Ask ChatGPT to recommend a real estate agent in your market. Your future clients already are.
The names that come back aren’t random, and nobody paid to be there. AI assistants build shortlists from public data — mostly your Google presence — and the same few agents in each market keep winning, while everyone else is invisible.
I’ve audited Google Business Profiles for agents in more than 160 U.S. and Canadian markets, and the agents AI recommends share a short list of traits. None of them are secrets. Most agents just never do the work.
The shortlist has replaced the search
Somewhere between 60 percent and 70 percent of AI users have already used AI for local search, per Near Media’s research. But here’s the nuance that matters for agents: AI is where the shortlist gets built, and Google is where it gets verified.
In a BrightLocal survey of 1,227 U.S. adults, only 18 percent said they would contact a business straight from an AI answer without checking further — and 71 percent still use Google at some point in the journey.
So the funnel now looks like this: A buyer asks ChatGPT or Gemini for two or three names, then looks those names up on Google before calling. A thin Google presence loses twice. AI never recommends you, and even when it does, the buyer bounces at the verification step.
What AI actually pulls about you
AI models don’t know you. They know what’s written about you.
When a buyer asks for an agent recommendation, these systems retrieve live data.
- Google’s AI Overviews and AI Mode lean on Google’s index and Business Profile data.
- ChatGPT pulls from Bing’s index and the open web, Yelp reviews, etc.
- Gemini draws from Google’s stack, including Maps and Google Business Profiles.
In practice, three sources decide your fate:
- Your Google Business Profile (primary category, description, review count, review text, ratings, photos, Q&A)
- Your website (your bio and your neighborhood and specialty pages)
- Third-party mentions and reviews on industry-specific platforms — brokerage team pages, “best agents in” roundups, local news coverage, podcast appearances and portal profiles.
The 4 factors that decide who gets picked
- Entity clarity. Google has to know exactly who you are, what you do and where you do it. That means the right primary category on your profile, a real description that names your specialties and service areas, and the same name, phone, and brokerage everywhere you appear online. Keyword-stuffing your profile name reads as spam to Google and as noise to a language model. Clean beats clever.
- Review depth, not just rating. Whitespark’s 2026 Local Search Ranking Factors puts high numerical ratings, the quantity of text reviews and review recency all among the top local ranking factors — and AI reads the text. Fifty reviews that just say “great agent” teach the machine nothing. A review that says “helped us buy our first home in Brookside area in Kansas City, Missouri, under $400K” teaches it exactly which questions you should be the answer to.
- Mentions across the independent web. The Whitespark report puts it plainly: In AI search, mentions are the new link. Every time your name appears next to your market and specialty on a site you don’t control — a news story, an association page, a “top agents” roundup, a podcast show-notes page — the model gets another vote of confidence. Agents with zero footprint outside their own website rarely make the answer.
- Consistency across all of it. When your profile says one thing, your website says another, and a portal says a third, the model hedges — and recommends someone whose story lines up. Language models are pattern matchers. Give them one clean, repeated pattern.
How to get recommended by AI: What to fix this quarter
Start with an honest audit: Open your Google Business Profile as if you were a stranger. Fix the primary category, service areas, service descriptions, hyperlocal posts and description first — the description should name the neighborhoods and buyer types you actually work with, in plain language.
Then change how you ask for reviews. Stop asking for “a quick review,” and start asking every closed client to mention what you helped them do and where. You cannot script their words, but you can ask the question that gets a specific answer. Ten specific reviews this quarter outweigh a hundred generic ones.
Finally, go earn two or three mentions. Pitch a comment to a local reporter, get on a local podcast, make sure your association and brokerage pages are current. And check your own results monthly: Ask ChatGPT and Gemini who the best agents in your market are, and ask what they know about you by name. That’s your new ranking report.
The math has changed
The map pack trained agents to think in rankings: 10 blue links, three pins, everyone gets a little traffic. An AI answer names two or three agents, with reasons. There is no Page 2. Either you’re in the answer, or you don’t exist.
The good news is that the inputs are still yours to control. You can’t buy your way into the recommendation. You earn it — with a clean profile, reviews that tell your story in your clients’ words and a name that shows up in more places than your own website.
Bobby Kerr is the founder of LOC8 My Business. Get connected on LinkedIn and Instagram.
