Teach the actual steps with worked examples

Prepare a representative task

Use a realistic but non-confidential input. State the goal and constraints before opening the tool. Explain why you chose the approach and what an acceptable result should contain.

Show the review process

Include an imperfect output and how you detect the problem. For an AI-generated answer, compare claims with sources or a task-specific rubric. Explain which corrections require human judgment.

Let learners adapt it

Give learners a variation that changes one constraint. Ask them to explain what they kept and what they changed. Share a short checklist they can use after the session. Avoid hiding essential decisions behind a polished demo that only works with the exact input you prepared.