AI Named the Best Agents in 100 Cities. 91.5% of You Were Never Mentioned Once.

August 25, 2026

Stop reading for a second and do this.

Open ChatGPT on your phone. Type: who is the best real estate agent in [your city]. Do not add anything. Do not clean it up. Hit enter.

Now look at the answer.

If your name is not there, you are in the 91.5 percent.

The study

Local Falcon, a platform that tracks brand visibility across AI answer engines, ran 37,500 AI searches in August 2026. One hundred largest US cities. Three phrases: “best realtor near me,” “best real estate agent near me,” “top rated real estate agent near me.” Twenty-five grid points per city across a two-mile radius. Five platforms: ChatGPT, Google AI Overviews, Google AI Mode, Gemini, and Grok.

Then they cross-referenced every agent named against RealTrends Verified production rankings.

Among established, website-publishing agents and teams, 91.5 percent were never cited once.

Narrow it to the agents who actually sell the most homes in their market, and only 21.2 percent got named by any platform at all.

Read that again. Being the top producer in your city does not get you named. It is not even close to a reliable predictor.

Production is invisible

Local Falcon’s CEO David Hunter put it about as plainly as it can be put.

"An agent can be number one in their city by volume and still not exist as far as ChatGPT is concerned. 
The models aren't reading MLS data or production rankings; they're reading pages."

That is the whole problem in two sentences.

Your MLS statistics are not in the training data. Your production ranking is not in the training data. Your closing volume, your years in the business, your client outcomes, none of it is legible to the model unless it appears on a page the model can read.

The machine is not evaluating who is good. It is summarizing what has been published.

Where the citations actually come from

This is the part that changes what you do on Monday.

Google Business Profile accounts for 36.6 percent of all agent citations. Zillow accounts for 15.9 percent. Your own website is roughly one in nine.

But the platform-by-platform split matters more than the average, because the platforms behave nothing alike.

ChatGPT pulls 39.8 percent of its citations from Zillow alone. It leans harder on portals than any other model.

Google AI Mode pulls 72.5 percent from Google Business Profile listings. It barely touches the open web.

Gemini pulls more from agents’ own websites than any other platform, though portals still lead.

So the honest answer to “where should I focus” is another question. Which model are your buyers using? Because the correct move on ChatGPT is different from the correct move on Google AI Mode, and most agents are optimizing for a platform their clients do not use.

Here is the twist inside the Zillow number. 86.9 percent of Zillow citations point to a citywide agent directory page, not to an individual profile. The AI is not using Zillow to say “here is who is good.” It is using Zillow to say “here is where to look.”

Your Zillow profile is not getting you cited. Zillow’s directory page is getting Zillow cited.

The brokerage numbers nobody wants to see

Compass: 1,510 citations across 86 of 100 cities.

Coldwell Banker: 567. REMAX: 454. Compass beat both combined.

Keller Williams, the largest brokerage in the country by agent count, came in eighth. Seventy-six citations across 35 cities.

And here is the one that should end the “we are too small” excuse. Shorewest, a Wisconsin-only brand, pulled 85 citations, every one of them out of Milwaukee. A single-state brokerage operating in one metro beat the biggest brokerage in America nationally.

Local density of cited pages beats national scale. Headcount does not translate. Published local authority does.

Now the uncomfortable part

Here is what the study found about the agents who did get named.

Eighteen of the twenty-five most-cited agent websites carry some version of a “best real estate agents in [city]” page. A page the agent wrote. About their own market. Frequently ranking themselves at the top of it.

So the single most reliable tactic for getting named by AI right now is publishing a page declaring yourself the best agent in town.

Now hold that next to this.

A tagline like “The No. 1 Real Estate Agent For You” is the textbook example real estate ethics trainers reach for when they teach Article 12. It is fine if you actually are number one and can prove it. It is a problem if you cannot.

Article 12 requires REALTORS to be honest and truthful in their communications and to present a true picture in their advertising. An unsubstantiated superiority claim is exactly the thing it exists to catch.

So the winning move and the reportable move are the same move.

That is not a gotcha. That is a genuine, unresolved collision between how these models work and how our profession is governed, and almost nobody is talking about it.

How to get cited without getting called in front of a grievance committee

You do not have to choose between invisible and indefensible. There is a version of the local authority page that gets cited and survives a complaint. It has five rules.

One. State your criteria before you state your ranking. Transaction volume. Homes closed. Years in market. Verified review count. Whatever your standard is, publish the standard first, then apply it. A ranking without disclosed criteria is an opinion wearing a number.

Two. Cite a verifiable source for every claim. RealTrends Verified. MLS statistics. Your board’s published production numbers. If you cannot point to where the number came from, do not publish the number.

Three. Name other agents. This is the one people resist and it is the most important. A best-of page with one name on it is an advertisement. A best-of page with eight names on it is a resource. Resources get cited more, and resources do not trip Article 12.

Four. Never make an unverifiable superiority claim about yourself. If you cannot substantiate it with a third-party source, it does not go on the page. Not in the headline, not in the meta description, not in the alt text.

Five. Date it and update it. A page from 2024 signals to the model that your information is stale. Put the date on it and refresh it quarterly.

Here is the punchline. Based on how these models weigh third-party consensus, the compliant version is probably more citable than the self-serving one. Models are getting better at discounting self-published superiority claims, and they are getting better fast. The agent who builds the substantiated version is not just safer. They are building the thing that still works in eighteen months.

Who paid for this research

I am going to do something here that most people writing about this study will not do.

Local Falcon sells AI visibility tracking. This study is excellent, the methodology is disclosed, and Inman independently cross-referenced the results against RealTrends. It is the best data available on this question right now. It is also research published by a company that sells the solution to the problem the research identifies.

That does not make it wrong. It makes it worth naming.

The same applies to the free scanning tools showing up in this space. There is a free AI-visibility scanner from Roomvu called Found. It takes about sixty seconds, requires no login, and returns a real scorecard. I have no problem recommending it. It is also a lead magnet built by a company that sells marketing software to agents.

Use the free tool. Know why it is free.

What to do this week

Run the five-platform check on your own name. ChatGPT, Gemini, Google AI Mode, Google AI Overviews, Grok. Same query on each. Write down what comes back. You will find the platforms barely agree with each other, and that disagreement tells you where your gaps are.

Fix your Google Business Profile. It is 36.6 percent of all citations and 72.5 percent of Google AI Mode’s. It is free. Most agents have not touched it in two years. Categories, service areas, description, hours, photos, the question and answer section. This is the highest-return hour you will spend on your marketing this month.

Draft the honest authority page. Criteria first. Sources cited. Other agents named. Dated. If you will not write it, that is a legitimate thing to hire out, but it needs to exist.

Put a reminder in your phone for thirty days. Same query, same phone. If the answer changed, the work is working. If it did not, you know which of the three things above you skipped.

The line worth remembering

The machines are not reading the MLS. They are reading pages.

And in most markets, the only page about the best agent in town is the one nobody bothered to write.

Production is invisible. Publishing is not.

Want help getting found?

Reading about AI visibility and actually fixing yours are two different sports.

Done for you: I build custom AI chatbots and visibility systems for agents at Big Brain Chatbots.

Learn it yourself: come into AI-Cademy, where we work through this live every week.

Ready to go all in? The 6-week AI-Cademy Bootcamp, where week one homework is running your own five-platform check and fixing what it finds.

Not ready yet? Start with my free guides.

One question before you go. Did your name come up when you ran the search at the top of this article? Either answer tells you exactly what to do next.

JMan is a national real estate AI keynote speaker and strategist who makes the complicated stuff simple and usable. Learn more at jmanai.com.