Before you start
Any chat assistant. Copy the fictional example below; no uploads, coding or paid features required.
Why this lesson exists
Across Patrick’s experiments, a reliable pattern repeats: give the tool a clear task, useful context, constraints and a way to check its work. This is the smallest version of that loop.
Do the exercise
Paste a small source
Use: “Maya drafts the invitation. Lee checks the venue details by Thursday. We have not chosen the event date.” Read it once yourself before asking the model anything.
Ask for an observable result
Request a table of action, owner and deadline using only the note. Explicitly require unknown where a field is missing. This is easier to check than an open-ended summary.
Check each cell
Maya owns the draft but has no deadline in the note. Lee checks venue details by Thursday. The event date is undecided. Compare the answer with these facts.
Make one targeted correction
If the model assigns Thursday to Maya, point to the exact mistake and ask it to revise only that row. Save the prompt and your checked answer as a reusable example.
A prompt to adapt
Replace the bracketed parts with your own practice details.
Turn this note into a table with action, owner and deadline. Use only facts in the note. Write “unknown” when a deadline or owner is missing. After the table list open questions. Note: Maya drafts the invitation. Lee checks the venue details by Thursday. We have not chosen the event date.
What this looks like
Expected result: invitation draft — Maya — unknown; venue-detail check — Lee — Thursday. Open question: what is the event date? Do not add a calendar date for “Thursday” without more context.
Check your result
Use evidence from your output. A confident explanation from the AI is not enough.
- No owner or deadline is invented.
- You can point to the source of every cell.
- You can explain one correction in plain language.
If it isn’t working
If the table includes invented details, shorten the source and repeat the use-only-the-note rule. Always read the final answer; a better prompt reduces mistakes but does not eliminate them.
Optional. Progress stays in this browser.
Where this came from
An original foundation lesson developed for this learning hub, informed by the site’s build and publishing workflow. Sources and approach.
Prepared September 2026. Tools and interfaces change; use current official setup instructions. Session lengths are estimates.