A participant should leave an AI workshop able to do something useful and judge whether the result is good enough. Watching an impressive demonstration is different from learning a repeatable skill.
If you want to teach paid workshops, begin with a task you understand well and an audience whose needs you have checked. Your professional or community experience may help you choose relevant examples, but enthusiasm about AI alone is not evidence that you can teach it responsibly.
Promise one observable outcome
Choose a specific result: participants will turn a set of fictional meeting notes into a draft action list and check it against the notes. That is a clearer promise than “master AI in an hour.” Explain what prior knowledge, device, account, and software access the session requires.
Ask the host about participants’ existing experience, accessibility needs, organizational AI policy, and the tools they are permitted to use. Do not infer ability from age. Some attendees will be experienced specialists using AI for the first time; others will already have useful techniques to share.
Be accurate about your credentials. A short vendor course can help you learn a product, but completing it does not automatically make you an accredited educator or certify your workshop. State the training and experience you actually have.
Design practice before slides
Build a small exercise with a known, checkable result. Use fictional or public material you have permission to share. Prepare a source sheet, a starter prompt, and a checklist participants can use without your help.
For the meeting-notes exercise, include one agreed action, one rejected suggestion, and one missing deadline. The expected result should preserve those differences. A good lesson includes a situation where AI should ask a question or admit uncertainty.
A possible 60-minute session is:
- 10 minutes: confirm the task, tool access, and information boundaries.
- 10 minutes: demonstrate one attempt and show how you check it.
- 20 minutes: participants try a related exercise themselves.
- 10 minutes: compare the output with the source and correct mistakes.
- 10 minutes: discuss appropriate use, remaining questions, and a next practice task.
This is an illustrative lesson structure, not a tested formula for learning outcomes. Adjust it after a small pilot and leave room for the time your audience actually needs.
Teach judgment alongside prompting
UNESCO’s AI competency framework for teachers includes human agency, ethics, AI foundations, and teaching practice. It is a useful reminder that education involves more than operating a tool. It does not endorse your course or establish a commercial credential.
Ask participants to identify where the model used source facts, where it inferred something, and where a person must decide. Show an error you have prepared transparently as a teaching example rather than pretending every live output will be correct.
Use questions such as: “What evidence supports this sentence?” and “What would happen if we acted on it without checking?” Keep the first exercise away from medical decisions, legal documents, financial recommendations, and sensitive workplace information.
Prepare for access and failure
Test the exercise on the actual tool and account tier before the session. Check current official instructions and limits instead of relying on screenshots from an old course. Tell participants beforehand if a paid plan is required, and avoid unexpected purchases during class.
Provide a way to participate without sharing passwords or private screens. Use readable text, explain unfamiliar terms, and offer written instructions as well as a spoken demonstration. For an online session, plan captions or other agreed accessibility arrangements.
Have an offline version using prepared outputs in case the service is unavailable. Label those outputs as prepared examples. Participants can still practice spotting unsupported claims and revising a prompt without a live connection.
Do not record participants, publish their questions, or upload their exercise work to an AI tool without the appropriate agreement. For workplace or school delivery, follow the host’s privacy, safeguarding, procurement, and accessibility requirements.
Price the complete job
Count preparation, research, rehearsal, delivery, travel or platform costs, materials, administration, and the follow-up you promise. A one-hour event can involve considerably more than one hour of work. Use actual quotes for paid services and keep the fee understandable.
Put the scope in writing: audience size, learning outcome, required equipment, deliverables, cancellation terms, and revision or rescheduling arrangements. Define whether post-session questions are included and for how long. Do not promise ongoing personal support if you cannot provide it.
Run a small pilot before promoting a larger program. Ask participants to complete a fresh example independently, explain their checking process, and identify a situation in which they would avoid using AI. That gives you more useful feedback than satisfaction ratings alone.
Leave participants with a next step
Provide the exercise, checking checklist, current source links, and one manageable practice suggestion. Invite feedback about confusing instructions or missing access arrangements. Revise the material when the tool or the audience changes.
For the learner’s perspective, see choosing an AI course that fits a goal. For a workplace exercise, see testing one AI workflow at work. A workshop earns its place through useful learning and honest delivery, not a prediction of the instructor’s income.
AI disclosure: This article was written and revised with AI assistance. Sources checked September 26, 2026. See our disclaimer.
Illustrative photograph by mouad bouallayel on Unsplash.
