Between October 2024 and the first quarter of 2026, the share of homebuyers using AI as their main tool for researching agents jumped from 17% to 67%, according to FlyDragon’s State of AI SEO in Real Estate report.
Over that same quarter, FlyDragon found 8.4% of practicing U.S. agents appeared in any AI-generated answer when people in their market searched for an agent to hire.
The other 91.6% weren’t named in a single answer.
BAM’s audit, “10 Reasons AI Isn’t Recommending You (And How to Fix Each One),” laid out why agents get passed over. Reason #10 was the easiest one to act on: run the prompts your clients are typing and read what comes back. Few agents have done it.
Below you’ll find 12 prompts to run in four groups, plus a guide to what each answer tells you. Every prompt connects to a fix from the original list.
How to Run the Test So the Results Mean Something
ChatGPT and other LLMs can remember details from your past chats. If you’ve used it to write listing descriptions or plan your marketing, it may tailor its answers to you, and you won’t see what a stranger sees.
Set up the test so you’re looking at the same answer a buyer would get:
- Open a Temporary Chat, or use a browser where you’re logged out of the ChatGPT or Claude
- Turn on web search so the answers pull from current information
- Run the same prompts in Perplexity and Google Gemini, since buyers use those tools too
- Screenshot each answer and note the date
- Keep every result in one document you can compare against next month
AI answers change from one session to the next. One run gives you a snapshot. A few runs over several months show you a pattern.
Questions That Test What AI Knows About You
The first five prompts use your name. They tell you whether AI can find you and whether it gets your details right.
Q1. “Should I hire [Your Name] to sell my home?”
This one comes from the original audit. Read the answer the way a seller would. Does it make you sound like someone they’d call?
Q2. “What do you know about [Your Name], a real estate agent in [Your City]?”
Compare the answer against the contact info you use everywhere else. If the brokerage or phone number is wrong, go back to reason #3 and make your details match on every platform.
Q3. “What do reviews say about [Your Name]?”
Look at the dates. If AI pulls reviews from 2023, you look inactive (reason #2). Check whether the reviews it mentions name a neighborhood or an outcome. Generic five-star reviews give AI little to work with.
Q4. “Which neighborhoods does [Your Name] specialize in?”
If AI names the wrong areas or none at all, your hyperlocal content is thin (reason #4) or your Zillow and Realtor.com service areas don’t match your website (reason #9).
Q5. “What sources did you use to answer that?”
Ask this as a follow-up to any of the first four prompts. The answer tells you which platforms AI is reading. If Bing Places or the Better Business Bureau don’t come up, check whether you’ve claimed and completed those profiles (reason #8).
Homes.com and your local Chamber of Commerce listing belong on that checklist too.
Questions Your Future Clients Are Typing
The next five prompts leave your name out. Type them the way a buyer or seller would, using the neighborhoods and suburbs you serve.
During a BAMx Office Hours session with host Krys Benyamein, AI strategist Phil Stringer explained what buyers get back when they ask these kinds of questions.
“When someone asks ChatGPT who the best Realtor is in their area, it doesn’t give them a list of websites. It gives them names.”
Q6. “Who is the best real estate agent in [Neighborhood]?”
Like the first prompt, this one comes from the original audit. If your name isn’t in the answer, you’ve found your starting point.
Q7. “Who’s the best agent for first-time buyers in [Neighborhood]?”
Swap in any client type you work with, such as investors or downsizers. Phil recommends a “reasons to choose” article for each core service you offer, described the way a client would say it. Those articles give AI a reason to connect your name to a specialty.
Q8. “I’m relocating to [Suburb]. Which agent should I call?”
Phil’s fix for this one is a set of local service articles that pair each of your services with each of the main areas you serve. If you’ve written a relocation guide for that suburb, this prompt tells you whether AI has found it.
Q9. “I need to sell my house in [Neighborhood] fast. Who should I hire?”
Sellers with a deadline ask AI questions like this one. An agent who has written about pricing for a quick sale in that area has a better shot at the recommendation.
Q10. “Who are the top real estate agents in [Your City], and how do they compare?”
A buyer who’s interviewing agents might type something like this. The answer tells you which competitors AI names first and what it says about each of them.
Look for press mentions or neighborhood guides they have that you don’t (reason #5). Phil suggests publishing your own comparison articles that line you up against two to four local agents using factual categories like review ratings and response time.
According to Phil, AI tends to reference content that reads like research.
Questions That Build Your Content Plan
The last two prompts turn ChatGPT into a research tool. You’re asking it what you should publish next.
Q11. “What questions do buyers and sellers in [Your Market] ask most when choosing a real estate agent?”
Phil suggested this move in the original audit. He also recommends AnswerThePublic for seeing the real prompts people type into AI search. Each question you find can become a “best choice” article, a 200-300 word post with the question as its title. The same questions belong on your website’s FAQ page and in your Google Business Profile Q&A (reasons #1 and #6).
Q12. “What would a real estate agent in [Neighborhood] need to have online for you to recommend them?”
AI will tell you what it looks for. Check its list against your own online presence and mark anything you’re missing.
What Your Results Are Telling You
Once you’ve run all 12 prompts, match what you found to the fix that goes with it:
- You don’t appear at all. Start with your Google Business Profile and the six directories AI pulls from (reasons #1 and #8).
- AI has your details wrong or mixed up. Fix your name, phone and address on every platform before anything else (reason #3).
- AI describes you with old information. Ask recent clients for detailed reviews and update your Zillow and Realtor.com profiles (reasons #2 and #9).
- AI knows who you are and gives a vague description. Build neighborhood guides and FAQ pages (reasons #4 and #6).
- A competitor comes up in your place. Look at what AI cites for them and go after the same kinds of sources, starting with local press (reason #5).
Pick the one or two results that worry you most and start there.
Phil described this work as a way to raise the odds AI recommends you, with no promise of the top spot in every answer. The monthly test is how you’ll know whether those odds are improving.
Put the Prompt Test on Your Calendar
Pick a day each month and run the same 12 prompts. Compare the new screenshots against last month’s to see whether your fixes are changing the answers.
Phil also recommends refreshing your content every six to 12 months, since AI tends to favor newer information when it builds an answer. Your monthly results will show you which pages need that refresh first.
The top spot is open in a lot of markets:
- In 71% of U.S. metros, no single agent holds more than 15% citation share
- Agents who started this work in 2025 hold 5.7x the citation share of agents who started 12 months later
AI authority builds over time. Each month you wait makes it harder to catch the agents who started before you.
If you want help building the content these prompts point to, Sharran Srivatsaa‘s AI SEO Playbook inside BAMx includes 20 customizable “best choice” blog templates and 10 customizable “reasons to choose” templates. Start your 7-day free trial here to get access.
