How Homebuyers Are Using ChatGPT to Find Agents

The DataHow Many Buyers Are Actually Doing This?

More than you would guess, and the growth curve is the part worth paying attention to.

Veterans United's 2026 homebuying survey found that 59% of respondents had used at least one AI platform to support their homebuying journey. ChatGPT led at 33% (up six points year over year), with Gemini at 20% (also up six) and Meta AI at 10%. That concentration is why ChatGPT specifically tends to be the first surface worth fixing in this industry.

NerdWallet's annual home buyer report, conducted by The Harris Poll across more than 2,000 U.S. adults, found 48% of people planning to buy within twelve months had used or planned to use AI tools during the process — 27% to estimate housing costs, 26% to guide them through the process itself, 25% to visualize renovations.

But the single most striking number comes from outside real estate. BrightLocal's 2026 Local Consumer Review Survey found the share of consumers who had used an AI tool to find a local business recommendation jumped from 6% in 2025 to 45% in 2026. That makes AI the third-most-used local discovery channel — ahead of Yelp, behind only Google and Facebook. Finding a real estate agent sits squarely in that category.

One more data point deserves a caveat, because it is being quoted everywhere. A 2026 report from the AI-SEO firm FlyDragon claims 67% of buyers now use an AI tool as their primary research method before contacting an agent, up from 17% eighteen months earlier. The underlying dataset is large. But FlyDragon sells AI-visibility services to agents, which makes a report titled around agent invisibility a marketing document as much as a research one. Treat that specific figure as directional. The independently-run surveys above tell a consistent enough story without it.

The NuanceAI Did Not Replace the Agent. It Replaced the Referral.

This is the part most panicked LinkedIn posts get wrong. NAR's 2025 Profile of Home Buyers and Sellers — the 44th edition, covering transactions from July 2024 to June 2025 — is unambiguous:

• 88% of buyers worked with a real estate agent or broker.
• 91% of sellers used an agent. For-sale-by-owner fell to 5%, the lowest share ever recorded.
• Buyers ranked agents as their most useful information source at 85% — ahead of every digital tool.
• 40% of buyers found their agent through a referral from a friend, neighbor, or relative.

So the agent's role is not shrinking. What is shifting is the step before the agent — the moment a buyer assembles their mental shortlist. Historically that came from a referral, a Google search, or a portal's agent directory. For a growing share of buyers, an AI assistant now does that job, and does it in one turn instead of twenty browser tabs.

You are not losing the deal to AI. You are losing the shortlist to whoever AI names instead of you.

Realtor.com's own consumer research points the same direction: a large majority of home shoppers report using AI for real estate insight, yet they still rate real estate professionals as the most trusted and accurate source. Buyers are using AI as a research layer that feeds a human decision. That is good news for agents — provided you are inside the research layer.

The PromptsWhat Buyers Actually Type Into ChatGPT

This matters more than the statistics, because it tells you what content to build. AI prompts look nothing like the short keyword strings agents have optimized for over the last decade. They are long, conversational, and loaded with context.

The patterns we see most often in this category:

• Agent discovery with constraints — "find me a real estate agent in [neighborhood] who has experience with first-time buyers"
• Process questions — "what questions should I ask a real estate agent before hiring them?"
• Commission and cost — "how does realtor commission work and can I negotiate it?"
• Neighborhood evaluation — "is [area] a good place to buy right now, and what should I know before making an offer?"
• Head-to-head comparison — "compare [Agent A] and [Agent B] — who has a better track record in [area]?"
• Timing and market — "should I buy now or wait six months in [city]?"

Notice that only the first and fifth mention an agent at all. The rest are the questions buyers ask on the way to needing an agent — and the sources cited in those answers are the brands that end up feeling familiar and credible when the buyer finally does search for representation. That is why answering process questions well is agent marketing, even though it never mentions your listings.

The ProblemWhy AI Skips Most Agents

Four reasons, in the order we usually find them during audits.

The portals still dominate broad queries

Ask for "real estate agents in Chicago" and you will mostly get Zillow, Realtor.com, and Redfin, because AI engines draw from pages that already rank and carry heavy entity signals. You will not beat them there. But ask a neighborhood-level or question-shaped version and the field opens up considerably, because national portals cannot manufacture genuine hyper-local expertise at that granularity. That is where individual agents actually win.

Brokerage template sites give agents nothing to stand on

If your site lives on your brokerage's domain using their template, two things happen. Your content is near-identical to dozens of other agents in your market, which is a duplicate-content problem. And any authority you build accrues to the brokerage, not to you — so it evaporates the day you switch firms. AI engines need a distinct, identifiable entity to cite. A templated subpage is not one.

The content answers keywords, not questions

Pages titled "Austin Homes For Sale | Top Realtor" are optimized for a search box, not for extraction into a generated answer. AI engines lift passages that directly answer a specific question in plain language. A page that opens with a one-sentence answer to "how long does it take to sell a house in Austin?" is far more citable than one stuffed with location keywords. This is the core mechanic behind generative engine optimization, and the published research backs it — keyword stuffing measurably reduces AI visibility.

Nobody else on the web mentions you

AI engines weight third-party corroboration heavily. If local publications, community sites, and industry directories mention several agents in your market and none of them mention you, the model has no reason to name you. This is the slowest fix and the one with the longest compounding payoff.

The FixWhat to Actually Do About It

In the order we would run it:

1. Check your robots.txt for blocked AI crawlers. Look for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended. If any are disallowed — often a leftover from a migration or a CMS default — that engine cannot use your content at all. Five-minute fix, entire category of invisibility removed.
2. Get off the brokerage template. Own your domain. This is the single highest-leverage structural change for an individual agent, and everything else compounds on top of it.
3. Build genuine neighborhood pages. Not thin location doorways — real guides with school data, commute times, price trends, and the kind of specific local knowledge a portal cannot fake. This is the content that wins hyper-local AI answers, and it overlaps heavily with local SEO for AI search.
4. Answer the process questions directly. One page per real buyer question, each opening with a clear one-or-two sentence answer before the supporting detail. Add FAQ and Article schema so engines can parse the structure confidently. This is the mechanic behind answer engine optimization — being the passage the AI lifts, not just a page it passed over.
5. Earn local third-party mentions. Local news, neighborhood associations, community sponsorships, industry directories, podcast appearances. Each one is a corroboration node.
6. Keep reviews active across multiple platforms. BrightLocal found 88% of AI users still fact-check the recommendation before acting, usually via reviews. Being named by AI gets you the shortlist slot; your reviews close it.

If you want the full diagnostic version of this — including how to tell which of these is actually broken on your site — we wrote it up in why your website isn't showing up in AI. For Google's AI answer box specifically, the AI Overviews playbook covers that surface in detail, and how to get cited by ChatGPT, Perplexity & Google AI goes deeper on the content patterns that earn citations. The property-specific version of all of it sits on our AI SEO for real estate page.

Reality CheckTimelines, and What This Doesn't Mean

Real estate SEO has a well-documented four-to-nine-month lag between investment and visible result. AI visibility builds on top of that foundation rather than bypassing it, so expect 45 to 90 days minimum after meaningful changes ship if your rankings are already decent — and add three to six months if they are not. Anyone promising AI citations in two weeks is selling you something.

It also does not mean referrals are dead. Forty percent of buyers still find their agent through someone they know, and that channel is not going anywhere. What has changed is that the buyer who doesn't have a referral — the relocating buyer, the first-timer, the one whose friend's agent retired — now has a fast, confident-sounding alternative. Eighteen months ago that person opened Google. Increasingly, they open ChatGPT.

The agents winning that buyer are not doing anything exotic. They own their domain, they publish real local expertise, they answer the questions buyers actually ask, and they show up in the places AI engines look. That work would have been good advice five years ago too. The difference is that now, skipping it makes you invisible rather than merely lower-ranked.

Worth saying that none of this is unique to property. The same shortlist shift is playing out in every industry where buyers research before they commit — healthcare, legal, financial services, construction. Real estate is just further along than most, because the purchase is large enough that people research it obsessively.