CMA + Market Report Toolkit
Pricing is one of the hardest conversations in real estate — not because agents don’t know the data, but because they can’t always translate it fast enough to use it in the room. These 4 AI prompts fix that. Paste in your MLS data, tell it your scenario, and get back a clean analysis and the actual words to say.
4 AI prompts
Works in ChatGPT or Claude
Copy and use
Power move — turn this into a project
If you’re going to use these regularly, create a Project in Claude (or a Custom GPT in ChatGPT) and paste all 4 prompts into the instructions. Then instead of copying a prompt each time, you just open the project and say “run tool 2 on this data” — it already knows what to do. Takes 5 minutes to set up and saves you that time before every appointment.
Step 1
Extract
Organize raw MLS data
Step 2
Analyze
Find the sweet spot + danger zone
Step 3
Translate
Turn data into words
Step 4
Position
Get your number + defense
Tool 01
Market Data Extractor
Paste in whatever your MLS gives you — this organizes it into a clean, usable snapshot.
The prompt — copy and paste into ChatGPT or Claude
Act as a real estate market analyst. I'm going to paste raw MLS data. Your only job is to organize it — do not give me advice yet.
DATA:
[paste your MLS export or stats here — active listings, sold listings, days on market, prices]
From this data, extract and present ONLY these fields in a clean table:
- Active listings: count + median list price + median DOM
- Sold (last 90 days): count + median sale price + median DOM
- List-to-sale ratio: average %
- Months of inventory: active listings ÷ (sold count ÷ 3)
- Price per square foot: median (if data includes it)
Below the table, write exactly 2 sentences summarizing the market in plain English — no jargon, no advice. Just what the numbers say.
If a field is missing from my data, flag it with "MISSING — need to pull from MLS" and move on.
Paste directly into ChatGPT or Claude
Tool 02
Price Band Analyzer
Breaks your sold data into price bands so you can see exactly where homes sell fast — and where they die on the market.
The prompt — copy and paste into ChatGPT or Claude
Act as a real estate pricing analyst. I'm going to give you individual sold property data. Do the math — don't summarize, calculate.
SOLD DATA (last 90 days):
[paste each property: list price, sale price, days on market — one per line or as a table]
TARGET PRICE RANGE: $[low] to $[high]
Instructions:
1. Sort properties into $[25,000 or 50,000]-wide price bands by LIST price within my target range
2. For each band calculate:
- Number of sales
- % sold at or above list price
- % sold below list price
- Median days on market
- Average list-to-sale ratio
3. Present as a table — one row per band
Then below the table, answer these three questions in plain sentences:
- Which band is the sweet spot? (fastest, highest list-to-sale)
- Where does velocity drop off? (slower DOM, more price reductions)
- If a seller insists on the top of the range, what does the data say happens?
No advice. No recommendations. Just what the numbers show.
Pro move: run this before every listing appointment
Tool 03
Talking Points Generator
Pick your seller situation, paste your numbers, get the actual words for the appointment. Not a summary — a script.
The prompt — copy and paste into ChatGPT or Claude
Act as a real estate listing coach preparing me for a seller conversation. I don't need data analysis — I need words.
MY STATS:
[paste 3–5 key numbers: median sale price, DOM, list-to-sale ratio, months of inventory, price band data if you have it]
MY RECOMMENDED PRICE: $[your number]
SELLER'S EXPECTATION: $[their number] (if same, write "aligned")
SELLER SITUATION — pick the one that fits:
[ ] Wants to list higher than the data supports
[ ] Nervous and needs confidence that the price is right
[ ] Comparing to a neighbor's list price (not what it sold for)
[ ] Already had one failed listing somewhere else
[ ] Shopping multiple agents / heard a higher number from someone else
Write me:
1. An opening statement (2–3 sentences) that starts with market reality, not my opinion
2. Three data-backed sentences I can say out loud — use the actual numbers I gave you, written conversationally
3. If there's a price gap: one reframe that doesn't sound like I'm talking them down
4. One closing question that moves toward a decision
Write it like I'm actually going to say it. Not a report. Not bullet points with labels.
Run this the morning of your appointment
Tool 04
Price Position Advisor
Give it the property and your comps. Get back one number, the reasoning, and exactly what to say if the seller pushes back.
The prompt — copy and paste into ChatGPT or Claude
Act as a real estate pricing expert doing a pre-appointment CMA review. I need a defensible number, not a range.
SUBJECT PROPERTY:
- Location/neighborhood: [area]
- Beds/baths: [e.g. 3/2]
- Square footage: [SF]
- Year built: [year]
- Condition and updates: [describe — cosmetic only, fully updated, needs work, etc.]
- Features that affect value: [list — view, garage, lot size, pool, etc.]
- Negatives: [busy street, dated systems, small lot, etc.]
COMPARABLE SALES (last 90 days, similar size and location):
[list each: address or description, sale price, SF, sold date, notable differences from subject]
ACTIVE COMPETITION (current listings):
[list if relevant — buyers are comparing these right now]
Based on this, give me:
1. One recommended list price — not a range — and the one sentence reason why
2. The two comps that most support this number and what makes them relevant
3. Any adjustments you made: what, why, and how much
4. What realistically happens if this is listed $25,000 higher — DOM estimate and likely outcome
5. One paragraph I can say if the seller brings up a higher sale they heard about
Be direct. I walk into this appointment in [X hours/days]. I need to own this number.
Combine with Tool 3: run this first, then get your words
