How AI assistants decide who to recommend, in plain English

People don't only search anymore. They ask.

"Who's a good orthodontist in Atlanta?" "Is Dr. Smith any good?" "Find me a tax strategist who works with business owners." They type it into ChatGPT, Claude, Gemini, or the AI answer that now sits at the top of Google, and they get a confident paragraph back.

So it's fair to wonder: where does that paragraph come from? And what decides whether you're in it?

They're not inventing it. They're summarizing.

When an AI assistant answers a question about a person or a business, it's working from what it can find and what it has learned from public information: websites, directory listings, official records, news, reviews, articles that mention you.

Think of it like a very fast, very confident assistant who's never met you, reading everything about you in a few seconds and writing a summary.

Which means the summary can only be as good as what's out there.

Why they get people mixed up

If two doctors share a name, or your details are different on every site, the assistant has to guess. Sometimes it guesses wrong and blends you with someone else into one paragraph that sounds completely sure of itself.

That's not the assistant being careless. It's the information being unclear.

The same things that confuse a person skimming search results confuse an AI:

• Your name spelled three different ways.
• An old practice or firm still listed in a few places.
• No single source that clearly says who you are and what you do.

What you actually control

You can't control the assistant. You can control what it's reading.

Fix the official record. Most directories copy from a source record: your NPI listing if you're a physician, your state bar listing if you're a lawyer, your registration if you're an advisor. Update that first and the corrections spread.

Claim and match your profiles. Same name, photo and details on the two or three profiles that show up for you.

Have one clear source of truth. A site of your own that says plainly who you are, what you do, who you do it for, and where. When everything else agrees with it, there's much less room to guess.

Keep it accurate. Every claim on your site should be one you can back up. Assistants repeat what they find, so make sure what they find is right.

What nobody can promise

You'll see people promising to get you "recommended by ChatGPT." Be careful.

Nobody controls what an AI assistant says, any more than anyone controls Google's rankings. What you can do is make it easy to get you right: clear, consistent, accurate information, with one obvious source.

That's not a trick. It's just being findable on purpose.

Try it

Ask an AI assistant about yourself: "Who is [your name], [your profession] in [your city]?"

Read the answer like a stranger would. Is it you? Is it current? Is anyone else mixed in?

Whatever's wrong in that answer is usually wrong somewhere online first. That's where to start.

---

Want a starting point? The 5-Second Name Test walks you through the searches and gives you your own fix list.