Library › Research AI › Customer support teams
Check a saved reply macro against the current refund policy
Practice: Catching mistakes and made-up facts. About 5 minutes.
About this exercise
Your team's "Refund request" macro was written last year, and the refund policy changed in October. You want one numbered line per claim in the macro, under 60 words total, each labeled Matches, Outdated, or Not in policy, with the current rule after any Outdated label. Use only the October 2026 policy, plain text, nothing else.
Common mistakes when you ask AI for this
- Not naming the current policy. The AI may assume the macro is right because it sounds official. Try: Say: "Check against the Oct 2026 policy only."
- Calling extras wrong. The coupon isn't in the policy, so it's unverified rather than outdated. Try: Add: "If the policy doesn't mention it, label it Not in policy."
- Getting a new macro. The AI rewrites the macro, and your lead can't see what changed. Try: Say: "Label each claim only, no rewrite."
- Trusting the AI to count words. AI is bad at counting. Ask for a total word limit and it often lands 10-20% over. Try: Cap each line: "each line under 10 words."
- Getting markdown back. AI defaults to bold labels and headings, which show up as stray symbols in the macro editor. Try: Ask for "plain text, no bold, no headings."
What a strong prompt includes
- Gives exact labels
- Names the current policy
- Asks for the current rule
- Sets a length limit
- Asks for plain text
Related exercises
- Alert sales to a competitor's price change, facts only (Research AI · Product managers)
- Answer a buyer's HOA questions from the bylaws (Research AI · Real estate agents)
- Answer a coworker's travel policy question in one Slack reply (Research AI · Office and admin)
- Answer a family's visiting questions from the current policy (Research AI · Nurses and healthcare workers)
- Answer winter break housing questions from the current policy (Research AI · College students)