Library › Research AI › Real estate agents

Catch a chatbot's school zone mistakes before telling your buyer

Practice: Catching mistakes and made-up facts. About 5 minutes.

About this exercise

Your buyers asked which schools serve a listing, and a chatbot answered. You want to check it against the district's boundary lookup before replying. You want only the errors, as numbered lines in the form "Chatbot says X; district says Y", under 50 words total, plain text. Use only the district lookup, skip lines that match, and don't rewrite the answer.

Common mistakes when you ask AI for this

  • Trusting old boundaries. Chatbots often know older zoning, and won't notice a redraw unless you check the lookup. Try: Say: "The 2026-27 district lookup is the only source of truth."
  • Flagging extras as errors. The rating and busing claims aren't in the lookup, so they're unverified, not contradicted. Try: Add: "Only list what the lookup contradicts."
  • Drafting the buyer reply. The AI writes a message to your clients instead of a list of errors. Try: Say: "Output only the error lines."
  • 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 14 words."
  • Getting markdown back. AI defaults to bold labels and headings, which show up as stray symbols in your client notes. Try: Ask for "plain text, no bold, no headings."

What a strong prompt includes

  • Names the source of truth
  • Asks for contradictions only
  • Gives the line format
  • Sets a length limit
  • Asks for plain text

Practice this now ↑

Related exercises