Library › Data AI › Real estate agents

Price a listing from comps after tossing a bad one

Practice: Keeping the AI to your sources. About 5 minutes.

About this exercise

You're preparing a suggested list price for a 1,900 sq ft home from 4 comps. Write exactly 4 numbered lines, one per comp, each as Address: $price ÷ sq ft = $/sq ft, marking any comp you exclude and why. Then one unnumbered last line averaging the valid comps and multiplying by 1,900. You pulled the county record: 7 Birch Ln is 1,820 permitted sq ft (the MLS figure includes an unpermitted garage conversion). Apply the notes too, plain text, under 100 words.

Common mistakes when you ask AI for this

  • Trusting the MLS square footage. The AI divides Birch's price by 2,100, which makes it look cheaper per foot. Try: Add: "Use 1,820 permitted sq ft for 7 Birch Ln."
  • Keeping the foreclosure. The AI averages in the bank sale and drags the price down. Try: Add: "Exclude 48 Elm Ct (foreclosure) and say so on its line."
  • Not showing the final math. A bare price is hard to defend to sellers who want $500k. Try: Add: "Last line: (a + b + c) ÷ 3 = $X/sq ft x 1,900 = $Y."
  • Trusting the AI to count words. AI is bad at counting. Ask for a total word limit and it often lands over it. Try: Cap each line: "each line under 16 words."
  • Forgetting where it goes. AI loves markdown tables and bold, which show up as clutter in a seller text. Try: Add: "Plain text, no bold, headings or tables."

What a strong prompt includes

  • Fixes the square footage
  • Excludes the foreclosure
  • Asks to show the math
  • Sets a word limit
  • Asks for plain text

Practice this now ↑

Related exercises