Does AI recommend your brand?
We ask ChatGPT, Perplexity, Gemini, Qwen, Doubao and DeepSeek five real buyer questions about your category and your brand, then show who gets recommended, whether they know you, and which websites the answers come from. Takes two to three minutes. Free, 2 tests per day.
One test is a snapshot. A GEO audit is the full picture.
We run dozens of buyer questions across ChatGPT, Perplexity, Gemini, Google AI Overviews and the main Chinese assistants, map the sources behind every answer, and tell you which three fixes move the most.
Comparing paid trackers? See the best AI brand monitoring tools of 2026.
What this test measures
Search engines show ten links; AI assistants give one answer and name a handful of brands. If you are not in that answer, the buyer never sees you. This test measures that directly. It sends five questions to six assistants and records what comes back:
| Question type | Example | What we measure |
|---|---|---|
| Category (×3) | “What are the best [category]?”, “Which ones have the best reputation?”, “Which gives the best value?” | Whether your brand is named, in which position, and which competitors are named instead |
| Long-tail (optional) | “I’m [who you sell to] looking for a [category]. Which fit me best?” | Whether you show up for the buyers you actually target |
| Brand (×2) | “What is [brand]? Is it any good?”, “How does [brand] compare with alternatives?” | Whether the assistant knows who you are and describes you correctly |
We also record the websites each assistant cited or retrieved while answering. Those are the pages that shape the answer, and they are where your next mention should come from.
Which AI assistants it tests, and why
The free test queries ChatGPT (OpenAI), Perplexity and Gemini (Google), the assistants overseas buyers use most, plus Qwen (Alibaba), Doubao (ByteDance) and DeepSeek, the largest in China. All are called through their APIs with web search switched on where available, so you see both your English-market and your Chinese-market visibility in one run.
It does not query Google AI Overviews, which has no public API. For that, read our guides on Google AI Overviews, ranking in ChatGPT search and getting cited by Perplexity, or ask for a free GEO audit, which covers it.
API answers are close to, but not identical with, what a logged-in user sees in each app: the apps add personalisation and conversation history. Treat the result as a strong signal, not a pixel-perfect copy.
How to read your result
- Category visibility is the share of the category answers (18, or 24 with a target customer) that mention you. Zero is normal for young brands; above 50% means you are part of the default shortlist.
- Brand recognised counts the brand answers where the assistant described you correctly. “Doesn’t know you” or “Partly” usually means there is too little consistent, third-party information about you online.
- Competitors named shows who the assistants put on the shortlist instead. These are your real AI-search competitors, which are often not the ones you track in classic SEO.
- Sources shows the sites behind the answers. A brand that appears on those sites tends to appear in the answer.
AI answers vary from run to run, so a single test is a snapshot. Run it again in a few weeks after changes, and judge trends rather than one number.
How to improve AI visibility
- Make sure crawlers can read you. Check robots.txt, CDN bot blocking and JavaScript-only content with the free AI crawler checker.
- State the facts plainly on your own site. What you do, for whom, where, at what scale, with proof. Assistants quote clear sentences.
- Earn mentions on the source sites. Industry media, directories, comparison articles and Q&A threads that the assistants already cite. Use genuine, disclosed participation.
- Publish comparison and alternatives pages. “X vs Y” and “best X for Y” pages answer the exact questions buyers ask.
- Re-test monthly. Track the same questions so you can see what moved.
Limits, privacy and fair use
The test is free and limited to two runs per network per day so it stays free for everyone. We store your email, brand, category and the test result so a strategist can follow up once if it is useful. We do not sell or share it. Answers are generated by third-party AI models and may contain mistakes, including about your brand; we show them unedited. Each report gets a private link that works for 30 days; it shows the brand, category and answers but never your email. From a report you can download a PDF, or re-test the same brand and see what changed since that run. If you turn on weekly monitoring, we also keep the contact you leave there and re-run the test once a week until the trial ends or you stop it; you can stop it from the report page at any time.
FAQ
Is the AI visibility test free?
Yes. Each network can run two tests a day, and there is a site-wide daily limit. An email is required so we can keep the test free and follow up if it helps.
Does it check ChatGPT?
Yes. The free test asks ChatGPT, Perplexity and Gemini alongside Qwen, Doubao and DeepSeek. Google AI Overviews has no public API, so it is covered in our GEO audit instead.
Is this an LLM visibility tool or an AI Overview checker?
It is a free LLM visibility checker: it measures how often large language model assistants (ChatGPT, Perplexity, Gemini, Qwen, Doubao and DeepSeek) name your brand for real buyer questions, and which sources they cite. Google AI Overviews has no public API, so the AI Overview check is part of our free GEO audit, where we run your own queries by hand.
Why does my result change if I run it again?
AI assistants sample their answers and search results change daily, so the brands named can shift. Compare several runs or track monthly rather than reading too much into one result.
Why does the AI not know my brand?
Usually because there is little consistent information about you on the websites the assistant reads. A clear About page and genuine third-party coverage on the source sites shown in your result fix this over time.
What category should I enter?
Use the words a buyer would type, such as “CRM for small business” or “cross-border ecommerce agency”, not your internal product name.
A free test covers six assistants and five questions. Want every buyer question in every engine?
Get a free GEO audit