Free tools › What AI knows about your brand
Free test · no signup · result in about a minute
What AI knows about your brand
People ask AI models about companies like yours. The model answers from what it remembers: sometimes with a competitor's name, sometimes with a city or an offer your site never mentions. Enter your domain and see what six AI models say about you from memory, without web search.
What the result shows
- How many models know your brand. Out of the models that answered: how many describe your company the way your site does, how many only in general terms, and how many have no information about it.
- How many name you in your category. We ask each model which companies it recommends in your category, without your brand in the question, and check whether your name is in the answer.
- Which facts are inconsistent with your site. Country, city, founding year, kind of business, offer. Where a model says something different from your own site, you see both versions side by side.
- How many describe a different company. A model can describe another business under your name, and sound sure about it.
- Counts, not a score. You get numbers of models and facts, not a 0 to 100 grade. The test is separate from the CitePulse visibility score.
Memory is half of the picture. Search is the other half.
This test measures what models remember with web search turned off. That is how a model answers when it is built into an app, a plugin or an agent that does not look anything up.
Assistants such as ChatGPT, Perplexity, Gemini and Google AI Overviews usually search the web first, and their answers can differ from model memory in both directions. CitePulse monitoring asks those assistants the questions your customers ask, on a schedule, and shows whether they name you and who they name instead.
If the test shows a fact inconsistent with your site, the monitoring plans with Fact Integrity check whether assistants that search the web repeat it.
What this test does not tell you: it does not show how many people see these answers, and a model not knowing a brand is not proof of lost sales. It shows what the model says when someone asks.
Frequently asked questions
Which AI models do you ask?
Six model families: GPT, Grok, DeepSeek, Llama, Mistral and Kimi. Each one is asked directly, without web search, so the answer comes from what the model learned in training.
Is this what ChatGPT says about my company?
No. This is the memory of the GPT model without web search. The ChatGPT app usually searches the web before it answers, so its answer can be different. CitePulse monitoring shows what assistants that search the web say.
What does “inconsistent with your site” mean?
The model states a fact, for example a city or a founding year, and your own site states a different one. We do not judge which of the two is true. If your site is out of date, the fix may be on your side.
Why does a result say “not measured”?
When fewer than four of the six models answer, or their answers cannot be compared with your site, we show “not measured” instead of a zero. A zero would look like a finding, and it is not one.
Does this test change my CitePulse score?
No. The test is separate. It does not enter the CitePulse score or the monitoring results.
How often can I test a domain?
A result is kept for 7 days. Testing the same domain again in that time shows the same result.
Do I have to leave my e-mail?
No. The result shows on the page without it. E-mail is asked only if you want the full details, and it is stored as described in the Privacy Policy. You get marketing e-mails only if you tick that box.