listicle · Real Estate Agents
ChatGPT Prompts for Real Estate Agents (2026): 25 Prompts That Stay Inside Fair Housing Rules
Twenty-five copy-and-paste ChatGPT prompts for real estate agents — listing copy, follow-ups, objections, CMA narratives and social captions — plus the fair housing rule that decides whether the output is publishable, and October 2026 pricing for the three assistants worth running them in.
Affiliate disclosure: This article contains affiliate links. If you buy through them, we may earn a commission at no extra cost to you. It never influences our rankings — read our full disclosure.
Here is the answer up front: the prompts that save agents the most time are not the ones that write marketing copy — they are the ones that force the model to describe the property instead of the buyer. The 25 prompts below are grouped by job: listing descriptions, follow-up email, objection handling, pricing narratives, social and open-house copy, and negotiation. Every one is written so the output has a chance of clearing the Fair Housing Act’s advertising rule, explained below with the statute and the HUD guidance linked.
This is not a record of agents trying these prompts and reporting back — we do not run tools on this site. What we do is read the vendor’s own pricing page and the government’s own text, stamp the date, and tell you where the lines are. Every price below was checked in October 2026.
The fair housing rule that decides whether your AI copy is publishable
The statute. 42 U.S.C. § 3604(c) makes it unlawful “to make, print, or publish, or cause to be made, printed, or published any notice, statement, or advertisement, with respect to the sale or rental of a dwelling that indicates any preference, limitation, or discrimination based on race, color, religion, sex, handicap, familial status, or national origin.” That is the whole test: does the advertisement indicate a preference or limitation?
The regulation. 24 CFR 100.75 defines what counts as an advertisement — flyers, brochures, signs, banners, posters, billboards, and “any documents used with respect to the sale or rental of a dwelling” — and states that discriminatory notices include “[u]sing words, phrases, photographs, illustrations, symbols or forms which convey that dwellings are available or not available to a particular group of persons.” It points to 24 CFR part 109, HUD’s advertising guidelines.
The AI clarification. HUD’s press release 24-098, dated 2 May 2024, states plainly: “Housing providers, tenant screening companies, advertisers, and online platforms should be aware that the Fair Housing Act applies to tenant screening and the advertising of housing, including when artificial intelligence and algorithms are used to perform these functions.” The model is not a defence: whoever publishes the listing is the one who published it.
What that looks like as a prompt rule
Describe the property. Never describe the people. The standard realtor-board advertising word list — the version distributed by Dayton Realtors prints a revision date of 05/15/06 and a legend marking bold entries “not acceptable” and italic entries “caution” — runs three pages, and the entries that get agents in trouble cluster in four groups: familial status (“children, no”, “families, no”, “adults only”, “couples only”, “empty nesters”); religion (“Churches, near”, “Mosque”, “Parish”, “Catholic”); race, national origin and disability (“Integrated”, “safe neighborhood”, “Exclusive”, “Restricted”, “physically fit”, “no section 8”); and proxy signals whose real function is to say who the neighbours are (“school name or school district”, “Prestigious”, “gated community”).
SmartMLS, a Connecticut multiple listing service, states the operating rule in one line on its compliance page: describe the property, not the people, and do not use ethnic references, nationalities, religious references, or sex, gender and sexual orientation terms in advertising. That is the rule the prompts below are built around. It is not legal advice and does not replace your brokerage’s compliance review.
The 25 prompts at a glance
| Prompt set | The job it takes off your desk | Prompts |
|---|---|---|
| Listing descriptions | First draft, property-first rewrite, fair housing audit | 5 |
| Follow-up email | Reviving cold leads, post-showing, price cuts, past clients | 4 |
| Objection handling | Price, rates, FSBO, photo-versus-reality | 4 |
| Pricing and CMA narratives | Explaining comps, days on market, credits, net sheets | 4 |
| Social and open house | Captions, invitations, neighbourhood copy, ad variants | 4 |
| Negotiation and repairs | Repair requests, multiple offers, counters, timelines | 4 |
The prompts are written to fail loudly if you leave a bracketed value blank, which is the behaviour you want — a model that invents a square footage is worse than one that asks.
1. Listing description prompts
1. First draft from the spec sheet. “Write a property listing description using only the facts I give you — add no features, dimensions or amenities not in my notes. 120 to 150 words, neutral and factual. Notes: [beds/baths, square footage, lot, year built, renovations, systems age, parking, HOA terms].” The no-invention instruction matters; without it you get a fireplace the house does not have.
2. Property-first rewrite. “Rewrite the description below so every adjective describes the building, the lot or the location’s physical features — never the buyer, the occupants or the neighbours. If a sentence only works by implying who should live there, delete it. Description: [paste].”
3. Fair housing self-audit. “Act as a fair housing reviewer for residential advertising. Flag every word, phrase or implication below that could indicate a preference or limitation based on race, color, religion, sex, handicap, familial status or national origin, or that references schools, churches or neighbourhood demographics. Quote the phrase, name the protected class it touches, and propose a property-based replacement. Text: [paste].” This is the prompt worth memorising. Run it on copy you wrote yourself, not just copy the model wrote.
4. Three registers, one property. “Write three versions of this description: portal listing, printed brochure, luxury placement. All three must describe only the property and keep the underlying facts identical. Facts: [paste].”
5. Compress for portals. “Cut this description to 60 words without losing a verifiable fact. Keep the numbers. Remove every adjective that cannot be checked against my notes. Text: [paste].”
2. Follow-up email prompts
6. Revive a cold lead. “Write a four-email sequence to a buyer lead who went quiet [N] weeks ago. Each gives one specific reason to reply — a matching listing, a rate change, a market fact — and none uses guilt or manufactured urgency. Criteria: [paste]. Sign as [name], [brokerage].”
7. Post-showing follow-up. “Draft a follow-up to a buyer who viewed [address] today: one sentence confirming what they saw, one offering two concrete next steps, one question that needs more than a yes. Direct tone, no exclamation marks. Showing notes: [paste].”
8. Price-reduction announcement. “Write the message to a seller explaining that a price reduction is the next lever, using only the market facts I supply. No hedging. Include one paragraph the seller can forward explaining the reasoning. Facts: [days on market, comparable sales, showing feedback, competing inventory].”
9. A past-client check-in that is not a mass email. “Write a short message to a past client [N] months after closing, referencing [specific thing]. Offer one genuinely useful item — a tax appeal deadline, a maintenance reminder, a comparable sale on their street — and nothing else. Do not ask for a referral.”
3. Objection-handling prompts
10. “Why is this priced above the comps?” “A buyer says the asking price is above recent comparable sales. Write a response separating the three things they may be conflating — list price, closed sales, active competition — using only these numbers: [paste]. End with a question that moves to offer strategy.”
11. “I will wait for rates to drop.” “Write a response to a buyer who wants to wait for lower rates. Cover the trade-off between rate and price in plain arithmetic using my numbers: [paste]. Do not predict where rates are going. Do not use the phrase ‘marry the house, date the rate’.”
12. The FSBO conversation. “Write a script for a first conversation with a for-sale-by-owner seller. Open with a question, not a pitch. Cover pricing accuracy, buyer-agent access and the paperwork timeline. No disparagement, no unsupplied statistics. Market facts: [paste].”
13. “The photos look better than the house.” “A buyer’s agent says the listing photos oversell the property. Write a reply that acknowledges the point, states what the photos actually show, and offers one specific remedy — a re-shoot, a virtual decluttering or an adjusted showing plan. Four sentences.”
4. Pricing and CMA narrative prompts
14. CMA narrative in plain English. “Turn this comparable-sales data into a two-paragraph pricing narrative a seller can follow without a spreadsheet. State the recommended range, the two sales that drive it, and the one adjustment that would change the recommendation. Data: [paste].”
15. Explaining days on market. “Explain why a listing at [address] has been on the market [N] days while comparable homes sold faster. Compare price position, condition and photography only. Do not speculate about buyer behaviour. Facts: [paste].”
16. Repair credit versus price cut. “A seller is choosing between a [amount] repair credit and a [amount] price reduction. Show the arithmetic both ways, say which the net proceeds favour, and name the one situation where the other choice wins. Numbers: [paste].”
17. Net sheet walkthrough. “Explain this seller net sheet line by line in the order the money leaves the transaction. Define each term in one sentence. Flag every line that is a range rather than a fixed number. Net sheet: [paste].”
5. Social and open-house prompts
18. Caption from listing copy. “Turn this description into three Instagram captions: one under 15 words, one leading with the strongest single feature, one written as a question. All three stay property-first. Description: [paste].”
19. Open-house invitation. “Write an open-house invitation for [address] on [date, time]. Include what a visitor will see, parking, and one reason to come this weekend rather than next. Do not describe the likely attendees. Details: [paste].”
20. Neighbourhood copy without the fair housing trap. “Write a 90-word description of the area around [address] using only physical, verifiable features: distances, transit, parks, amenities, zoning. Do not name schools or churches or say anything about residents. Leave a blank rather than guessing. Notes: [paste].”
21. Ad variants. “Write five variants of this ad headline and first line for A/B use. Each leads with a different property attribute. Do not vary anything that touches a protected class, and do not change the facts between variants. Base copy: [paste].”
6. Negotiation and repair prompts
22. Repair request. “Draft a repair request from these inspection findings. Group items into safety, systems and cosmetic. Ask for repair on the first two and a credit on the third, citing the inspection page number for each item. Findings: [paste].”
23. Multiple-offer explanation. “Explain to a buyer, without disclosing any other offer’s terms, why the seller chose a different offer. Cover price, financing type, contingency structure and closing date, and state plainly that the specific terms are confidential. Deal facts: [paste].”
24. Counter-offer email. “Write a counter-offer email covering price, closing date and one contingency change, giving the reasoning for each in one clause. No rhetorical questions, no pressure framing. Terms: [paste].”
25. Closing-timeline message. “Write the closing-week message to a buyer: every remaining item, who owns it, its deadline. Flag the two items most likely to slip. Timeline: [paste].”
Which assistant to run these in
The prompts are plain text and work anywhere. These three are the ones worth paying for.
- Best for
- Drafting listing copy, follow-ups and objection replies at volume
- Pricing
- Free; Go $8/mo, Plus $20/mo, Pro from $100/mo (verified Oct 2026)
The free tier runs every prompt on this page. The cap that bites is usage on the strongest models during a heavy listing week — exactly when you are drafting most. It wins here on the rewrite loop: prompts 2, 3 and 5 all take text you paste back in, and ChatGPT is fast at that. The caveat is that default output reads generic, and clients notice unedited AI email.
Strengths
- Free tier genuinely usable — no card, no trial clock, no watermark
- Handles the rewrite-and-recheck loop the audit prompt depends on
- Go at $8/mo is a cheap step up when the free caps start to bite
Trade-offs
- Free tier throttles the strongest models during busy listing weeks
- Default prose is bland and recognisable — rewrite before sending
- No MLS or comps data; it will invent facts you do not supply
- Best for
- Working through long inspection reports, disclosures and addenda
- Pricing
- Free; Pro $20/mo or $17/mo billed annually; Max from $100/mo (verified Oct 2026)
The free plan is narrower on model access than ChatGPT’s, but it ships what matters for document work: file handling, web search, memory, connectors and Projects capped at 5. For prompts 17 and 22 — the net sheet walkthrough and the repair request that cites inspection page numbers — that beats raw model strength, because the job is holding one long document steady across several follow-ups. If you want only one assistant, take ChatGPT.
Strengths
- Free tier includes file handling, web search, memory, connectors and Projects
- Noticeably better at holding a long document in context across follow-ups
- Pro is $17/mo when billed annually, below ChatGPT Plus
Trade-offs
- No access to the strongest model on the free tier
- Projects capped at 5, which runs out fast at one per listing
- No Research mode on free
- Best for
- Keeping prompt output grounded in your own documents
- Pricing
- Free Standard plan, no credit card; paid tier via Google AI Plus (verified Oct 2026)
It answers only from the documents you upload and cites the passage it used, which makes it the right home for the facts your prompts depend on: your board’s MLS rules, your brokerage’s advertising policy, a rate sheet, the seller’s disclosure packet. Free with no credit card; daily caps apply to the heavier features. Google now surfaces it as Gemini Notebook in some menus — a rename, not a shutdown, whatever a 2025 article told you.
Strengths
- Free with no credit card, and answers are grounded in your uploaded documents
- Cites the exact passage behind each answer, so you can verify before quoting
- Turns a document set into a briefing in one pass
Trade-offs
- Cannot pull in outside knowledge — it only knows what you upload
- Daily caps on audio overviews and heavier features
- Not a drafting tool; it is where the facts live, not where the copy gets written
How to get a usable draft out of any of these
Four habits do most of the work. Give it the facts and forbid invention — a model that fills a gap with a plausible number creates a disclosure problem in the direction that makes the listing look better. Run the fair housing audit as a separate pass — asking one prompt to write and self-censor at once produces bland copy that still slips. Never let it describe the buyer — the failure mode is not slurs, it is the friendly sentence about growing families that reads as service and functions as a preference statement. Rewrite the first sentence yourself — model output opens generically, and the first line is the one a buyer reads.
For the wider paid stack, start from our best AI tools for real estate agents pillar. If listing copy is the main job you want covered, our roundup of the best AI listing description generators compares the dedicated tools against doing it in a general assistant, and the free AI tools roundup covers what you can run at $0. When the copy is done and the photos are not, our head-to-head on BoxBrownie vs Virtual Staging AI prices the two staging routes.
The bottom line
A prompt library is worth having not because it writes faster but because a good prompt encodes the rule you keep forgetting. The Fair Housing Act’s advertising provision, 24 CFR 100.75 and HUD’s May 2024 AI guidance point the same way: the advertisement must not indicate a preference, whether a person or a model typed it. Keep prompt 3 on your phone, run it before anything reaches the MLS, and the rest of the list saves you an afternoon a week.
Related guides
- The 7 Best AI Tools for Real Estate Agents in 2026 (Compared)
- 6 Best AI Listing Description Generators for Agents (2026)
- The 8 Best Free AI Tools for Real Estate Agents in 2026 (Free Tiers, Checked)
- AI Email Marketing for Real Estate Agents in 2026: 7 Platforms, Priced and Compared
Frequently asked questions
Can I publish AI-written listing descriptions as they come out of ChatGPT?
Not without reading them line by line. The Fair Housing Act makes it unlawful to publish any advertisement that indicates a preference, limitation or discrimination based on race, color, religion, sex, handicap, familial status or national origin, and HUD confirmed in its May 2024 guidance that this applies to advertising delivered online, including when artificial intelligence and algorithms are used to produce or target it. A model will happily write that a home is perfect for a growing family or a young couple if you let it — those are familial-status statements about people, not facts about the property.
What is the single most useful prompt to keep on hand?
The fair housing self-audit prompt in section 1. Paste your finished description back in and ask the model to flag anything that describes people rather than the property, references schools, churches, ethnic or national origin, or uses coded words like exclusive, safe or integrated. It is not a legal review and it will miss things, but it catches the obvious violations before a brokerage compliance officer or an MLS does.
Do I need ChatGPT Plus, or is the free tier enough?
For prompt work, the free tier covers most of it — ChatGPT Plus is $20 a month and the free plan handles drafting and rewriting. Plus buys higher usage limits and broader model access, which matters if you are drafting for a full listing inventory in one sitting. ChatGPT Go is the cheaper paid step and Pro starts at $100 a month. All prices verified October 2026.
Which assistant is best for the contract-and-disclosure side of the job?
Claude, because it holds a long document in context across follow-ups better than the alternatives and its free plan still includes file handling and Projects. Claude Pro is $20 a month, or $17 a month when billed annually. Use ChatGPT for volume drafting and Claude when you are reasoning through a 60-page inspection report.
Is it safe to paste client details, addresses or offer terms into a chatbot?
Assume no until your brokerage says otherwise. Consumer free tiers are not built for confidential data, and seller names, unlisted addresses and offer terms should not go into a personal account. Check whether your brokerage has an approved-tools list first, and prefer tools that let you opt out of having your conversations used for model training.
Keep reading
→Best AI Tools for Real Estate Investors in 2026: 6 Tools Compared
Most "AI" in investor software is a calculator with a marketing budget. These six tools are priced from the vendors' own pages in October 2026, with a clear line between the tools whose AI predicts a number you would otherwise be guessing and the ones that only crunch it.
→Best AI Chatbots for Real Estate Websites (2026): 6 Tools Ranked on Lead Capture
Your website gets visited at 11pm, when you are asleep. Six AI chatbots for real estate websites compared on lead qualification, MLS awareness, human handoff, and pricing — and which ones are built to turn that night traffic into appointments.
→Best AI CRM for Real Estate Agents in 2026: 6 Tools Compared
Six real estate CRMs compared on the three AI features that genuinely change a working week — lead scoring, smart follow-up, and behaviour-triggered drip — judged on documented capability and price rather than demo polish.