TL;DR: Real estate agents use AI for four things on repeat: listing descriptions, buyer/seller follow-up, market update content, and open-house promo. Below are copy-paste prompts for all four, plus the fair-housing wording rules AI doesn't know to follow and won't warn you about before you hit publish.
What Should Real Estate Agents Actually Use AI Prompts For?
Four jobs, on a loop, all year: writing the listing, following up with the lead, sending the market update, and promoting the open house. None of them are complicated writing tasks on their own; the problem is volume. A busy agent might draft fifteen listing descriptions, forty follow-up messages, and a dozen social posts in a single month, and rewriting each one from a blank page, instead of from a working template, is where the real time actually goes. Multiply that across a full pipeline of active listings, past clients, and open houses, and the writing alone can eat a morning that should have gone to showings.
AI collapses the blank-page part. It doesn't collapse the part where you're legally responsible for what gets published under your name, which is the one caveat that has to come before any of the prompts below rather than after them.
What Words Can Get Your Listing Description in Trouble?
The federal list is short, and most violations aren't intentional at all; they start out as a well-meaning adjective that happens to describe a buyer instead of a property. HUD's own overview of the Fair Housing Act names the protected categories as race, color, national origin, religion, sex, familial status, and disability, and its digital-advertising guidance specifically calls out "words, phrases, photographs, illustrations, symbols or forms which convey that dwellings are available or not available to a particular group of persons" as the kind of language that draws a complaint.
In practice, that means the following phrases (all things an AI model will cheerfully generate if you simply ask for copy in that direction) are the ones to strip out before publishing:
- Familial status: "family-friendly," "no kids," "perfect for empty nesters," "great for a bachelor"
- National origin / religion: "walking distance to [a specific church/temple/mosque]," "Hispanic neighborhood," "in the Jewish area"
- Disability: "able-bodied tenants only," "must be able to climb stairs" (describe the actual stairs instead, since that is a property fact, not a preference)
- Sex / race: "safe for a single woman," "quiet, mature neighborhood" used as coded shorthand
Many states and cities add protected categories the federal Act doesn't cover: source of income, sexual orientation, gender identity, and marital status are common additions. Check your own state's fair housing statute and your MLS's compliance rules, not just the federal floor above, since a state list is almost always longer than the federal one, never shorter. If you list outside the US, your jurisdiction's own equivalent law governs instead, and you, not the AI and not this post, are the one who has to know which one applies to you.
Here's what the rewrite looks like end to end. Ask a model for a description that's "perfect for a growing family," and it will produce exactly that phrase, or something close enough that a fair housing officer reads it the same way. Ask instead for a description that lists four bedrooms, a fenced backyard, and a finished basement, and it can't drift into protected-class territory, because there's nothing in the prompt to drift from. The compliant version isn't blander copy; it's just more specific, and specific is what the listing needed in the first place.
How Do You Prompt AI for a Listing Description That Stays Compliant?
Build the fair-housing instruction into the prompt itself, every time, rather than trusting yourself to catch it on review. It's one extra sentence and it's cheaper than a complaint.
1. Standard listing description. What it's for: the everyday MLS/Zillow description. What to change: property details, the one feature you want leading.
Write a 150-word MLS listing description for a 3-bed, 2-bath single-family home, 1,850 sq ft,
built 2015, updated kitchen with quartz counters, fenced backyard, attached 2-car garage. Lead
with the updated kitchen. Describe only the property's physical features and condition — do not
mention or imply any preference about who the buyer should be (no references to family status,
religion, national origin, sex, or disability, and no coded phrases like "walkable" or "quiet,
mature" used as a stand-in for those). Confident, warm tone, no exclamation points.
2. Luxury listing with a lifestyle angle. What it's for: higher-end copy that still needs to sell a feeling, without selling it via an implied buyer. What to change: the standout feature, the lifestyle angle.
Write a 200-word luxury listing description for a 5-bed waterfront estate with a private dock
and a home theater. Sell the lifestyle through the property's amenities and setting (the dock,
the view, the theater) — never through a description of who would live there. Avoid "family,"
"empty nester," "bachelor," or any age-, religion-, or ability-coded language. Sophisticated,
understated tone.
3. Fair-housing self-check pass. What it's for: a second prompt that reviews the first draft before you publish it. Treat this pass as mandatory, not optional; it's the cheapest insurance in this whole workflow. What to change: paste in your draft.
Review this real estate listing description for fair housing compliance. Flag any word or
phrase — even an indirect or "coded" one — that could imply a preference or limitation based on
race, color, national origin, religion, sex, familial status, or disability. For each flagged
phrase, explain what it implies and suggest a replacement that describes only the property.
Listing description: [paste draft here]
Save whichever version of prompt 1 works for your market as a reusable prompt template, so you're editing four details instead of rewriting the whole brief every time a new listing lands. The self-check prompt is the one worth saving hardest: it's the only step in this section that exists purely to catch a mistake before a client, a broker, or a fair housing complaint catches it for you, and it takes about as long to run as it takes to read this sentence.
What Are the Best AI Prompts for Buyer and Seller Follow-Up?
Follow-up is the highest-volume writing job most agents have, and it is the one where a generic AI reply reads as generic AI the fastest, because every buyer can tell when they've received the same three sentences as the last twenty leads. Persona prompting (telling the model who it is writing as, not just what to write) closes that gap; see persona prompting: make ChatGPT think like an expert for the underlying technique.
4. Post-showing follow-up. What it's for: same-day contact after a buyer tours a property. What to change: property address, what they responded to, next step.
Write a warm, brief follow-up text (under 50 words) to a buyer who toured 42 Elm Street this
afternoon and seemed most interested in the finished basement. Reference the basement
specifically, ask one open question about their timeline, and suggest a next step. No emojis,
no exclamation points, sounds like a real text from a person.
5. Cold-to-warm seller lead nurture. What it's for: a lead who requested a home value estimate but hasn't responded since. What to change: how long since they inquired, what you know about their property.
Write a short follow-up email to a homeowner who requested a home value estimate 10 days ago and
hasn't replied. Lead with one specific, current piece of market information relevant to their
neighborhood (not a generic "the market is hot" line), ask a low-pressure question about their
timeline, and keep it under 100 words. Confident but not pushy.
6. Buyer drip sequence, three touches. What it's for: a short nurture sequence for a lead not ready to tour yet. What to change: what triggered the lead, their stated budget range.
Write a 3-email nurture sequence for a buyer lead who saved a search for homes under $450,000 in
[neighborhood] but hasn't booked a showing. Email 1 (day 1): welcome and set expectations. Email
2 (day 4): one useful piece of neighborhood-specific information. Email 3 (day 9): a direct,
low-pressure ask to schedule a call. Each under 120 words, same sign-off.
If your leads live inside a CRM rather than a spreadsheet, the AI in that system already has context a general model does not. Follow Up Boss's own product page describes its AI as one that "summarizes your calls and recaps the lead's entire history at a glance", and says "Important tasks are suggested automatically and activated with a single click" (accessed September 2026). That is a narrower job than the prompts above: it reads your existing history instead of writing from a blank prompt, which is exactly what makes it useful for follow-up specifically and not much help for anything you haven't already logged, like a market update or a first listing draft.
What Are the Best AI Prompts for Market Update Emails and Social Posts?
Market updates are the content most agents mean to send monthly and actually send twice a year, because assembling the numbers into readable prose is tedious. AI doesn't have your MLS data, so the prompt has to carry it; this is exactly where a reusable structure earns its keep, since from one send to the next you're only ever swapping four or five numbers, not rewriting the whole email. Our email marketing prompts cover the same "supply your own numbers" discipline for non-real-estate audiences.
7. Monthly market update email. What it's for: the recurring newsletter your sphere expects. What to change: the actual stats.
Write a 250-word monthly market update email for [neighborhood/city]. Use these figures: median
sale price $[X], up/down [Y]% from last month; average days on market [Z]; inventory [rising/
falling/flat]. Explain what each number means for a buyer and what it means for a seller in one
sentence each. End with a single clear call to action to reply with questions. Plain, factual
tone — no "market is on fire" clichés.
8. Market update social post (3 platforms). What it's for: turning the same update into short-form social copy without starting over. What to change: the one headline stat.
Take this market stat — [median price up 4% year over year in ZIP 90210] — and write three
short social posts: a 2-sentence LinkedIn version (professional, data-first), a casual
Instagram caption under 40 words with one relevant hashtag, and a single punchy sentence for a
text-message blast. No emojis in the LinkedIn version.
9. Quarterly report for past clients. What it's for: a check-in that isn't a sales pitch, sent to people who already bought or sold with you. What to change: the client's original purchase date, current comparable stats.
Write a 150-word quarterly check-in email to a past client who bought their home in [month/
year]. Mention one relevant local market data point (without pressuring a sale), ask how the
home is treating them, and offer to answer any questions about their home's current value with
no obligation. Warm, low-pressure, personal tone.
All three of the prompts above share one habit worth naming: they ask for a specific number or a specific date, not a general "the market is strong" claim. That's the difference between a market update a past client actually reads and one that gets filed next to the last five that said the same thing in different words.
What Are the Best AI Prompts for Open House Promotion?
Open-house copy is short, repeats weekly, and is the easiest place to let a template do the work, because the only things that change from one Saturday to the next are the address, the time, and the one feature worth leading with. See ad copy for email, ads, and landing pages for the broader promotional-copy patterns this borrows from.
10. Open house social invite. What it's for: the announcement post that goes out a few days before. What to change: address, date/time, one standout feature.
Write an Instagram caption (under 60 words) announcing an open house at 118 Maple Court this
Saturday 1–3pm. Lead with the renovated primary suite. Include the date, time, and address
clearly, end with "link in bio for details," one relevant hashtag, no more than one emoji.
11. Open house email invite to your sphere. What it's for: a direct invite to your existing contact list, not cold social reach. What to change: address, date/time, why this one is worth a visit.
Write a short email (under 100 words) inviting my contact list to an open house this weekend.
Mention the address, date, and time in the first two sentences, name one specific reason this
home is worth seeing in person, and close with a line inviting them to bring a friend who might
be house-hunting.
12. Post-open-house follow-up (attendees). What it's for: same-day or next-day contact with everyone who signed in. What to change: the property, anything specific an attendee mentioned on the sign-in sheet.
Write a brief, friendly follow-up text (under 50 words) for someone who attended Saturday's open
house at 118 Maple Court. Thank them for coming, ask one open question about their impression,
and offer to send comparable listings if this one isn't quite the fit. No pressure language.
Which AI Tool Should Agents Use for Which Task?
There's no single best tool here, and the honest answer changes by job. Two of the rows below aren't prompt-driven at all; they're platform features you switch on rather than write a prompt for. General-purpose models (ChatGPT, Claude, Gemini) cover most of the twelve prompts above, since none of that writing needs real estate-specific training to execute well. The tools worth naming below are the ones doing something a generic prompt cannot: reading your actual lead history, or reaching buyers who never typed a prompt at all.
| Task | Best Fit | Why |
|---|---|---|
| Listing description draft | General model (ChatGPT, Claude, Gemini) | You control and review every word before it's published — non-negotiable given the fair-housing exposure above |
| Buyer/seller follow-up from existing lead history | CRM-native AI (e.g., Follow Up Boss) | Reads your actual call notes and lead activity; a general model can't see that context unless you paste it in |
| Market update email/social copy | General model | Needs current MLS stats you supply — no AI tool has your local market data baked in |
| Open house promo (social, email) | General model | Short, templated, low compliance risk, high repetition — the ideal shape for a saved prompt |
| Buyer-facing home search and lead routing | Platform AI (e.g., Zillow AI Mode) | Zillow describes its AI mode as helping "buyers and renters discover homes through conversation, receive guidance and take real steps such as scheduling tours or connecting with a local agent" (Zillow Front Porch, March 25, 2026) — this is a consumer-facing feature that routes leads to you, not something you prompt directly |
Where Does a Prompt Manager Fit Into an Agent's Day?
It sits underneath all twelve prompts above, not next to them. The recurring cost in a real estate AI workflow is not writing one good prompt; it is rewriting the same well-structured prompt every single listing, every single follow-up, every single week, because it lives in a note app or a group chat instead of somewhere reusable. That cost gets worse, not better, once you're bouncing between tools mid-day: drafting a listing in ChatGPT before a showing, then switching to Claude at your desk because that's what your brokerage happens to license.
Prompt Architects turns a rough one-line request ("write me an open house post") into the structured version above, and saves it as a reusable prompt library synced across ChatGPT, Claude, and Gemini, so the fair-housing self-check prompt and the four listing templates travel with you instead of getting rebuilt from memory on a Tuesday. The free plan includes 5 prompt enhancements a day, per its FAQ page; if you're running dozens of listings and follow-ups a month, the paid plans remove that cap.
Stop rewriting prompts. Start shipping.
Works with ChatGPT, Claude, Gemini, Grok, Midjourney, Ideogram, Veo3 & Kling. 4.8★ on the Chrome Web Store.
Create An AccountNone of this replaces judgment. The prompts above get you a strong first draft in seconds instead of twenty minutes; they don't know your MLS's specific compliance rules, your brokerage's review process, or which sentence in a draft is the one that would read wrong to a fair housing enforcement attorney. That review is still the job, on every listing, every single time. AI just gets you a much better first draft to start that review from, which, across a full month of listings, follow-ups, and open houses, is still a real amount of time back.