Does ChatGPT recommend small businesses? How to check your own data, and what decides who gets named

"AI only picks big brands" is an untested assumption

Most people who argue this have run one or two prompts. One prompt is an anecdote, not a test.

They also tend to run the wrong kind of prompt. A brand-name query ("tell me about Nike") asks about something the chatbot already knows. Of course the big names show up.

A buying-question query is different. It is the question a real buyer types before they have picked anyone: what to choose, who to hire, what fits their situation. That is where a smaller business can get named, and that is the only kind of query worth testing.

So the question is not "does AI know big brands." It is "does AI name a business like mine when my buyer asks their real question." You can answer that with data.

Step 1: Check your analytics for real ChatGPT referral traffic

Open your analytics. Filter your referrals and traffic sources for chatgpt.com.

Then look past the visit count. Check what those sessions did: signups, leads, form fills, calls booked. Visits show curiosity. Signups and leads show a buyer who arrived through an AI answer and acted.

This is the strongest evidence you can get because it comes from your own site, not from a screenshot of one prompt. If you see chatgpt.com in your referrals, ChatGPT is already sending people to you.

If you see nothing, that is useful too. It tells you where you start, and Step 2 tells you why.

Step 2: Run a 10-question audit with the questions your buyers actually ask

Write 10 questions your buyers actually ask. Pull them from sales calls, support emails, search queries, and the first thing people say when they contact you. Do not write questions that flatter you.

Run each question in a fresh chat. A fresh chat keeps earlier context from steering the answer.

For every question, log three things: who gets named, how they are described, and what is missing about you.

That third column is the point of the audit. It turns a vague feeling ("AI ignores us") into a list of specific gaps you can fix.

If ChatGPT does not name you, test these four things

Treat each of these as a hypothesis to test, not a guarantee. Change one, then re-run the audit and see what moves.

First: does your site state plainly what you do, who it is for, and where you operate? If a person has to dig for it, a machine summarizing your site may miss it too.

Second: does your description match across directories and profiles? Conflicting descriptions make it harder for anyone, human or AI, to say what you are.

Third: do third-party sources mention you? Independent mentions give an answer engine something to draw on beyond your own claims.

Fourth: do you have pages that answer the buyer's question directly? If your 10 audit questions have no matching page on your site, you have no answer to be found.

What referral data proves, and what it does not

Referrals prove it can happen. If chatgpt.com shows up in your sources with signups attached, a chatbot named a real business and a real person acted on it.

They do not promise it will happen for you. Which businesses get named depends on the question, the answer, and what is findable about you. Nobody controls that, and anyone who guarantees a result is guessing.

That is why you run both steps. Referral data tells you what is happening now. The audit tells you why, and what to test next.

Run the audit this week, then run it again in 30 days

This week: check your referrals for chatgpt.com, write your 10 buyer questions, run each in a fresh chat, and log the results.

Then fix the gaps in your third column, starting with whatever your site does not say plainly.

In 30 days, run the same 10 questions again and compare. One run is a snapshot. Two runs show you whether anything you changed made a difference.

P.S. Run the 10-question audit before you read another opinion on this.