For Property Appraisers ·
What you'll accomplish
Writing all the narrative sections of a URAR report from scratch takes 2-3 hours. This workflow drops that to 30-45 minutes of reviewing and editing. You'll generate neighborhood descriptions, market conditions, comp selection rationale, adjustment explanations, and reconciliation all at once from a single structured data input.
What you'll need
Instead of prompting ChatGPT once per narrative section (5 separate conversations), you'll build a single "report package" prompt that generates all sections at once from a structured data input. This means you enter your property data once and get a complete first draft of all narratives in one session.
What you should see after setup: You type your property data in a consistent format, paste it into ChatGPT, and receive 5-6 complete narrative sections ready to review and paste into your forms software.
Troubleshooting: If you get generic output, your data input is too vague; the more specific your property facts, the better the output.
Open a notes app or document and type this prompt template. You'll save this permanently and paste it at the start of each report session:
You are an assistant helping a state-licensed certified residential appraiser write professional report narratives. All output must use appropriate appraisal terminology, be objective and factual, and comply with USPAP standards. Do not include opinions of value: only describe the market, property, and comparable sales.
For the following property data, generate these sections in order:
1. NEIGHBORHOOD DESCRIPTION (URAR format: location, boundaries, employment access, amenities, schools, market character)
2. MARKET CONDITIONS (describe market trend using the statistics provided, mention supply/demand, price trend, DOM trend)
3. COMPARABLE SELECTION RATIONALE (explain why these comps were selected and note any limitations)
4. ADJUSTMENT EXPLANATIONS (for each adjustment listed, write 2-3 sentences explaining market support)
5. RECONCILIATION (summarize the weight given to each approach and explain final value conclusion)
PROPERTY DATA:
[You will paste your data here each time]
Now create a standard data input format. This goes at the bottom of the master prompt, after "PROPERTY DATA:"; you fill this in for each specific report:
Subject Property:
- Address: [address]
- Type: [SFR/condo/etc]
- GLA: [sqft]
- Beds/Baths: [#/#]
- Year Built: [year]
- Condition: [C1-C6 or describe]
- Notable features: [pool/garage/renovations/etc]
- Location: [city, neighborhood name if applicable]
Neighborhood:
- City/Metro: [city, state]
- Type: [suburban/urban/rural]
- Nearby employment: [employers, distance]
- Schools: [district name, above/below/avg rating]
- Amenities: [shopping, transit, parks within X miles]
- Market trend: [increasing/stable/declining]
Market Statistics (from your MLS):
- Median sale price now: $[amount]
- Median sale price 12 months ago: $[amount]
- Average DOM: [days]
- List-to-sale ratio: [%]
- Active listings: [#, trending up/down]
Comparable Sales:
COMP 1: [address], [GLA sqft], [beds/baths], sold $[price] on [date], [distance] miles
COMP 2: [address], [GLA sqft], [beds/baths], sold $[price] on [date], [distance] miles
COMP 3: [address], [GLA sqft], [beds/baths], sold $[price] on [date], [distance] miles
Adjustments:
- [Feature]: +$[amount] or -$[amount] ([brief reason])
- [Feature]: +$[amount] or -$[amount] ([brief reason])
Reconciliation notes:
- Final value: $[amount]
- Primary weight given to: [which comp(s) and why]
Open ChatGPT. Start a new conversation. Paste your Master Report Prompt, then immediately below it paste your Data Input Template filled in with real data from your current report.
Press Enter and wait 30-45 seconds.
What you should see: Five labeled sections appear, each written in professional appraisal language. The neighborhood description will reference your specific location. Market conditions will use your statistics. Adjustments will have professional explanations.
Troubleshooting: If the output feels generic, add more specific local details to your data input (name specific employers, name the school district, mention a nearby landmark). Specificity drives output quality.
Read each section critically. As a licensed appraiser, you're responsible for every word. Check:
Make edits where needed (usually minor word changes for accuracy), then copy each section and paste into the corresponding field in Total, ACI, or your forms software.
Time check: Initial review and editing of AI-generated narratives typically takes 15-20 minutes, dramatically faster than writing from scratch.
For a declining market:
Add to market stats: "Note that absorption has increased 20% in past 90 days. Market is declining. Emphasize conditions exposure period lengthening."
For limited comparable sales:
Add: "Note limited comparable sales in this market, only 3 sales within 1 mile in past 12 months. Explain why expanded geographic or time search was necessary."
For a non-conforming property:
Save this prompt. You'll paste it into a fresh ChatGPT conversation at the start of each report.
Add: "Subject has a non-conforming feature [describe]. Include a brief note in the neighborhood section about this and how it was handled."
For a condo:
Change "NEIGHBORHOOD DESCRIPTION" to include: "Add a brief project overview section describing the condominium project, homeowner association, and any project-level conditions."