All articles
Search visibility and AI search · · 6 min read

How to Check if ChatGPT Mentions Your Brand

ChatGPT answers differ every time you ask. Here is how to test whether your store appears, with enough runs to mean something, and how to track it over time.

Spreadsheet and keyboard on a desk
On this page

To check whether ChatGPT mentions your brand, ask it the questions your customers ask, many times each, in fresh sessions, and log how often your store appears. Because ChatGPT writes a new answer on every request, a single check proves nothing. What you need is a rate: for example, named in 4 of 20 runs for a given question.

Why the answers change every time

ChatGPT does not return a fixed list of links. It generates each response, and generation includes randomness, so the same prompt can produce a different set of recommendations from one run to the next. Whether it searches the web for a given question can also change what it says. That is why one manual check is a coin flip rather than a measurement.

How to check manually

  1. Write the prompts. Do not only ask "What is [Brand]?", which tells you whether the model knows you exist. Ask what your customers ask, such as:
    • "What are the best manual espresso makers under $500?"
    • "Recommend online stores that sell high-quality burr grinders."
    • "What is a good alternative to the Comandante C40?"
  2. Remove personal context. Use a temporary or logged-out chat so saved memory and earlier conversations do not shape the result. Note your location and whether search was on.
  3. Run each prompt at least 10 times, each in a new session.
  4. Log every run in a spreadsheet: date, prompt, run number, whether your store or product was named, how it was described and what else was named.
  5. Calculate the rate for each prompt, and check the descriptions for errors. Wrong prices, policies or product details are a finding in themselves.

Three runs per prompt is too few to tell a difference between a rate of 20 percent and 40 percent. Ten runs is a minimum, and small differences should be treated as noise.

Competitors in the same answers

You read the whole answer, not just your own name. If a competitor appears consistently and you do not, look at how it is described and what sources come with it. Often an answer draws on review sites, buying guides and discussions rather than on the brand's own pages. If a competitor is in several roundups and you are in none, that is a likely reason.

Shipping label on a cardboard box
Clear product data helps AI models map your inventory to user queries.

When a tracking tool makes sense

Manual testing gets tedious with a large catalog or many categories. Several SEO suites now include AI tracking that runs prompts on a schedule and reports how often a brand is mentioned. Ahrefs Brand Radar is one example. Check what a tool actually does before you buy: whether it queries the same product you care about, how many runs it samples, and whether it shows the answers so you can read them. Tools that watch a web page for changes cannot submit questions to ChatGPT and are not a substitute.

SEO Karbon's free Shopify SEO audit includes an AI visibility check as one of 71 checks across 10 areas. It asks assistants buying questions about your category and reports how often your store is named, with no login.

How to improve your odds

Nothing here can promise a mention, but these are the levers you control.

  • Be mentioned on credible third-party sites. Reviews, buying guides and industry roundups are the kind of sources AI answers often cite.
  • Keep your product data clear and consistent. State what each product is, who it is for and its specifications, and keep price and availability consistent between your pages, your structured data and your Merchant Center feed. See the product schema guide.
  • Let crawlers in. Check robots.txt for rules about AI crawlers such as GPTBot and OAI-SearchBot if you want to be found.
  • Write plain descriptions. If someone asks for a waterproof running jacket with high visibility and your description is vague marketing copy, a model has little to match.

Next steps

Pick the 10 most important non-branded questions for your store and run them this week, 10 times each. Record the baseline and repeat monthly. For the full method, read how to run an AI visibility audit, and for the bigger picture, what AI search visibility means for a Shopify store and how AI search differs from traditional SEO. If you want help acting on the results, Digital Karbon works with Shopify brands on product data and visibility.

Next step

See how your own store scores

The free audit checks your store against the issues in this article — indexation, duplicate collections, page speed and everything else that holds a Shopify store back — and shows you what to fix first.