; Enter opened the correction; Help opened the different-answer guidance and returned correctly. The table was inspected in desktop Chrome. The SBA business-planning resource was rechecked for the existing cost-research advice.

Not tested: actual demand, market rates, customer understanding, a paid job, profitability, or tax treatment. Recheck when scope, hours, prices, or included expenses change.

AI disclosure: describe completed checks precisely

Open the disclosure exercise. Added a five-part fictional work log, full prompt, worked AI-use note, pending-check list, and overclaim repair. The note distinguishes AI drafting, the writer’s changes, the two checked details, unchecked facts and links, an unrecalculated total, and pending publication approval.

Checked: the 1,279-character prompt copied exactly and selected in full; Enter opened the correction; Help opened the invented-details answer and returned to the practice block. The article’s YouTube disclosure guidance, MLA explanation of substantive AI use, and MLA citation guidance were rechecked. Those examples remain specific to their platforms or academic context.

Not tested: acceptance by any editor, employer, instructor, or platform; this is not a universal disclosure form. Recheck the applicable rules and work log before real use.

Whole-task time calculator

Open the time calculator. Added six labeled minute inputs, first-attempt and later-attempt totals, a setup-recovery projection, explicit assumptions, copy/select controls, clear/example buttons, per-field errors, and a static fallback. The calculator adds no account connection, storage, or network request for entries. A later-attempt projection is not presented as measured performance.

Checked in desktop Chrome: Enter submitted the form. The fictional 20-minute baseline with 5 setup, 0 preparation, 8 drafting, 9 checking, and 0 saving produced 22 minutes first and 17 later; two projected tasks totaled 39 versus 40 minutes. A 10-minute baseline with 5 setup and 2 + 8 + 9 + 1 recurring minutes produced 25 first and 20 later, with no setup recovery. Equal repeat time with setup, all-zero inputs, and a decimal case with 1-minute baseline, 0.02 setup and 0.99 repeat time produced the expected comparisons; the decimal case reached equal cumulative time after two tasks.

Blank, negative, over-maximum, and three-decimal inputs showed associated errors, focused the first invalid input, and suppressed results. Editing hid stale results. Clear reset fields and errors; the example restored valid results. The 861-character default summary copied exactly and selected in full. Enter opened the assumptions, and the corrected Help link opened the different-answer guidance and returned to the calculator. The existing FTC subscription guidance was rechecked.

Not tested: actual AI-product performance, a real subscription, a no-JavaScript browser session, or clipboard-denied behavior. Recheck formulas, input limits, copy controls, and Help after any calculator change.

Shared limits: no physical-phone, tablet, screen-reader, or reader-comprehension test was performed in this update. Desktop checks do not establish whole-site accessibility. Excerpts were updated to describe the new exercises and calculator.

October 14, 2026: cost calculator and three practice exercises

October 14, 2026: cost calculator and three practice exercises

Evidence type: fictional examples with AI-assisted editorial, official-source and desktop-browser checks. The subscription calculator performs arithmetic in the page. The other worked examples are illustrative; no fresh-chat AI outputs or reader-trial results were collected.

Subscription-cost calculator

Keep Your Personal AI Tools Within Budget now separates the monthly equivalent, 12-month total, first-month scenario and other monthly payments. With fictional inputs of a $50 ceiling, $20 monthly charges and $180 annual charges, the results are $35 monthly equivalent, $420 yearly total, $200 first-month scenario and $20 in each other month. These are invented amounts, not current vendor prices.

Checked: the example, over-budget and zero-cost cases; decimal rounding ($10.01 monthly plus $100 annually gives $220.12 yearly, $18.34 rounded monthly and $110.01 in the first-month scenario); blank, negative and overly precise input rejection; individual field recovery; stale-result removal; Clear amounts; and keyboard Enter submission. The selected and copied cost summary matched all 868 characters. The live example was visually inspected on desktop.

Scope and source: the calculation assumes unchanged monthly and annual charges in US dollars, with annual charges grouped into the first month for that scenario. It does not retrieve prices, connect a bank, determine actual renewal dates or add unentered taxes and usage fees. The FTC subscription guidance was rechecked for the existing advice about terms, cancellation records and monitoring charges.

Journal transcript rehearsal

Transcribe Old Family Journals With AI now has a complete three-line text rehearsal and a worked transcript note with a separate uncertainty list. Checked that JournalA_0003, both possible years, [illegible], line order and “did not visit” survive, without an invented gender, companion or reason for travel. The full copied exercise matched all 920 displayed characters.

Scope and source: no handwriting image was inspected and no recognition accuracy was measured. The National Archives’ digitization guidance was rechecked for the existing handling advice about book supports, fragile papers and retaining originals.

Greeting-card brief and wording

Create AI-Assisted Greeting Cards Worth Printing now includes a complete fictional brief, sample wording, an illustration brief and unresolved print checks. The front contains 4 words and the inside 25, within the 6-word and 30-word limits. The wording uses the supplied helpfulness detail without promising business success. The full copied exercise matched all 1,118 displayed characters, and the sample was visually inspected on desktop.

Scope and sources: no illustration, print-ready file or physical card was produced or tested. WHCC’s card guidelines were rechecked for the existing trim, fold and guide-layer statements; the U.S. Copyright Office’s AI copyrightability notice was rechecked for the existing human-authorship statement. These sources are now linked beside the corresponding claims. A reader’s actual printer specifications and asset permissions still need checking.

Photo-restoration report review

Check AI Photo Restoration Before Sharing or Selling It now includes a complete fictional brief, four reported changes, worked decisions and a current-draft note. Checked that permitted brightness and dust work still needs visual inspection; the invented brooch, unsupported year and unapproved crop require correction; the original stays identified; and private-review permission is not expanded into public use or sale. The full copied exercise matched all 1,513 displayed characters.

Scope and source: no photograph was uploaded, inspected, edited or shared. The Library of Congress photograph-care guidance was rechecked for the existing careful-handling and clean-work-area advice.

Shared browser checks: desktop Chrome keyboard activation copied all three complete prompts and the calculator summary exactly. Manual selection selected each complete text. Answer disclosures opened by keyboard. All four contextual Help links opened the relevant answer and returned to their exercise or calculator.

Limits and recheck triggers: real-phone, screen-reader and reader-comprehension checks remain outstanding. Recheck after changes to the calculator, prompts, controls, cited guidance or printer requirements, or when a reader reports an unclear step or incorrect calculation.

October 14, 2026: three practical worked examples

October 14, 2026: three practical worked examples

Evidence type: fictional teaching exercises with AI-assisted editorial, source and desktop-browser checks. The worked answers are illustrative. They are not recordings from an AI product or results from reader trials.

Home-maintenance record

Use AI to Organize Home Maintenance now includes four fictional source records, a complete copyable exercise, a source-linked worked record, three questions to confirm and a mistaken schedule to repair. Checked that the last inspection date comes from R2, the W10/W20 model mismatch stays visible, and no future service date or appointment is invented. The full copied prompt matched all 1,206 displayed characters.

Source checked: the FTC home-improvement guidance supports the existing advice about licensing and insurance, multiple written estimates and examining differences between bids. No appliance was inspected, repair diagnosed or real maintenance interval established.

Short-course pilot

Build a Short Course With AI now includes a complete prompt, a three-lesson plan, separate demonstration and practice notes, an answer guide and specific corrective feedback. Checked the proposed 10 + 10 + 10 = 30-minute total, the Maya and Noor commitments, and the unresolved refreshments and projector suggestions. The full copied prompt matched all 1,275 displayed characters.

Source checked: Carnegie Mellon’s learning-objective guidance supports aligning the objective, activity and assessment. No learner trial, learning outcome or actual completion time was measured. Source and permission checks are still required for a reader’s own course materials.

Family-memory storyboard

Make a Family Memory Video With AI Help now includes a complete five-file prompt and a timed storyboard with source IDs and factual captions. Checked that 12 + 18 + 15 + 15 + 15 = 75 seconds, the scene boundaries join, and the watering clip’s unknown usable length stays visible. The table has a caption and column-heading markup. The full copied prompt matched all 1,316 displayed characters.

Source checked: Google Photos’ Android creations guidance supports the existing highlight-video and accompanying-music statements. This was a documentation check; the app was not operated. No family media was uploaded, video exported or sharing permission established.

Browser checks: desktop Chrome keyboard activation copied each full exercise exactly. The manual-selection buttons selected each complete prompt. The answer disclosures opened by keyboard, and each contextual Help link opened the relevant answer and returned to its lesson. The storyboard was visually inspected at desktop width.

Limits and recheck triggers: real-phone, screen-reader and reader-comprehension checks remain outstanding. No fresh-chat AI output was collected. Recheck when these prompts, controls or cited official instructions change, or when a reader reports an unclear step or incorrect source, time or commitment.

October 14, 2026: progress checklist and email builder

October 14, 2026: progress checklist and email builder

Changed: Practice AI now offers a six-task checklist with optional saving in the current browser, a suggested next task, reset and a text download. The email prompt builder now has an explicit Select prompt text button and a specific explanation beside each missing required field.

Checked: AI-assisted checks on the public site in desktop Chrome confirmed that remembering starts off, chosen ticks restore after a reload, reset removes the saved ticks, and the suggested next task changes with completion. The downloaded checklist matched the selected text. In the email builder, only missing fields were flagged; correcting one cleared its error while preserving other outstanding errors. Keyboard selection covered all 750 characters of an edited fictional prompt, and automatic copying matched that edited text. Changing the brief disabled both copying controls until the prompt was rebuilt or edited.

Limits: controlled offline checks covered unavailable storage and invalid saved data; those failure conditions were not forced in the live browser. Phone, screen-reader and real-reader return-use testing remain outstanding. The text download was checked for content; it was not printed on a physical printer. No AI service, email sending or personal source material was used.

Recheck: after changes to the lesson list, browser storage, copying, validation or page rendering. A reader tick is a personal progress note, not proof of learning or a certificate.

October 14, 2026: homepage and first-lesson keyboard checks

October 14, 2026: homepage and first-lesson keyboard checks

Changed: the three homepage examples and both starting-point panels in the first AI lesson now use standard expandable controls with a visible keyboard-focus outline. This replaces controls whose accessibility states conflicted when expanded.

Checked: an AI-assisted review used the public website in desktop Chrome. Tab reached the controls in order; Enter and Space opened the example and setup panels. The homepage and first-lesson skip links led to the main learning actions. The full first-lesson prompt copied correctly by keyboard and matched the text on the page. The answer-checking Help link opened the relevant answer and returned by keyboard to the exercise.

Limits: this was a targeted desktop keyboard and page-markup check, not a complete accessibility assessment. It did not include a phone, zoom/reflow testing, an independent screen-reader user or an AI-product test. The automated homepage audit that identified the issue was not re-run after these changes; no new accessibility score is claimed.

Recheck: after a theme, page-builder, expandable-panel, copy-control or Help-route change. Device, zoom and screen-reader checks remain to be completed.

October 14, 2026: three practice guides

October 14, 2026: three practice guides

These records describe the published lessons checked on this date. They show what the check covered and what should prompt another review.

Reviewer and conditions: AI-assisted website checks for AI Over 40, using desktop Chrome on the public pages. No AI product or paid plan was tested. No independent human review, phone test, screen-reader test or client-data security assessment is claimed.

Plan a Small Project With AI

Purpose and evidence: practise separating work estimates from waiting time and unconfirmed commitments. The source is the complete fictional welcome-guide prompt in the lesson; the worked budget is illustrative teaching material.

Checked: the 20 + 50 + 20-minute estimates total 90 minutes of the reader’s work. The colleague’s 30 minutes stays separate; review timing, approval and the deadline remain unconfirmed. Both copying buttons put their full prompt text on the clipboard. The mistake reveal opens with Enter. Copying Help returns to this exercise, and the next-lesson link opens the weekly-planning guide.

Recheck when: the prompt, estimates, copy controls, Help route or next lesson changes, or a reader reports a mismatch.

Open the project-planning exercise

Does Running AI Locally Keep Client Data Private?

Purpose and evidence: check task extraction without real client information. The answer-checking sources are the invented handover paragraphs P1–P4. This record covers the practice exercise, not the privacy or security of a local application.

Checked: the worked answer preserves the missing calendar date, supplier quote, task owners, approval and sharing permission. Jordan remains proposed rather than confirmed. The copy button copies all instructions and four paragraphs. The mistake reveal opens with Enter. The personal-information Help answer returns to the exercise, and the next link opens the client-permissions guide.

Recheck when: P1–P4, the worked answer, copy controls or linked Help changes. Recheck the article’s product-specific background separately if the linked provider documentation or application features change.

Open the fictional handover exercise

Build a Small AI Reference Library

Purpose and evidence: trace an answer to its documents and preserve conflicts or gaps. The sources are the fictional venue handbook A, draft event plan B and equipment list C, included in the copyable exercise.

Checked: the worked answer retains the 09:00/09:30 conflict, the unconfirmed equipment-check owner and the remote-control information gap. Copying includes all three sources. The manual-copy control selects the entire exercise in a read-only text box. The mistake reveal opens with Enter. Invented-details Help returns to the exercise, and the next link opens the AI-search guide.

Recheck when: a source, worked answer, copy control or linked Help changes. Review the separate notebook setup instructions when the provider changes its interface or source-import behavior.

Open the three-source exercise

Found a problem? Tell us the guide URL and the step that differs. Include your device and browser if relevant; leave out passwords and private records.

Core guide review register

Core guide review register

These records cover our five starting guides. Website and source checks were recorded on October 16, 2026; actual signed-out ChatGPT text demonstrations were added on October 19, 2026. The linked transcripts keep observed responses separate from the illustrative worked examples.

Publisher and reviewer
AI Over 40 is the publisher. An AI assistant performed the source, example, website and ChatGPT text checks listed here. No independent human review is recorded.
Conditions shared by these records
Desktop Chrome at a 1,363 × 936 CSS-pixel viewport, using supplied fictional practice material. The October 16 review covered the website. On October 19, each of the five exact exercise texts was submitted in an empty signed-out ChatGPT conversation; follow-ups stayed in that conversation. No account, paid subscription, upload or provider-setting change was used; the model/version was not displayed. Read the actual test conditions, outputs and corrections.
Verification still outstanding
Real-phone and tablet use, remaining difficult-control screenshots, screen-reader checks and independent reader trials. Five AI-operated text demonstrations are now recorded; they do not establish general accuracy, privacy, reliability or independent learner success.

First useful AI task

Open this practice lesson

Purpose and evidence
Turn fictional notes into a checklist, preserve the facts and improve it once. The written exercise’s sample answer is illustrative. The separate five-screen walkthrough contains actual ChatGPT outputs from October 21, 2026.
Tool, plan and input
The no-account on-page practice remains available. An AI assistant ran the exact fictional prompt and follow-up in signed-out desktop ChatGPT on October 21. The model/version was not displayed. Current input, output, screenshots and check record; earlier October 19 record.
Publication and lesson change
Page published October 4, 2026. Latest recorded substantive lesson change: October 21, 2026 — added five actual captures, keyboard navigation, larger instructional images a complete written walkthrough and an 84-second captioned video.
Sources reviewed October 21, 2026
Checks and limits
Both observed answers kept Saturday at 9:30 a.m., all three packing tasks and the unresolved return train. The full 394-character prompt was manually copied and pasted. Five image loads, desktop keyboard navigation, visible captions and the text alternative were checked. Independent reader, phone/touch and screen-reader checks remain outstanding; the MP4 is published, and desktop playback, pause, end seeking and keyboard transcript access were verified.
Next review trigger
Recheck when OpenAI changes the pictured controls, access or setup guidance; when this lesson’s prompt, copying or navigation changes; or when a reader reports an unclear step or mismatch.

Write a clear email

Open this practice lesson

Purpose and evidence
Write and check a reply that preserves the supplied deadline and makes no invented commitments. The Morgan brief, sample reply and mistake exercise are fictional. They teach preserving facts and commitments; the sample is not a recorded answer from an AI product.
Tool, plan and input
The no-account on-page practice remains available. On October 19, 2026, an AI assistant also ran the exact fictional text prompt and follow-ups in signed-out desktop ChatGPT. Read this actual input, output and check record. The earlier worked example remains illustrative.
Publication and lesson change
Page published June 28, 2025. Latest recorded substantive lesson change: October 26, 2026 — clarified builder steps and recovery, added a changed-details review, and labeled the saved-file timestamp. The actual text demonstration remains dated October 19.
Sources reviewed October 16, 2026
The fictional Morgan brief and worked replies in the guide. No external provider feature claim is part of this exercise review.
Checks and limits
The Morgan example preserves the Friday deadline, offers Tuesday and asks whether Tuesday works. Builder and text-download checks are recorded separately. Desktop review-note, copying and navigation checks: October 16, 2026. Conditions and outstanding verification are listed above.
Next review trigger
Recheck if the fictional brief, sample reply, builder fields, copying, saving or lesson links change. Also revisit the record after a reader reports an unclear step or mismatch.

Get help with a confusing screen

Open this practice lesson

Purpose and evidence
Ask for one screen-help step and stop when the suggested control does not match the visible screen. EventDesk is an invented screen rehearsal. The exercise teaches checking visible labels and stopping at a mismatch; no screenshot-recognition product test is recorded.
Tool, plan and input
The no-account on-page practice remains available. On October 19, 2026, an AI assistant also ran the exact fictional text prompt and follow-ups in signed-out desktop ChatGPT. Read this actual input, output and check record. The earlier worked example remains illustrative.
Publication and lesson change
Page published October 6, 2026. Latest recorded substantive lesson change: October 19, 2026 — added the actual text demonstration, findings and limitations.
Sources reviewed October 16, 2026
Checks and limits
The EventDesk rehearsal uses the displayed choices and distinguishes selecting a PDF option from actually creating a file. Desktop review-note, copying and navigation checks: October 16, 2026. Conditions and outstanding verification are listed above.
Next review trigger
Recheck if the cited screenshot or image-input instructions change, or the rehearsal labels, answers or Help routes change. Also revisit the record after a reader reports an unclear step or mismatch.

Choose what to share with AI

Open this practice lesson

Purpose and evidence
Choose the minimum information needed for a task and identify details that should stay out. The gathering example is fictional practice in sharing less information. It does not test provider privacy settings or approve a service for sensitive records.
Tool, plan and input
The no-account on-page practice remains available. On October 19, 2026, an AI assistant also ran the exact fictional text prompt and follow-ups in signed-out desktop ChatGPT. Read this actual input, output and check record. The earlier worked example remains illustrative.
Publication and lesson change
Page published July 25, 2024. Latest recorded substantive lesson change: October 19, 2026 — added the actual text demonstration, findings and limitations.
Sources reviewed October 16, 2026
Checks and limits
The fictional gathering exercise distinguishes task-relevant information from unnecessary private details. The source review covers the two linked general risk statements. Desktop review-note, copying and navigation checks: October 16, 2026. Conditions and outstanding verification are listed above.
Next review trigger
Recheck if NCSC or NIST revises the cited guidance, or the exercise, explanation, permissions wording or Help routes change. Also revisit the record after a reader reports an unclear step or mismatch.

Find an answer in a manual

Open this practice lesson

Purpose and evidence
Find an answer in a supplied manual passage and keep its conditions, warnings and unresolved details. The Luma manual excerpt and answers are fictional. They teach source checking; no real lamp or AI file-upload service was tested.
Tool, plan and input
The no-account on-page practice remains available. On October 19, 2026, an AI assistant also ran the exact fictional text prompt and follow-ups in signed-out desktop ChatGPT. Read this actual input, output and check record. The earlier worked example remains illustrative.
Publication and lesson change
Page published October 8, 2026. Latest recorded substantive lesson change: October 19, 2026 — added the actual text demonstration, findings and limitations.
Sources reviewed October 16, 2026
Checks and limits
The Luma exercise requires the supplied source, retains conditions and warnings, and marks missing information. No real appliance instructions were verified. Desktop review-note, copying and navigation checks: October 16, 2026. Conditions and outstanding verification are listed above.
Next review trigger
Recheck if OpenAI changes file/image handling, or the fictional excerpt, answer, reference card, copy controls or Help routes change. Also revisit the record after a reader reports an unclear step or mismatch.

Find earlier checks and their limitations

Finding and keeping a useful result — October 12, 2026

Finding and keeping a useful result

October 12, 2026 · AI Over 40
An AI-assisted check of the public desktop website covered these changes:

  • Library searches for “reply to email” and “understand a manual” each put the relevant lesson first.
  • The revised notebook exercise displays the full instructions and fictional notes together. Keyboard activation of its copy control shows a completion message.
  • The manual reference card preserves the fictional source’s model, page reference and cleaning conditions.
  • The main navigation includes Help, with direct lesson links for readers who cannot find a guide through search.

Limits: this checks published lesson content and desktop page interactions. It does not establish mobile usability, screen-reader support, clipboard behavior on every device, or an AI product’s accuracy. The worked examples remain labeled teaching material. No independent human review is claimed.