Stop Writing CMAs That Take All Afternoon

Back when I still thought AI was a toy for tech bros, I spent a Saturday afternoon pulling comps for a listing presentation. Three hours hunched over spreadsheets, MLS tabs open like a dozen browser windows was some kind of productivity. My wife walked past the office twice without saying a word, which is how you know you're being an idiot.

The presentation was fine. Clean. Professional. The seller chose another agent anyway. Someone younger who got her CMA to her inside half a day.

"You were my second choice," the seller told me on the phone. "She just got back to me faster."

Nineteen years in the business and I lost a listing on speed. Not expertise. Not pricing accuracy. Speed. That one stung in a way that took me a while to shake.

Here's what I didn't know then: the comps, the adjustments, the narrative that explains why this price makes sense. A good AI prompt can do the heavy lifting in about ninety seconds. Not the final judgment call at the end. That's still your name on the line and your experience in play. But the grunt work that eats your Saturday? That's midnight paperwork, my friend, and you don't have to do it anymore.

Here is the prompt I use now. Copy it. Paste it. Fill in the brackets with your property details and let the new junior assistant earn its keep.

You are an experienced real estate appraiser and CMA specialist helping a listing agent prepare a Comparative Market Analysis. Your job is to produce a thorough, client-ready CMA that a seller would actually read and trust.

PROPERTY DETAILS:
- Address: [full property address]
- Bedrooms: [#] | Bathrooms: [#] | Square Footage: [#,###]
- Lot Size: [if relevant]
- Year Built: [####]
- Notable Features: [list 3-5: pool, renovated kitchen, ADU, view, etc.]
- Property Condition: [fixer / average / well-maintained / fully renovated]
- Days on Market Goal: [e.g., 30 days]

INSTRUCTIONS:

1. COMPS ANALYSIS
Identify 5-7 comparable sold properties from the last 6 months within [X]-mile radius. For each comp, include:
- Address and sale date
- Bed/Bath/SqFt
- Sold price and price per square foot
- Key differences from subject property (bigger lot, older kitchen, etc.)
- Adjustment notes (add or subtract value for each meaningful difference)

2. ACTIVE COMPETITION
List 3-5 currently active listings that a buyer would cross-shop against this property. Include asking price, days on market, and one sentence on how the subject property compares.

3. PRICE RANGE RECOMMENDATION
Give a recommended list price with a supporting range (e.g., "$425,000 to $445,000, recommended at $435,000"). Explain your reasoning in plain language. What comps support this, where the risks are, and what would need to happen for a higher or lower price to work.

4. MARKET NARRATIVE
Write 3-4 paragraphs a seller would actually read and understand. Cover: what's happening in this neighborhood right now (buyer demand, inventory, days on market trends), what similar homes are actually selling for, and why your recommended price positions this property to sell. No real estate jargon unless you explain it.

5. PRICING STRATEGY OPTIONS
Briefly outline three scenarios:
- Aggressive (sell fast, below market)
- Market-aligned (your recommendation)
- Aspirational (test a higher price, risk sitting)

For each, note the likely timeline and trade-off.

FORMAT: Use clear section headers. Use bullet points for comp details. Write the market narrative in full paragraphs. No fluff, no filler. Sound like a professional who respects the seller's intelligence.

A couple notes on making this work for you.

If you are in a market where six-month-old comps are useless because inventory moves fast, change that window to three months in the first line of the Comps section. Same goes for the radius. Tighten it for a dense urban neighborhood, widen it for rural. The prompt adapts fine either way.

Also worth knowing: the AI does not have live MLS access, so you still need to pull the actual comp data from your MLS and feed it in. But here is the thing. Once you paste those comps in as context above the property details, the AI handles the adjustments, the narrative, and the strategy framing in under two minutes. You just saved yourself two hours of formatting and writing.

Try this on your next listing presentation. Even if you already have a CMA template you like, run this alongside it tonight with a property you know cold. See what it gives you that you would not have written yourself. The worst case is you spent ninety seconds learning something. The best case is you never lose a listing on speed again.