How a real estate agent gets recommended by an AI assistant

A client asked me last month why she showed up in Google for her own name but never came up when someone asked ChatGPT for an agent in her area. Fair question, and the answer is that those are two different systems with two different appetites. Ranking is a contest. Being recommended is a verification problem.

What an answer engine is actually doing

When someone asks an assistant for a good agent in a specific neighbourhood, it is not sorting a list of ten links. It is assembling a short answer, usually naming two or three people, and it will only name someone it can corroborate. It is looking for a business it can identify with confidence, facts it can find stated plainly in more than one place, and content that answers the question directly rather than circling it.

That last part is why so much real estate content earns nothing. A page titled "Thinking of Selling? Here's What You Need to Know!" answers no specific question. A page that says what homes in a named ZIP code actually sold for last month, and what that means for a seller, answers several.

The three things that decide it

One entity, stated the same way everywhere. Your name, your brokerage, your phone number, and your service area should be byte-for-byte identical on your site, your Google Business Profile, your Zillow profile, and anywhere else you appear. Not roughly the same. The same. A machine cross-checking three sources that disagree resolves the conflict by choosing someone else.

Structured data that names you as one business. Most agent websites either have no structured data or, worse, have it scattered so that every page declares a separate unnamed business. Done properly, every page on your site points to a single business record with a stable identifier, and that record connects to your Google Business Profile. This is invisible to visitors and decisive to machines.

Answers, written as answers. Not "contact me to learn more about the Sellwood market" but the actual median sale price, the actual days on market, and a plain sentence about what that means if you are buying there. Then mark it up as a question and answer, which is the format these systems lift verbatim.

Why this is more winnable than ranking

You are not going to outrank Zillow for "homes for sale in Portland". Nobody is. Those results belong to the aggregators and they will keep belonging to them.

But an assistant asked to recommend a specific person cannot recommend Zillow. It has to name a human being, and the pool of agents who have made themselves machine-verifiable is still remarkably small. The work is unglamorous and mostly invisible, which is exactly why the field is open.

What I would do first

Fix the profile and the consistency before touching anything clever. Claim and complete the Google Business Profile, make every mention of your details match it exactly, then put real numbers on real pages about the specific areas you actually work. The structured data goes on top of that, and it only helps if the facts underneath it are already straight.

If you want the machinery behind this rather than the theory, it lives on my page about custom software for real estate agents.