Key takeaways
- More than half of AI-using REALTORS already draft social posts and emails with AI; the gains come from better inputs, not cleverer prompts[1][2].
- Keep a voice guide and a fact library, and let AI work only from them.
- Never use AI to write reviews or testimonials; the FTC's 2024 rule bans fake reviews, including AI-generated ones[3].
- Every post is advertising: the true-picture standard and fair housing rules apply to captions and scripts as much as to listings[4][5].
Drafts, not autopilot
Marketing is where most agents first try AI. In NAR's 2026 technology survey, 56% of AI-using REALTORS said they use AI for social media posts and 52% for emails and follow-ups[1][2]. The appeal is obvious: a blank caption box at the end of a long day is hard, and a draft that is 80% right is a relief.
The trouble is that AI-written marketing tends to sound like everyone else's AI-written marketing. Generic captions about dream homes and hot markets do little for your reputation, and invented details can create real problems. The aim is to let AI do the first draft and the reformatting, while your voice, your local knowledge and your facts carry the content.
Build two inputs first
Before you ask for a single post, build two documents you will paste into every request. They take an hour to write and save that hour back within a week.
- A voice guide: three or four sentences about how you write, five phrases you use, five you never use, your typical length on each platform and your sign-off. Include two or three past posts you liked as examples.
- A fact library: current listing fact sheets, your recent sales you are allowed to share, local market figures with their source and date, and brokerage details that must appear in ads.
- Standing rules: no exclamation-mark hype, no claims about neighborhoods' residents, no numbers that are not in the fact library, and the brokerage name in every ad as your state requires.
With those in place, your requests can be short: write three Instagram captions for the listing in the fact sheet, using the voice guide, each under 60 words, one focused on the kitchen. The model has what it needs and very little room to invent.
Content types and who does what
| Content | AI's job | Your job | Check before posting |
|---|---|---|---|
| Listing social posts | Caption variations and platform lengths | Choose the angle and photos | Every fact against the listing sheet |
| Market update email | Explain figures you supply in plain words | Supply the figures and the local take | Every number and its source date |
| Newsletter | Structure, headlines, summaries of your notes | The opinions and stories | Tone and accuracy |
| Short video script | Outline, hook options, timing | Delivery and on-camera judgment | Claims you say out loud |
| Review replies | Polite draft responses | Approve or rewrite each one | No private details revealed |
| Blog or guide | Outline and first draft from your notes | Local knowledge, examples, final edit | Facts, links and fair housing wording |
A weekly social workflow
Plan
On Monday, list this week's three to five topics: a new listing, a just-sold, a local event, a market figure and a how-to tip.
Draft in one session
Paste the voice guide and the relevant facts, ask for all drafts at once, and request two options for each.
Edit for you
Cut generic phrases, add one local detail only you would know, and check every fact.
Compliance pass
Confirm brokerage identification, fair housing wording and that no photo or caption misrepresents the property.
Schedule
Queue posts for the week, and keep the final versions as examples for next week's voice guide.
Market updates: your numbers, not the model's
Market updates are where AI does the most damage when used carelessly, because models will happily produce plausible statistics. Never ask AI for market figures. Pull them yourself from your MLS reports or a named public source, paste them in with the source and date, and ask AI only to explain them in plain language for buyers or sellers.
Then add the part no model can supply: what the numbers mean on your streets this month. A sentence about how homes near the new school sold faster than the area average, from your own observation, is worth more than a paragraph of national trends.
Email follow-ups that do not read as automated
Email is where AI can either deepen relationships or quietly erode them. A past client who receives a generic monthly blast learns to ignore it; one who receives a short note about a sale on their street, written in your voice, remembers you. AI makes the second kind practical at volume, as long as the input is specific.
- Segment first: past clients, active buyers, sphere and new leads each need different messages and frequency.
- Give AI one specific hook per segment, such as a nearby sale, a seasonal maintenance tip or a rate change, with its source.
- Keep each email under 150 words with one clear reason to reply.
- Personalize from CRM facts only, such as the closing anniversary or the neighborhood, never from guesses.
- Send from a person, sign as that person and make sure replies reach them.
If a segment stops replying, the fix is rarely a new subject line. It is usually that the content stopped being about them. Read a sample of last month's emails as if you were the client and ask whether you would have opened them.
Video scripts
For short videos, AI is good at structure: a hook in the first two seconds, three points, a call to action, timed to the platform's length. Ask for three hook options and a 45-second script, then rehearse it in your own words. Viewers can tell when an agent is reading lines written for them, so treat the script as notes rather than a teleprompter.
- Keep claims in scripts as strict as captions; something said on camera is still advertising.
- Do not use AI-generated voices or faces that could be mistaken for you or for a real client.
- Show the real property; if footage or images are edited or virtually staged, say so on screen.
Reviews and testimonials
Reviews carry weight with consumers. Pew Research found 82% of US adults at least sometimes read online ratings or reviews before buying something for the first time[6], and people pay more attention to extremely negative reviews than to extremely positive ones[7].
Source: Pew Research Center, 2016[7]
Show the data
| Item | Value |
|---|---|
| Extremely negative reviews | 54% |
| Extremely positive reviews | 43% |
This is the one area where AI must not create content. The FTC's final rule, announced August 14, 2024, bans businesses from creating or selling fake reviews and testimonials, including AI-generated fake reviews[3]. It also prohibits offering incentives conditioned on writing reviews[8] and restricts reviews written by company insiders without disclosure of the connection[9]. Do not let AI write reviews, polish a client's review into something they did not say or produce testimonials for marketing pieces.
AI can help with replies. Ask for a short, calm draft response to a negative review that thanks the reviewer, acknowledges the concern without revealing private transaction details and invites them to talk directly. Then rewrite it so it sounds like you, and post it only after a night's sleep.
The rules every post must meet
Social posts, emails and scripts are advertising. The REALTOR Code of Ethics requires a true picture in advertising, marketing and other representations, and its Standard of Practice 12-10 covers internet content and prohibits misleading images[4]. Fair housing rules apply too: the Fair Housing Act prohibits discrimination because of race, color, national origin, religion, sex, familial status and disability[5], so captions should describe homes and amenities, not the people who should live there.
Check your state's advertising rules as well. Many require the brokerage name to appear in every ad, including social posts, and some regulate how team names are shown. Put those requirements in your standing rules so AI includes them by default.
Example: one listing, one week
Example (fictional): an agent listing 22 Example Way pasted her voice guide and the fact sheet into one request and asked for a launch caption, a kitchen detail post, an open house reminder and a short video script. She rejected two drafts that called the street quiet and perfect for families, replaced them with the cul-de-sac location and the fenced yard dimensions, and added a note about the farmers market two blocks away that she had checked herself.
Example result: four posts and a script were ready in under an hour, every fact matched the listing sheet, and the brokerage name appeared in each post as her state requires.
Frequently asked questions
Can I use AI to write my real estate social media posts?
Yes, as drafts built from your voice guide and verified facts. Edit each one, check every fact and confirm your brokerage identification and fair housing wording.
Is it legal to use AI to write reviews or testimonials?
No. The FTC's 2024 rule bans fake reviews and testimonials, including AI-generated ones. Use AI only to help draft your replies to genuine reviews.
Where should market statistics in my posts come from?
From your MLS reports or a named public source, with the date. Never ask AI for market figures; give it the numbers and ask it to explain them.
Do AI-written posts need a disclosure?
Most rules focus on accuracy rather than authorship, but altered images and AI-generated people or voices should be labeled, and you remain responsible for every claim.
References
National Association of REALTORS® (2026). REALTORS® Adopt Technology to Save Time and Improve the Client Experience, NAR Report Finds (2026 REALTORS® Technology Report) (opens in a new tab) ↑
National Association of REALTORS® (2026). REALTORS® Adopt Technology to Save Time and Improve the Client Experience, NAR Report Finds (2026 REALTORS® Technology Report) (opens in a new tab) ↑
Federal Trade Commission (2024). Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials (opens in a new tab) ↑
National Association of REALTORS® (2026). 2026 Code of Ethics & Standards of Practice (opens in a new tab) ↑
U.S. Department of Housing and Urban Development (2026). Housing Discrimination Under the Fair Housing Act (opens in a new tab) ↑
Pew Research Center (2016). Online reviews (Online Shopping and E-Commerce) (opens in a new tab) ↑
Pew Research Center (2016). Online reviews (Online Shopping and E-Commerce) (opens in a new tab) ↑
Federal Trade Commission (2024). Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials (opens in a new tab) ↑
Federal Trade Commission (2024). Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials (opens in a new tab) ↑