What is AI visibility?
AI visibility is whether AI engines like ChatGPT, Claude, Perplexity, and Gemini can find, describe, and recommend your brand when buyers ask about your category. It has two layers: the answer stage, whether you appear and get cited, and the decision stage, whether you are the name the engine picks.
Buyers now research inside AI engines before they reach your site. They describe a problem, ask which options are worth considering, and the engine names a few vendors and drops the rest. AI visibility is the term for how your brand fares in that conversation: whether the engine can reach your content, whether it describes you accurately, and whether it puts you forward when someone asks what to choose.
The two layers of AI visibility
Most tools measure one layer and report it as the whole picture. Separating the two is what makes the term useful, because a brand can do well at the first and disappear at the second.
| Dimension | Answer stage | Decision stage |
|---|---|---|
| What the buyer asks | What are the options for my use case? | Which one should I pick? |
| What it measures | Whether you appear and get cited | Whether the engine names you as the answer |
| Typical metric | Mentions, citations, share of voice | Whether you survive the final question |
| What it tells you | You are retrievable | You are chosen |
The distinction matters because the two behave differently. A broad question lets the engine surface many brands, so appearing is mostly a retrieval problem. When the buyer adds constraints or asks it to pick, the answer has to take a position: prefer one option, recommend another only under a condition, rule some out, or decline to choose. Mention counts do not capture that difference.
"Being visible means the engine can find you. Being chosen means the engine names you as the fit for that buyer. Only the second one shapes the shortlist."
of B2B buyers used generative AI during a recent purchase, mostly to research vendors and products
Source: Gartner, 2026
What AI visibility is not
- Not your Google ranking. An engine can cite a page that ranks poorly and ignore one that ranks first.
- Not website traffic. A recommendation can happen with no click at all, which is why analytics alone will not show it.
- Not brand awareness. Engines describe you from what they can retrieve and verify, not from what your market already believes.
- Not something you control. You influence the inputs the engine reasons over, you do not control the output it produces.
How AI visibility is measured
It is measurable because the inputs are inspectable. Three kinds of signal do the work: deterministic checks where there is a right answer, such as whether AI crawlers can reach your site, whether your entity is declared in structured data, and whether an llms.txt exists. Live measurement where the outcome is what matters, meaning querying the engines and recording what they actually say about you. And stabilized judgment, used only where the question is genuinely subjective.
Radar grades the answer-stage signals across ChatGPT, Claude, Perplexity, and Gemini, and shows the methodology behind each grade so the score is one you can defend rather than one you have to trust. The decision-stage frame above is the direction the product is built toward, not a live measurement today.
Results come from provider APIs. Consumer chat interfaces can answer differently because of personalization, web search, memory, and model choice, so treat an audit as a measurement of the underlying signals rather than a screenshot of one session.
How to check yours
Start with the technical layer, because it gates everything else: if AI crawlers cannot reach your content, nothing downstream can be fixed by writing more of it. A free Radar audit grades crawler access, entity clarity, and structured data, then shows which answer-stage signals are missing before you spend anything.
Sources
Want to take action on this?