VercatusAI visibility score
What is an AI visibility score?
An AI visibility score is one number for how often AI engines such as ChatGPT, Perplexity and Google's AI Overviews mention or cite your brand. Tools calculate it in very different ways, so two scores for the same brand rarely agree. Before you trust one, ask what it counts, against whom, and how much it can move by chance.
By Norbert Szabo, founder of Vercatus · Updated
In short
- There are three kinds of score: a share of real answers, a readiness score made of site checks, and an index against your competitors. They measure different things.
- Scores disagree widely. Orbit Media ran one site through 18 audit tools and got anything from 28 to 100.
- A score worth reading is built from real answers, counts mentions and citations apart, compares you with the brands actually named, and shows a range.
- A good score is relative: the gap to the brand that leads each of your topics, per engine, over weeks.
- Vercatus reports Answer Share: the share of answers on your topics that name or cite you, with its range, next to each competitor. Not a composite number.
Three kinds of AI visibility score
Tools use the same name for very different numbers. Before comparing two, work out which kind each one is.
| Kind | What it counts | Useful for | Watch out for |
|---|---|---|---|
| Share of answers | The share of sampled AI answers that mention or cite you | Knowing where you stand in the answers buyers see | How many distinct questions, which engines, and whether mentions and links are counted apart |
| Readiness score | Technical and content checks of your site, added up | A to-do list for access and page structure | It says little about whether engines actually name you |
| Index against competitors | Your mentions compared with a competitor benchmark, such as the median of your top rivals | A quick relative read across a category | Who chose the competitors, and on which prompts |
The IAB's August 2026 framework, Measuring Visibility in the AI Era, sorts the underlying metrics into presence (mention rate, citation rate, share of voice), prominence, portrayal and persuasion. A single score usually blends some of these; knowing which ones is the first step to reading it.
Why scores for the same brand disagree
In October 2026 Orbit Media ran its own site through 18 free AI visibility audit tools and got scores from 28 to 100. Across the 14 tools that publish their method, technical setup made up 46% of the score on average and the actual AI answers 19% (Orbit Media). The IAB notes that more than 20 companies sell AI visibility tools and produce different answers for the same brand.
The causes are easy to name:
- Different questions. Many tools write prompts from your own website rather than from what buyers ask.
- Different engines, read differently. In our 1 October 2026 experiments, ChatGPT's app and its API named only 30% of the same brands and cited 4% of the same sites. A score built on the API describes an engine your buyers never use.
- Different counting. A mention, a link and a first recommendation can each count as "visible", or be weighted differently.
- Too few answers. Two answers to the same question shared 68% of their brands in our tests, so a score from a few repeated prompts moves a lot by chance.
Six questions to ask of any score
- Is it made of real answers? Or of site checks presented as visibility?
- Whose questions? Your buyers' questions, including comparisons and alternatives, or prompts generated from your own pages?
- Which engines, read how? Each engine named, read with web search on, the way people use it.
- Mentioned or cited? The two are counted apart, and you can see which.
- Against whom? The brands the answers actually named, not a list someone picked.
- How sure is it? A range, or the number of answers behind it. Without one, treat a weekly change as noise.
What is a good AI visibility score?
There is no universal benchmark, and anyone who quotes one is guessing. A good score is relative, and the useful comparison is the brand that leads each of your topics.
Say the leader is named in four of every ten answers on "payroll software for small businesses" and you in one. That gap, on that topic, is what to work on, whatever your overall number says. Read it per engine too: being strong in ChatGPT and absent from Google's AI Overviews is a different problem from the reverse. And read it over weeks, because answers drift on their own.
Why Vercatus reports Answer Share instead
Answer Share is the share of AI answers on your topics that name or cite you, shown with its range and next to your competitors on the same scale. We report it this way, rather than as a composite score out of 100, for three reasons.
- It is a count you can check. Every share is made of answers you can open and read.
- It keeps the range. A share from fewer answers shows a wider range, so a move inside it is not reported as progress.
- It sits next to your competitors on one scale, topic by topic and engine by engine, so the gap is visible and so is the brand you are losing to.
Answer Share answers "how visible am I?". The next questions, why engines pick the leader and what to change, are what the rest of Vercatus is for.
Questions about AI visibility scores.
Anything else, ask. We reply within one business day.
Ask a questionWhat is a good AI visibility score?
There is no universal benchmark, and a number on its own says little. Compare yourself with the brand that leads each of your topics: the gap between your share and the leader's, topic by topic and engine by engine, is the figure that tells you where to work.
How is an AI visibility score calculated?
It depends on the tool. Answer-based scores divide the answers that mention or cite you by all the answers sampled. Readiness scores add up technical and content checks of your site. Index scores compare your mentions with your competitors'. Ask which kind you are looking at before you compare two.
Why do different tools give me different AI visibility scores?
They ask different questions, of different engines, collected in different ways, and count different things. Orbit Media ran one site through 18 audit tools and got scores from 28 to 100. Compare a score only with itself over time, or with competitors measured the same way.
Is an AI visibility score the same as Share of Voice?
Close. Share of voice in AI usually means your share of the brand mentions in sampled answers, next to your competitors'. Some tools call that their AI visibility score; others mix in technical checks. Answer Share in Vercatus is a share of answers on your topics that name or cite you.
Can I check my AI visibility score for free?
Vercatus runs a free audit once per domain: real answers from ChatGPT, Google AI Overviews, Google AI Mode and Perplexity on your topics, with your share, the brands named instead and the sites cited. The free checker at /check/ shows whether AI crawlers can read your site.
Where to go from here
- Product
AI visibility tracking
What to measure, and the six checks that make the numbers trustworthy.
- Free
The free AI visibility audit
Your share of real answers from ChatGPT, Google AI Overviews, Google AI Mode and Perplexity, once per domain.
- Compare
The best AI visibility tools
Eleven tools compared by job, engines and published price.
Measure the answers, not a proxy.
Create your account and add your domain: the free audit shows your share of real AI answers, with its range, and who is named instead.
Cancel any time · Data stored in the EU