ChatGPT Prompts for Real Estate Agents That Don't Sound Like AI

The best ChatGPT prompts for real estate agents give the AI three things before asking for anything: your verified facts, two or three samples of how you actually write, and clear rules on length, tone and Fair Housing. Prompts that skip those produce the generic copy buyers scroll past. The ten prompts here follow that pattern and work in Claude and Gemini too.

ChatGPT is the default for most agents. NAR’s 2026 REALTORS® Technology Report, released in September 2026, found that 91% of agents who use AI use ChatGPT, 42% use Gemini, 27% Copilot and 25% Claude. Many prompt lists for this search are old or are paid packs of one-liners. These are longer on purpose: each one tells the model what to use, what not to invent and what to hand back.

For which tool to pay for, see our guide to the best AI tools for real estate agents by job.

How do you make ChatGPT sound like you instead of AI?

Make ChatGPT sound like you by giving it real samples of your writing once, asking it to describe your voice, and saving that description as standing instructions. Lists of banned words help, but samples do most of the work.

Prompt 1: Build your voice profile

I'm a real estate agent in [city/area]. Below are three things I wrote myself:
an email to a past client, a social post, and a text to a buyer.

Read them and describe my writing voice in 8-10 bullet points: sentence length,
words I use often, words I never use, how I open and close, humor, formality.
Then list 10 phrases that would sound wrong coming from me.

[paste sample 1]
[paste sample 2]
[paste sample 3]

What good output looks like: specific bullets you recognize (“short sentences, opens with the client’s name, no exclamation marks, uses ‘honestly’ a lot”), not flattery like “warm and professional.” If the bullets could describe any agent, add longer samples and run it again.

Save the result. ChatGPT takes it as custom instructions (OpenAI no longer lets personal accounts create new custom GPTs and plans to retire them, per its custom GPT retirement FAQ, checked October 2026). Claude keeps it in a Project (Claude’s pricing page lists up to five Projects on the free plan and unlimited on Pro as of October 2026). Gemini saves it as a Gem (Gems are listed on the free plan). Every prompt below then starts with your voice already loaded.

Prompt 2: Strip the AI sound from any draft

Rewrite the draft below in my voice (see my voice profile). Rules:
- Keep every fact. Add none.
- Cut filler openings and sign-offs.
- No "nestled," "boasts," "dream home," "don't miss out," "in today's market."
- Max one exclamation mark. Prefer short sentences.
- Return the rewrite, then list what you changed.

[paste draft]

What good output looks like: a shorter draft with the same facts and a change list you can scan. If facts moved or vanished, say so and ask it to try again.

What are the best ChatGPT prompts for real estate marketing?

The best ChatGPT prompts for real estate marketing turn facts you already have (a listing sheet, your MLS numbers, a past client’s story told with permission) into drafts you edit. These four cover the jobs agents use AI for most.

Prompt 3: Listing description from a feature sheet

Write MLS public remarks using ONLY the facts below. Max [MLS character limit]
characters. Describe the property, not the buyer. No claims about schools,
safety or the neighborhood's residents. Open with the strongest verified
feature. If a useful fact is missing, list it under "Questions for the agent."

[paste feature sheet]

What good output looks like: concrete features in plain language, nothing you didn’t supply, and a short list of questions. The full workflow, with a fact-check list and a before/after, is in our guide to AI listing descriptions and Fair Housing checks.

Prompt 4: Monthly market update from your own numbers

Using only the numbers below (pulled from my MLS for [area], [month year]),
write a 150-word market update for my email list in my voice. Explain what
changed from last month in plain English, and what it means for a seller
and for a buyer in one sentence each. Do not predict prices. Do not add
any number I didn't give you. Cite "[MLS name], [date pulled]" at the end.

[paste: median price, active listings, days on market, months of supply,
this month and last month]

What good output looks like: the same numbers you pasted, a clear comparison and no forecasts. Check each figure against your MLS report before sending.

Prompt 5: Follow-up after an open house

Write a short follow-up email to a visitor from my open house at [address]
on [day]. They asked about [question they asked]. Answer it using these
facts: [facts]. Offer one next step: [private showing / list of similar
homes / call]. Under 120 words. In my voice. No pressure language.

What good output looks like: an email that mentions the visitor’s actual question in the first line and reads like you typed it on your phone. For sequences and timing, see our guide to real estate email marketing.

Prompt 6: Objection-handling role play

Play a homeowner in [area] who is thinking about selling but says:
"[objection, e.g. We'll wait until rates drop]". Stay in character, push
back realistically, and don't make it easy. After 6 exchanges, step out
of character and give me three specific notes on what I said well and
what I should change.

What good output looks like: a homeowner who argues back with plausible reasons, then feedback that quotes your own lines. NAR’s report found 13% of AI users already use it for role play and objection handling.

How do you turn a podcast interview into content with ChatGPT?

Turn a podcast interview into content by pasting the transcript and asking for one format at a time, using only what was said. A transcript is the best input you can give an AI because it already holds your opinions, stories and local knowledge in your own words.

Prompt 7: Transcript to blog post

Below is a transcript of an interview where I was the guest. Write a
900-word blog post for my website based ONLY on what I said. Use my
exact phrasing where it's strong and mark direct quotes. Structure it
around the 3-4 questions a buyer or seller in [area] would search.
Answer each question in the first sentence under its heading. Don't add
statistics or claims that aren't in the transcript.

[paste transcript]

What good output looks like: headings phrased as real questions, answers in your words, quotes you can verify against the recording. Our step-by-step on turning a podcast episode into a blog post covers formatting and SEO.

Prompt 8: Transcript to social posts and newsletter

From the same transcript, give me:
1. Five social post drafts, each built around one specific point I made,
   under 60 words, with the timestamp of the moment it came from.
2. A 200-word newsletter section for my sphere: one story I told, why it
   matters to them, and a link placeholder to the full episode.
Use only what I said.

What good output looks like: posts tied to timestamps you can check and clip, and a newsletter section that sounds like a note from you rather than a press release.

You need a transcript first. If you don’t host a show, the quickest way to get one is to be interviewed: request a free guest spot on the ICONS of Real Estate Podcast and feed the transcript into these prompts. If you want a steady supply, Icons’ Concierge guesting plan books you on two podcasts a month.

How do you check AI drafts for Fair Housing problems?

Check AI drafts for Fair Housing problems by running a review prompt on anything client-facing, then reading the flagged lines yourself. The Fair Housing Act bars ads that indicate “any preference, limitation, or discrimination” based on race, color, religion, sex, handicap, familial status or national origin (42 U.S.C. § 3604(c)), and state laws often add more. An AI can repeat old listing language that does exactly that.

Prompt 9: Fair Housing review

Review the draft below for language that could indicate a preference or
limitation based on race, color, religion, sex, disability, familial status,
national origin, or any class protected in [state]. Flag:
- phrases describing the ideal buyer or renter instead of the property
- words about safety, exclusivity, or the people in the area
- references to places of worship or ethnic landmarks
For each flag, quote the line, explain the concern in one sentence, and
suggest a property-focused rewrite. If nothing is flagged, say so.

[paste draft]

What good output looks like: quoted lines with plain rewrites (“perfect for families” becomes “four bedrooms and a fenced yard”). This is a second pair of eyes, not a compliance sign-off. AI can miss problems and flag harmless lines. The duty stays with you and your broker; when unsure, check with your broker before publishing. Our list of Fair Housing advertising words goes further.

Prompt 10: Fact check before sending

Compare the draft with the source facts below. List every claim in the
draft, mark each as "in source," "not in source," or "contradicts source,"
and quote the source line for each match.

Draft: [paste]
Source facts: [paste]

What good output looks like: a table of claims you can scan in a minute. Anything marked “not in source” comes out unless you can verify it.

What should you never put in a ChatGPT prompt?

Never put client personal or financial information, lockbox codes, showing instructions or transaction documents into a consumer AI tool your brokerage hasn’t approved. NAR’s guidance on brokerage AI policies warns that many consumer AI tools store prompts or use them to improve their models, so that data could be exposed or reused.

Safe habits:

• Use first names or “the buyer” instead of full names
• Remove addresses from anything that isn’t public listing copy
• Turn off model training in the tool’s data settings if it offers that option
• Ask your broker which tools are approved for client data
• Never ask the AI to set a price, give legal advice or decide anything for a client