article topic
sections to work through
last updated
AI platform features change quickly. This article describes what was verified on the date of the last update. Check the provider's current documentation before making an implementation decision.
When a prospective buyer asks ChatGPT about suppliers in your category, the model names specific companies - usually three to six of them. You are either on that list or you are not. This guide shows how to check that yourself in about 30 minutes: which questions to ask, how to avoid distorting the results, and how to record them so they can be compared over time. If the result disappoints, see how to build visibility in ChatGPT systematically.
Before you start: ChatGPT is two different mechanisms
The classic ChatGPT answers from training data, so its knowledge of brands can be many months old. ChatGPT Search - triggered automatically or via the globe icon - searches the live web and cites pages. Your brand can be visible in one mechanism and absent from the other, so check both.
Step 1: turn off personalisation and work in a clean context
ChatGPT remembers earlier conversations and adapts its answers. If you have previously asked about your own company, the model will name it more often and your results will be distorted. Use Temporary Chat, or sign out and ask in a private window. Ask every question in a new thread.
The most common mistake is asking "do you know company X?" in a normal chat. The model will almost always answer yes, politely - that is not a visibility test. A visibility test asks for recommendations in a category without naming the company.
Step 2: ask buying questions, not reputation questions
- Which companies do you recommend for [your service category]?
- I am choosing a supplier of [service] for a B2B company in [industry]. Who should we consider?
- Build a shortlist of five companies offering [category], with a short rationale for each.
- Which [category] company would suit [company size / budget / specific constraint]?
- How do the leading suppliers of [category] differ from one another?
Ask every question twice: once in the classic chat and once with search enabled. Note the differences - this is often the moment teams discover that the model "knows" the brand from training, but cites only competitors when searching live.
Step 3: assess the quality of the description, not just presence
If the brand does appear, look at how it is described. Possible problems include the wrong category, an outdated offer, or a mention with no supporting rationale. Each case should be documented against the complete answer and assessed in the context of the question.
Step 4: record the results and calculate Share of Model
| Question | ChatGPT (training) | ChatGPT Search | Competitor A | How the brand is described |
|---|---|---|---|---|
| Recommended [category] companies | no | no | yes (2x) | - |
| Shortlist of five [category] companies | yes | no | yes | wrong service category |
| Supplier for [industry] | no | yes | yes | no rationale given |
Share of Model is the percentage of answers containing your brand. Measure across a fixed set of at least ten questions and repeat monthly - a single measurement is a snapshot, and only a series shows the trend and the effect of the changes you made.
What to do when ChatGPT does not recommend you
- 1Sharpen the category, target audience and service scope on your home page - models do not recommend companies they cannot classify.
- 2Publish citable content: comparisons, FAQs and case studies with numbers - this is what ChatGPT Search builds answers from.
- 3Work on presence in industry rankings and round-ups, which are common sources for recommendations.
- 4Implement schema.org structured data so the model understands unambiguously what your company is.
ChatGPT is only one AI answer surface. The equivalent instructions for Perplexity are in the guide to checking brand visibility in Perplexity, and the full measurement rules are in the methodology. The scope of a paid Snapshot or audit is matched to the category and the hypothesis.
Assess whether your category is suitable for AI Visibility measurement.
We assess category fit free of charge. Full visibility and competitor measurement is delivered within a Snapshot or a strategic audit.
0 PLN · Fit assessment, without a promise of a full report
Frequently asked questions
Can I check visibility in ChatGPT for free?
Yes. The free version of ChatGPT is enough - it has access both to answers from training data and to web search. The key is the method: a clean context, buying questions and a fixed set of prompts.
Why does ChatGPT name my company sometimes and not others?
Model answers are non-deterministic - the same question can produce different results. That is why a single test settles nothing. Measure across a set of at least ten questions and look at the percentage of appearances (Share of Model), not at individual answers.
How long does it take to improve visibility in ChatGPT?
In ChatGPT Search, the effects of site changes can appear within weeks, because the mechanism reads the live web. In the classic ChatGPT, changes happen more slowly, alongside model data updates. A typical horizon for measurable improvement is two to four months of systematic work.
Is ChatGPT visibility the same as SEO?
No, although the two overlap. SEO optimises positions in Google results, whereas ChatGPT visibility depends on whether the model can classify the brand, cite it and justify a recommendation. Good SEO helps, but it is not sufficient on its own.