91% of Agents Are Invisible in AI Search. Here’s How to Fix That.

Two-thirds of homebuyers now use ChatGPT, Perplexity, or Google AI Overviews to research agents before making contact. 

That’s 67%, up from just 17% eighteen months ago, according to a new report from Intero Digital. 

Here’s the real problem: 91% of agents don’t show up in those results at all.

Sixty-one percent of buyer-side real estate searches now start in an AI engine rather than a traditional search engine. When someone types “best real estate agent in [your neighborhood]” into ChatGPT, they’re getting a direct answer that names specific agents and explains why they’re worth calling.

If your name isn’t in that answer, you’re invisible to the buyers who are ready to move.

Why does AI search work differently than Google?

Traditional SEO was a relevance game. The search engine asked one question: does this page match the query? 

AI search asks something harder. Who does the model believe is genuinely authoritative about this place?

The portals already figured this out. Within a five-month window, Zillow, Redfin, and Realtor.com all launched apps inside ChatGPT, positioning themselves at the top of the AI buyer’s journey before most agents even noticed.

Here’s the other number that should get your attention: 83% of searches that trigger a Google AI Overview result in zero clicks.

Getting cited has replaced getting clicked.

How do you get AI tools to recommend you?

Intero Digital’s report lays out a five-layer framework they call the “local authority stack.” Think of it as the new version of your Google Business Profile strategy, except it goes much deeper:

1. Hyperlocal content. AI models can tell the difference between generic city-level pages and genuinely local knowledge. Neighborhood guides, micro-market reports with real MLS data, school breakdowns for relocating families. Ask yourself: could a competitor copy your page by just swapping the city name? If yes, it’s not local enough.
2. GEO-specific reviews. “Great agent, highly recommend” does almost nothing for AI authority. A review that names the neighborhood, describes the problem you solved, and states the outcome is what AI models actually weigh. Update your post-closing ask to prompt that kind of specificity.
3. Local press mentions. Most agents are sleeping on this one. When a local newspaper or business journal quotes you on market conditions, AI models treat it as an independent endorsement. Build a list of 10 local journalists who cover real estate and position yourself as their go-to source for market commentary.
4. Structured data. Your name, brokerage, phone number, and service area need to be consistent across every platform. AI models build an identity profile from dozens of sources, and fragmented information means fragmented authority. Add RealEstateAgent schema markup to your website if you haven’t already.
5. Community authority. AI models read HOA newsletters, community Facebook posts, local event coverage, and civic organization sites. Agents who show up in those places as community participants, not salespeople, earn a kind of trust that marketing content alone can’t build for them.

Is it too late to start?

No, but the window is closing. 

Agents who began optimizing for AI search in early 2025 now hold 5.7x the citation share of agents who started the same work 12 months later, despite the latter group spending more money. 

AI authority compounds over time, and early movers are becoming the default answer for their markets.

The good news: according to that 91% stat, almost nobody in your market has started yet.

Here’s the play

Most agents won’t do any of this. Right now, that means a mediocre agent who does will get more calls than great agents who don’t. 

1. Pick your top three neighborhoods. 
2. Build one definitive, data-rich guide for each. 
3. Update your review request. 
4. Get your structured data locked down. 
5. Start showing up in local press. 

Do this over the next 90 days, and you’ll be ahead of virtually every agent in your market when a buyer asks AI who to call.

Bottom line: The agents who build local authority now will own the answer when that question gets asked. And it’s getting asked more every day.