How do I find out if ChatGPT recommends my product?

checked July 14, 2026 · CrediGeo method

Open ChatGPT in a fresh session, ask the question your buyers ask, “best [your category] software,” “top tools for [the job],” “alternatives to [your main competitor]”. And see whether your product is named. Then do it again, four more times, each in a new chat. If you’re named in most of the five answers, the AI is putting you on the shortlist. If you’re named once or not at all, you’re invisible in a channel your buyers are already using. ChatGPT alone reports over 800 million weekly users and still drives roughly 77% of AI-chatbot referral traffic. That’s the whole check. The details below are about doing it right, because one run will mislead you.

How to check it yourself in ten minutes

  1. Use a fresh session. Log out or open a new chat with history and memory off. If you’ve ever discussed your company with ChatGPT, it will politely include you. That’s not what your buyers see.
  2. Ask like a buyer, not like a vendor. “Best [category] software for [use case],” “alternatives to [competitor]”, the plain phrasing a prospect uses. Don’t ask “is [my product] good?”; that’s a different question with a flattering answer.
  3. Repeat it five times, one new chat each. This is the step everyone skips, and it’s the one that matters (see below).
  4. Do it on Google’s AI Mode too. ChatGPT and Google’s AI answer from different sources and disagree, checking only one hides half the picture.
  5. Screenshot every answer and note the date. Answers change; an undated screenshot proves nothing in a month.
  6. Count the names. For each product that appears, tally how many of your five answers included it. Your tally, and where you sit in it, is your actual AI visibility.

Why you’ll get a different answer each time

ChatGPT doesn’t keep a fixed top-ten list of tools in your category. It composes an answer each time from what it reads in the moment, so the named products shift from run to run. Some come up in every answer, others appear once and never again. That spread is why a single check misleads in both directions: one lucky mention convinces an invisible product it’s visible, and one unlucky miss convinces a well-covered product it has a problem. Five runs per engine cost nothing but minutes, and they separate the products the AI consistently names from the noise.

What to do if you’re not named

Don’t buy anything yet, first understand what you’re looking at. ChatGPT’s answers about software lean on the public record: the review platforms and comparison pages it cites (G2-, Capterra-, TrustRadius-type sources), the “best [category] tools” roundups, community threads, and what your own site plainly says about what the product does and for whom. If you’re invisible, it’s almost always because that record is thin, inconsistent, or unreadable, not because of a secret ranking.

A few things you can do yourself: state your category, use case, and buyer in plain text on real pages; claim and complete your profiles on the major review platforms for your category, identically everywhere; and earn honest presence in the roundups and threads buyers actually read. The longer version, with the four most common reasons, is here. If you’d rather see exactly where you stand first, across both engines, run-by-run, with the sources the AI cited. That’s the check we run.

Checked July 14, 2026. AI answers change continuously, re-run the check before acting on any of this.