How Home Buyers Use ChatGPT to Choose an Agent, and How to Show Up

Most home buyers do not ask ChatGPT to pick their agent. They ask it to explain the process, sort out what they can afford and tell them what to ask. Then a name comes up, from a friend, a portal or the AI itself, and they go back and ask about that name. That second conversation is where agents win or lose the call. Below: what buyers and sellers type, the follow up questions that decide who stays on the list, and a seven week plan for spring 2027.
For how ChatGPT and Google AI find agents in the first place, see How Real Estate Agents Get Recommended by ChatGPT and Google AI. This one is about the buyer's side of the screen.
How many buyers are using AI, and for what?
Most buyers now use AI, and a majority of those used it while buying a home. Very few hired an agent straight from an AI answer. Both facts matter.
- Broad use. Zillow's Consumer Housing Trends Report 2026, published October 14, 2026, found 82% of 2026 buyers use AI in general, and 61% of those users used it while buying a home.
- What they used it for. Among those AI users, Zillow found 74% used it to learn the home buying process, 70% to get home recommendations and 54% for help with an offer.
- Which tools. In a Realtor.com survey of 1,000 US adults buying or selling now or in the past two years (fielded August 7 to 8, 2025), 82% said they use AI for real estate insights, with ChatGPT (67%) and Gemini (54%) leading.
- Still rising. In a Veterans United survey of 859 people fielded June 1 to 17, 2026, 45% of prospective buyers were using AI tools in the homebuying process, up from 37% a year earlier. ChatGPT was used by 33%.
- Agents see it too. In a Q3 2026 survey of 208 agents and brokers by ResiClub and Zoodealio, 78% said clients had used AI tools to research pricing, listings or the transaction in the past year, and 23% said clients used AI to challenge pricing or valuation.
Now the other side. In NAR's 2026 Profile of Home Buyers and Sellers, WAV Group reports that just 1% of buyers found their agent through AI search, and the same 1% of sellers. Zillow found the top ways buyers first found an agent were a real estate website or app (28%) and a referral from friends, family, neighbors or colleagues (19%).
So AI is not replacing referrals. It is where buyers check a referral and prepare for the first call.
What do buyers and sellers actually type?
The prompts move from general to specific. Here is the pattern, grouped by stage. These are examples, not quotes from a dataset.
| Stage | Typical prompts | What the answer needs from you |
|---|---|---|
| Learning | "What does a buyer's agent do now and how do they get paid?" "How do I compare listing agents?" | Nothing yet, but this sets the checklist they will hold you to. |
| Shortlist | "Good buyer's agents in Chandler for relocation buyers?" "Who sells a lot of townhomes in Old Town Scottsdale?" | Public pages and reviews that tie your name to that city and that kind of client. |
| Vetting | "What do reviews say about [your name]?" "Is [your name] a good fit for a first time buyer under $400,000?" "[Agent A] or [Agent B] for selling a condo?" | Specific reviews, a clear price range and property type, and the same facts on every profile. |
| Preparing | "What should I ask a realtor on the first call?" "What should a listing agent's marketing plan include?" "Is the commission this agent quoted normal?" | A plain page on how you work, what you charge and what happens in the first 30 days. |
Sellers follow the same arc: pricing, marketing plans and fees, then a comparison of agents they heard about from neighbors.
What does one of these conversations look like?
Illustration only. This is a composite conversation we wrote to show the pattern. It is not a real transcript, and the agents are placeholders.
Buyer: We are moving to Chandler, Arizona in March for a job in Tempe. Budget is around $550,000. How do we find a good buyer's agent?
AI: Gives general advice: look for local experience in your price range, read recent reviews, ask how they handle remote showings, interview two or three agents and ask them to explain the buyer agreement.
Buyer: Can you name a few agents who help relocation buyers in Chandler?
AI: Names three. Agent A has a Chandler relocation guide on their site and reviews that mention video tours and a closing completed from out of state. Agent B is with a large brokerage and has many reviews, mostly general. Agent C is named with a note that little public detail was found.
Buyer: Which of them explains fees clearly? And who answers fastest?
AI: Says it cannot verify response times. Points to Agent A's page that walks through the buyer agreement in plain language.
Buyer: Write me five questions for a first call with Agent A and Agent B.
Agent C dropped out on the second question because there was nothing to go on. Agent B stayed but did not stand out. Agent A won the first call with things any agent can publish: a relocation page, reviews that describe the work and a fee explanation a stranger can read at 10 pm.
What makes a buyer trust a name, or drop it?
Specifics that match their situation. When Zillow asked 2026 buyers why they chose their agent, the top answers were local expertise (38%), specialized skills (33%) and personality and trust (22%). Only 7% named a technology edge, including AI.
Buyers also know AI can be wrong. In Cotality's AI in Housing 2026 report, as covered by HousingWire (a survey of buyers in the U.S., Canada, the U.K. and Australia), trust in AI to help find a home fell to 16% in 2026, 64% of buyers worried AI may recycle unverified information, and 44% of respondents said they would pay an extra fee for a human expert to verify AI generated housing decisions. In the Realtor.com survey, agents (62%) edged out AI (61%) as a source that made people smarter about the market. That is your opening: buyers use AI to get oriented, then look for a person they can verify.
Names tend to get dropped for a few repeat reasons:
- The AI found little about you, or describes you vaguely.
- Your areas, brokerage or phone number differ from one profile to the next.
- Nothing public mentions the buyer's price range, property type or situation.
- The newest reviews are old, or all say "great agent" with no detail.
- Your site never answers the fee and buyer agreement question.
Speed matters. Zillow found that 54% of buyers who contacted an agent contacted only one before hiring them, and the typical buyer contacted their agent twice before hiring. Text was the most preferred way to communicate with an agent (43%), ahead of phone calls (31%).
A fair housing note on neighborhood questions
Buyers often ask AI which areas are "best for families" or "safest." You cannot control what the AI says, but you control your own pages and replies. The Fair Housing Act, as NAR summarizes it, bars advertising that expresses a preference for or against protected classes, including familial status. Describe homes, commutes, parks, school district boundaries and official data sources, never who lives somewhere or who an area suits. Run neighborhood copy past your broker or attorney.
How do you show up in these conversations?
Make the vetting answers easy to find. The full setup, including how ChatGPT and Google AI source information and how to read your analytics, is in our November 6 guide. For the buyer side, focus on four things:
- Answer the follow up questions in public. One page on how you work with buyers, one for sellers. Cover fees and agreements in plain words your broker has approved, your price ranges, the areas you cover and what the first 30 days look like.
- Make reviews specific. Ask recent clients to mention the city, the kind of home and one thing you handled. Never write reviews for clients or offer anything for one.
- Keep facts identical everywhere. Name, brokerage, phone, areas and specialty should match on your site, your brokerage page, your Google profile and the portals you use.
- Answer fast, by text if they prefer it. More than half of buyers in Zillow's survey contacted only one agent, so the first reply counts.
For filming those answers, see Short Form Video for Real Estate Agents. For the full year, see the 2027 marketing plan for agents.
Your seven week checklist before spring 2027
This plan runs November 19, 2026 to January 8, 2027, so the work is done before buyers start searching in earnest.
| Dates | What to do |
|---|---|
| Nov 19 to 21 | Run ten prompts in ChatGPT, Gemini and Google AI Mode, using the four stages above for your two main cities. Include two that use your name. Save the answers and note who gets named and why. |
| Nov 23 to 25 | List the ten questions buyers and sellers asked you most in 2026. Mark which ones your website answers today. |
| Nov 24 to Dec 5 | Write or rewrite your "how I work with buyers" and "how I work with sellers" pages. Get fee and agreement wording approved by your broker. |
| Dec 1 to 12 | Send review requests to every 2026 client, with a short note asking them to mention the city, the home type and what you helped with. |
| Dec 8 to 19 | Audit your name, brokerage, phone, areas and specialty across every profile. Fix every mismatch. |
| Dec 15 to 23 | Build one page for your strongest niche, such as relocation buyers in one city or condos in one district, with real detail and fair housing safe language. |
| Dec 28 to Jan 2 | Film five short answers to your top buyer questions. Post one a week through spring. |
| Jan 5 to 8 | Rerun the same prompts from November. Compare who is named and what the AI says about you. Set a monthly reminder to repeat. |
Questions we hear
Do home buyers really ask ChatGPT to recommend a real estate agent?
Some do, but it is still a small share of how agents get hired. In NAR's 2026 Profile of Home Buyers and Sellers, as reported by WAV Group, 1% of buyers found their agent through AI search. Far more buyers use AI to learn the process and to check out agents they already heard about.
What do buyers ask AI after it gives them a list of agents?
Usually narrower questions: which agent fits their price range or situation, what reviews say, how the agent explains fees and the buyer agreement, and what to ask on a first call. Your public pages and reviews are what the AI has to work with for those answers.
Why does ChatGPT name a competitor and not me?
Usually because the competitor has more specific, consistent public information that matches the question, such as a page about relocation buyers in that city or reviews that describe similar deals. Answers also change with wording and location, so test several prompts before drawing conclusions.
Should I worry about fair housing when buyers ask AI about neighborhoods?
Yes, in your own content and conversations. Describe homes, amenities, commutes and data sources, not the people who live somewhere or who a place is right for. The Fair Housing Act bars advertising that expresses a preference based on protected classes. Check wording with your broker or attorney.
Want to see what buyers see when they ask AI about agents in your market? Get a free AI Visibility Report at LocalLift Signature, see our real estate marketing services, or talk with us about a plan for spring 2027.


