SEO for realtors in 2026 comes down to being one clear entity: one name, one Google Business Profile, one website, and reviews that tie that name to neighbourhoods and deal types. Get that right and you cover the map pack, Google’s AI answers and ChatGPT at once. 45% of US consumers already use AI to find local businesses (BrightLocal, 2026). Agents tell the same story in their own words: a client says they found a competitor “through ChatGPT”, and when the agent asks the same question, their own name isn’t there.
Where AI gets its information about agents
For agent recommendations, AI reads your Google profile and your website, and very little else. OmniEclipse analysed 334 ChatGPT answers about agents in 197 US, UK and UAE markets (April–May 2026; vendor study, small sample):
Source: OmniEclipse Real Estate AI Search Report 2026, 334 answers, Apr–May 2026.
Across all local businesses the mix is broader: websites make up 42% of AI citations, Google Business Profiles 28.6% and Yelp 9.5% (BrightLocal, Sep 2026). For agents the portals feed listing search; who gets recommended as an agent is still decided on your own surfaces.
What AI does with that information
Three things we measured across local industries matter directly for agents (our checks, 3 Oct 2026, small samples):
- The list changes from run to run. In our first round of checks, half of all businesses ChatGPT named appeared in only one of five runs. An agent who was named once, or missed once, has learned very little.
- Each engine has its own list. ChatGPT and Gemini shared about a fifth of their names on average. Gemini leaned almost entirely on Google Maps.
- The card’s rating depends on the country. In the UK and EU, ChatGPT’s business cards matched the Google profile exactly. In the US, close to half of the cards (159 of 341) showed Yelp’s rating and review count instead, often lower than Google’s.
Map rankings don’t transfer automatically either: tracked businesses appeared in 66% of Google Maps results but only 33% of ChatGPT answers and 32% of AI Mode answers (BrightLocal, Sep 2026). A strong map position is a start, not a guarantee.
The entity problem: agent, team or brokerage?
AI has to decide which name your reviews, address and sales belong to. Real estate makes that harder than most industries:
- Solo agent under a brokerage: are your reviews on the brokerage’s profile or your own? Google’s guidelines allow individual practitioners their own profile under conditions. Check them before creating one, and don’t duplicate the office address in a misleading way.
- A team with a brand name: the team, the lead agent and each member can end up as separate, half-complete entities. Decide which name is the public one and use it everywhere.
- An agent who changed brokerages: old reviews stay on the old profile, and old directory pages keep the old office. In our brand checks, AI surfaced outdated addresses and phone numbers that old directories still held.
- The same person, spelled three ways on Zillow, the brokerage roster and your site. AI may treat them as three people.
The fix list: one canonical bio page per agent, the same name and phone everywhere, links from that page to your Google profile, portal profiles and brokerage page (and back), and machine-readable markup (RealEstateAgent with sameAs) for clarity. Across nine businesses we checked by name, six had their own sources (website, Google profile, booking system, directory) disagreeing on at least one fact, and the engines picked a version.
Real estate SEO basics that still decide who gets named
Google Business Profile
Pick the right primary category, define the service areas you work, list services (buyer representation, listing, relocation, property management), and add real photos, answered questions and replies to every review. This one profile is the largest single source in the OmniEclipse data.
Reviews that mention places and deal types
Ask every client after closing. A review that says “sold our condo in South Austin in 12 days” connects you to an area and a deal type; “great agent!” doesn’t. No incentives, no filtering out unhappy clients, no reviews written for clients. 97% of people who use AI sometimes check its recommendation against real reviews (BrightLocal, 2026), so reviews do double duty.
A website AI can read
An agent bio page, buyer and seller service pages, and listing (IDX) pages that don’t bury everything else. Check that your listings are in your sitemap: in the Vienna audit linked below, 574 of 674 listing pages weren’t.
Mentions beyond your own site
Local news quotes, school and community sponsorships, podcasts. Across 75,000 brands, web mentions correlated more strongly with AI visibility than backlinks did (Ahrefs, Dec 2025; correlation, not proof of cause). One agent described on Reddit how their name started appearing after a local news site quoted them on school districts. That is the kind of mention worth pursuing.
Neighbourhood pages that AI can quote
Generic “local content” doesn’t help. A neighbourhood page with your own dated numbers does, because it gives AI and buyers a fact they can’t get elsewhere.
| Block | What it says | Rule |
|---|---|---|
| Your recent deals here | Number of sales, property types, typical days on market | From your own transactions, dated |
| Market figures | Median sold price, inventory | Follow your MLS’s display and attribution rules |
| Places | Schools, transit, parks, commute times | Describe places and facts, never who lives there or who should |
| Who to talk to | The agent who covers the area, with a link to the bio | Same name as on the Google profile |
Fair housing rules bar statements that indicate a preference based on protected characteristics. Keep neighbourhood copy to amenities and facts, and have your broker’s compliance contact review the template.
Getting the numbers right: commission, fees, taxes
AI repeats whatever number it finds about you: an outdated commission statement on an old page, a fee table from a previous year, a directory profile with a superseded office. Commission practices changed after the 2024 NAR settlement, so pre-settlement wording on your site is a liability twice over. Keep one dated “costs of buying and selling with us” page, review it yearly, and remove or redirect old versions.
The machine-readable layer matters as much as the visible one. In our audit of a Vienna real-estate agency the site looked fine to people, but its valuation form was dead, 405 client reviews were labelled as news articles with no ratings attached, and the company name was stored with a broken character. To AI, the agency had no reviews on its own site.
International and English-speaking buyers
The answer changes with the language of the question. When we asked ChatGPT for a dentist in Vienna in German and in English, the two answers shared only 3 of 21 names. The English answers favoured providers that state “English-speaking” on their profile and site. The same logic applies to agents in markets with foreign buyers: list your languages on the Google profile, the bio page and the relevant service pages, and write at least the key pages in the buyers’ language.
How to get recommended by ChatGPT as an agent, and why it names someone else
No one can register you with ChatGPT or guarantee a mention, and portal spend doesn’t carry over. When another agent is named and you aren’t, the reasons are usually concrete:
- Your reviews sit on a brokerage or old-team profile, not on the entity AI matches to your name.
- No page connects your name to the neighbourhood in the question.
- Your profile and site disagree on phone, office or name, so AI can’t be sure you are one agent.
- In the US, your Yelp page is empty or outdated, and that is the card ChatGPT shows.
A common tip is to post on Reddit because “LLMs read it”. For local recommendations our data doesn’t support it: Reddit was cited in 3 of 104 local answers in our first round of checks, all from Perplexity. Your profile, site and local mentions are a better use of the time. Google’s guidance agrees that optimizing for AI search is still SEO (Google Search Central, 2026). More in our guides on getting recommended by ChatGPT and ChatGPT ads vs. organic answers.
What we do for agents and brokerages
- Audit: we ask ChatGPT, Gemini and AI Mode your clients’ questions several times, check what they say about you, and show which agents are named instead and why.
- Fix the entity: Google profile, bio page, portal and directory profiles, Yelp in the US, one consistent identity.
- Publish what AI can quote: neighbourhood pages with your own data, a dated costs page, service pages for buyers and sellers.
- Measure the same way every month, reported as “named in k of n answers” per engine.
Everything is set up in your name, so it stays yours if you leave us or your brokerage. Monthly plans have no minimum term. Deep audits start at $1,200, monitoring at $149 a month and maintenance at $790 a month. See pricing · Local SEO for AI search · How we measure · Vienna case study
How we work
We don’t buy or write reviews, don’t post on Reddit or in agent groups under fake names, and don’t promise placements. Neighbourhood content stays within fair housing and MLS rules; your broker reviews it before it goes live.
FAQ
What is AI SEO for real estate agents?
Making sure AI assistants describe and recommend you correctly. For agents that mostly means one complete, consistent Google Business Profile, a website with a clear bio and neighbourhood pages, and reviews that mention areas and deal types, because those were the sources ChatGPT cited for agents in the largest study so far.
Why does ChatGPT recommend an agent with two years’ experience instead of me?
AI doesn’t see your sales record unless it is published somewhere it reads. It names agents whose Google profile, website and reviews clearly connect a name to an area and a type of deal. A newer agent with a tidy profile and neighbourhood pages can be easier to match than a top producer whose reviews sit on an old brokerage profile.
Does ChatGPT use Google reviews when it recommends agents?
Often, but not only. In our UK and EU checks, ChatGPT’s business cards matched the Google profile exactly. In the US, close to half of the cards we captured showed Yelp’s rating and review count instead. Keep Google strong and check what your Yelp page says.
Do Zillow and Realtor.com profiles matter for AI answers?
For agent recommendations, portals made up 0.5% of the sources ChatGPT cited in a 334-answer study by OmniEclipse, a vendor. Portals still matter for listings and as a place clients double-check you, so keep them consistent, but don’t expect them to get you named.
Can I pay for ChatGPT ads instead?
An ad is not a recommendation. In 197 ChatGPT answers to local questions we collected on 3 October 2026, none contained an ad, and we found no link between buying ads and being named. Like portal leads, ads stop when you stop paying; your profile, site and reviews stay yours.
How can I check whether AI recommends me?
Ask the same question several times in ChatGPT, Gemini and Google AI Mode, for example “who is a good listing agent in [neighbourhood]?”. Answers change between runs, so count how often you are named. Our free audit runs repeated checks and reports “named in k of n”, not a made-up percentage.