What is AI visibility?

AI visibility is how often AI assistants name your business when someone asks a question you would want to be the answer to. Ask ChatGPT for a wedding venue in the Cotswolds and it does not return ten blue links. It names three to five places. Being one of them is AI visibility. It is measured by asking the assistants real questions and counting how often you come up.

Why it is not the same as SEO

A search engine returns a ranked list, and page two still exists. An assistant returns a short answer, usually naming a handful of businesses, and there is no page two. If you are not in the answer you are not in the consideration set at all — and nothing in your analytics will tell you, because the customer never reached your site.

The four main assistants also disagree with each other. The same question put to ChatGPT, Perplexity, Gemini and Claude routinely gets four different shortlists, because they read the web differently and were trained on different data. Checking one of them tells you about that one.

The two halves of it

AI visibility has a cause and an effect, and it helps to measure both.

• Visibility — the effect. When the question is asked, are you named? Measured by asking and counting.
• Readiness — the cause you control. Can an assistant actually read what you do, what you charge and where you are? Measured by auditing your own website the way a machine reads it.

Readiness is the part you can fix this week. Visibility is the part that tells you whether it worked.

What UK businesses actually score

We audited 82 UK businesses' websites across four sectors between August and September 2026. Three gaps were near-universal:

Agentic commerce

0 of 82

Publish llms.txt

~15%

Structured data

0.25–0.32

Structured data — the schema.org markup that tells a machine what a business is — was the single biggest gap in every sector we measured. Without it, an assistant has to infer what you do from prose. The full figures are on the benchmarks pages.

The question decides the score

This is the finding that matters most if you are checking your own visibility. We asked the four assistants about 39 UK businesses, 464 answers in all. Asked a question shaped around what each business actually does, the median business was named in 75% of answers.

Asked a generic category question instead, many of the same businesses vanished. A Guildford live-music venue was named in 0 of 3 answers to "what's the best wedding venue in London" — a question it has no reason to answer — and in 12 of 12 when asked about live music venues in Guildford.

So a low AI-visibility score from a tool asking a generic question may mean nothing at all. The right question is the one your customer types.

Questions people also ask

Is AI visibility the same as GEO or AEO?

Close. GEO (generative engine optimisation) and AEO (answer engine optimisation) are the work of improving it. AI visibility is the thing being improved — whether you get named.

Does ranking well on Google mean AI will recommend me?

Not reliably. Assistants that search the web often start from search results, so it helps, but they then read a handful of pages and name the businesses those pages name. Plenty of well-ranked businesses are never named.

Which AI assistants matter?

ChatGPT, Perplexity, Gemini and Claude account for most of the use. They disagree with each other often enough that checking only one gives you a partial picture.

How often does it change?

Constantly. The same assistant asked the same question disagrees with itself about 4.7% of the time in our measurements, and every model update can move the answer. That is why one check is a snapshot rather than a verdict.