RoleReady

Fair Housing

ChatGPT for Real Estate Listings: A Fair-Housing-Safe Workflow for Agents

Published by RoleReady · 8 min read · Practical Workplace Guidance

Planning to use AI at work?

Answer five questions about your tool, task, and data. No workplace text upload needed.

Try Safe to Paste Check →

Listing copy is the easiest place to lose an afternoon and the fastest place to inherit a complaint. AI tools will happily produce three hundred words of warm, fluent property prose in seconds. That fluency is the problem: the phrases that create the most risk are also the phrases that sound the most like good marketing.

Quick answer: Use AI to describe the property, never the people who might live in it. Feed it verified features, not assumptions. Run every draft through a fair-housing pass that removes buyer-describing language, neighbourhood quality and safety claims, and exclusivity wording. Then verify each factual claim, add the brokerage and licence identification your rules require, and confirm MLS-specific disclaimers before publishing.

Two things outrank everything in this guide: your brokerage policy and your MLS rules. Where they are stricter, they win. This article is educational material about writing habits, not legal advice, and it does not decide whether any particular phrase is lawful.

What AI is genuinely good at in listing copy

AI performs well when the raw material is factual and the task is transformation. Reliable uses:

  • rewriting a plain feature list into readable sentences;
  • translating a feature into a benefit without inventing anything;
  • producing headline variations so you can choose, not accept;
  • tightening copy that has run long for a portal character limit;
  • converting a listing into a walkthrough script, a social caption, or a follow-up note;
  • reviewing your own draft for words that describe a buyer rather than a home;
  • keeping tone consistent across a batch of listings.

AI performs badly when you ask it to supply judgment, facts, or market claims. Avoid:

  • inventing features, upgrades, or amenities that are not in your notes;
  • comparing neighbourhoods, schools, or crime levels;
  • claiming an area is safe, quiet, family-friendly, or exclusive;
  • producing price opinions, investment projections, or "hot market" claims;
  • describing the ideal buyer, tenant, or household;
  • writing anything about protected characteristics, even as a compliment.

A useful mental model: the model may rephrase what you verified. It may not add facts or opinions. If a sentence contains a claim, that claim has to come from your own records.

The core rule: describe the property, never the people

Fair-housing advertising rules exist to stop housing marketing from steering people toward or away from homes because of who they are. In practice, the writing rule is simple and it eliminates most risk in one stroke.

Describe the house. Do not describe the household.

Sentences about the home are usually fine: three bedrooms, south-facing garden, renovated kitchen, two-car driveway, finished basement, step-free entry, ground-floor bedroom. These are verifiable facts about a building, and any buyer can evaluate them.

Sentences about the people are where problems start: perfect for a young family, ideal for a professional couple, great for retirees, safe for children, a quiet street for people who value privacy.

The second group feels warmer and converts better in the writer's imagination. It also does something the first group never does: it tells a reader who the home is not for. That is the risk shift worth internalising.

The phrases to strike

Treat this as a risk-awareness list, not a list of legal rulings. Different jurisdictions, brokerages, and MLSs handle wording differently, and only your broker or counsel can tell you what is acceptable in your market.

Buyer-describing language

  • ideal for families, couples, singles, students, or retirees;
  • perfect for a young professional, a growing family, or an active adult;
  • great starter home for a young couple;
  • safe for children, child-friendly, or kid-friendly;
  • suits a mature buyer, empty nesters, or downsizers;
  • no children, adults only, or professional tenants preferred.

Neighbourhood quality, safety, and school claims

  • safe neighbourhood, low crime, or the safest street in town;
  • good schools, top-rated schools, or best school district;
  • family neighbourhood or a great place to raise children;
  • exclusive area, the right kind of neighbours, or prestigious postcode;
  • walking distance to a place of worship, or a description of the area's demographic makeup.

Exclusivity and filtering language

  • exclusive community, private enclave, or members only;
  • no vouchers, no housing assistance, or employed applicants only;
  • must be financially qualified before viewing, used as a screening message rather than a process note;
  • references to religion, nationality, marital status, disability, or source of income in the copy;
  • any phrasing that implies a preference about who should apply.

Two habits remove most of this. First, replace people-words with property-words: "great for families" becomes "three bedrooms and a fenced garden". Second, replace quality claims with verifiable facts: "good schools" becomes nothing at all, or a neutral pointer telling buyers to check official sources themselves.

A repeatable workflow: features in, safe copy out

Use the same five stages for every listing. The sequence matters more than the prompts.

  1. Features in. Start from a verified feature list built during the walkthrough: rooms, dimensions, orientation, condition, updates with dates, appliances included, parking, access, and any known limitations. No assumptions, no hearsay from the seller about work that was never done.
  2. Benefit translation. Ask the AI to turn each feature into a benefit that stays strictly inside the fact. "New boiler, replaced last year" becomes "recently replaced boiler" — not "worry-free living for years to come".
  3. Fair-housing pass. Run the draft through a dedicated review prompt that flags buyer-describing language, neighbourhood quality and safety claims, and exclusivity wording. You decide what stays.
  4. Verify every claim. Check dimensions, ages, inclusions, and any superlative against your notes. Delete anything you cannot confirm. A model will not tell you that a claim is unverifiable unless you order it to.
  5. Add required identification and disclaimers. Insert your brokerage name, licence identification, and any MLS-required notices, then check the final text against the current MLS field rules and character limits before it goes live.

Keep the raw feature list and the review notes. If a complaint ever arrives, a documented process is worth more than a memory of what you meant.

Eight copy-ready prompts

Every prompt below is written to keep the AI inside facts you supply.

1. Feature to benefit

Turn this verified feature list into listing copy.
Rules: do not add any feature, upgrade, or claim that is not in the list. Do not mention who the home might suit. Do not describe the neighbourhood, schools, safety, or the people who live nearby.
For each feature, give one factual benefit sentence.
Feature list: [PASTE YOUR VERIFIED FEATURES]

2. Headline variations

Write eight listing headline options for this property.
Base them only on these verified features: [FEATURE LIST]
Rules: no superlatives, no "best" or "safest", no mention of families, couples, retirees, children, schools, or neighbourhood character. Maximum 12 words each.
Label each option with the feature it leads on.

3. Walkthrough script

Write a 90-second walkthrough script for a video tour.
Order: exterior and approach, main living space, kitchen, bedrooms, bathroom, outdoor space, parking.
Use only these verified details: [DETAILS]
Rules: describe the property, not the ideal buyer. No safety, school, or neighbourhood-quality claims. Flag any point where I should add a fact with [VERIFY].

4. Social captions

Write five social captions for this listing, each under 40 words.
Use only these verified details: [DETAILS]
Rules: no buyer descriptions, no neighbourhood quality or safety claims, no exclusivity language.
Include one caption that leads with a single standout feature and one that simply invites a viewing.

5. Open-house follow-up

Write a short follow-up message to people who attended an open house.
Include: a thank-you line, one factual reminder about the property, an invitation to ask questions, and a neutral next step.
Rules: do not assume the recipient's circumstances, household, budget, or intentions. Do not mention other attendees. Keep it under 120 words and leave placeholders for names and dates.

6. Market update to a seller

Help me draft a market update for a seller client.
Structure: what has changed in the local market category, what it means for their asking position, what we will adjust, and what happens next.
Rules: use only the figures I supply. Do not invent statistics, averages, or forecasts. Where I have not supplied a number, write [NEEDS DATA]. Make clear that this is an opinion based on supplied information, not a guarantee.

7. Fair-housing self-review of your own copy

Review this listing copy as a fair-housing risk check.
Flag, with the exact phrase quoted: (a) language describing or addressing a type of buyer, household, or person; (b) claims about neighbourhood safety, quality, schools, or character; (c) exclusivity or filtering language; (d) any factual claim that cannot be verified from the text itself.
For each flag, suggest a factual, property-focused replacement. Do not tell me whether anything is legal. Output a table: phrase, category, suggested rewrite.

8. Client consent note for using an AI tool

Draft a short note to a seller explaining that AI tools may be used to help draft and proofread marketing text for their property.
Include: what the tools are used for, what is not shared (private documents, personal information, financial details), that final copy is reviewed and approved by me before publication, and that all factual claims are verified.
Keep it under 150 words, in plain language, with placeholders for the brokerage name and my contact details. Do not make promises about any third-party provider's data practices.

Pre-publish checklist

Run this on every listing, every time. It takes two minutes.

  • Every feature, dimension, and upgrade is verified against my own notes or records.
  • Nothing in the copy describes or addresses a type of buyer, household, or person.
  • No claims about safety, crime, schools, neighbourhood character, or who lives nearby.
  • No exclusivity, screening, or filtering language in the marketing text.
  • No invented statistics, market figures, or price opinions.
  • Superlatives removed or replaced with a specific, provable fact.
  • Brokerage name and licence identification present as required.
  • MLS-required disclaimers, field rules, and character limits satisfied.
  • Fair-housing review pass completed, and I have made the final call on every flag.
  • Photos, floor plans, and copy agree with each other.

Brokerage policy and MLS rules always come first

This workflow is a drafting habit, not a compliance programme. Your brokerage's advertising policy, your MLS's rules and field definitions, your state or provincial requirements, and any applicable fair-housing advertising rules all take precedence over anything suggested here. If your broker says a phrase is off limits in your market, it is off limits, whatever a model tells you.

Two practical habits make that easier to live with. First, confirm with your broker which AI tools, if any, are approved for marketing work, and under what conditions. Second, keep a copy of the approved wording for identification and disclaimers somewhere you can paste it in seconds, so the compliant version is also the quickest version.

If your brokerage has not yet set a position on AI-assisted copy, that is worth raising at your next sales meeting. A five-minute conversation now is cheaper than rebuilding a listing description after publication.

If you work with confidential client material as well as listings, the Safe to Paste check walks through five questions about your tool, task, and data before you type.

The Safe AI at Work: Employee Quickstart Kit collects the wider system in one download: the green, yellow, and red data guide, the five-question Safe to Paste check, an abstract-first workflow, 30 copy-ready prompts, a printable Safe Task Worksheet, review reminders, and an AI Wins Tracker. Agents use it to build the habit of checking before pasting.

Get the $19 Safe AI at Work Employee Quickstart Kit

The RoleReady books page lists the full handbook and prompt collections if you would rather start with the reference shelf.

This guide is educational material about writing and workflow habits, not legal advice. It does not interpret fair-housing law, MLS rules, or brokerage policy, and it is not a substitute for guidance from your broker, your MLS, or a qualified lawyer in your jurisdiction.

Take Safe AI Further in Your Daily Workflow

Get complete copy-ready prompt libraries and reference handbooks designed specifically for workplace professionals:

Toolkit

Safe AI at Work: Quickstart Kit

15-page guide, 30 copy-ready safe prompts, printable desk worksheets, and the 5-point human review protocol.

$19 one-time Get the Kit ($19) →
Kindle Handbook

Safe AI at Work (Kindle Edition)

The complete 91-page digital handbook on Amazon. Instant delivery to your Kindle or Kindle app.

$2.99 on Amazon View on Amazon ($2.99) →
Educational disclaimer: This guide provides practical productivity education and risk-reduction methods, not legal, cybersecurity, compliance, or regulatory advice. Employer policy, approved tooling, contracts, and accountable managers always come first.