An AI-Visibility Score is a number on a scale of 0-to-100 calculated by Auditaar which shows how often a website shows up in the answers that AI engines give. It is a different question from search ranking: a page can rank first on Google and still never be quoted by ChatGPT or Google's own AI Overviews. Auditaar computes the score to make that gap measurable.
What the score actually measures
A useful AI-Visibility Score is built from signals that can be observed, not guessed. Auditaar weights five, each drawn from data the audit already collects, and publishes the weights so the number is never a black box:
AI-Visibility Score · the five weighted signals (sum to 100)
- Citations in AI answers (30). How often the site is cited as a source across a large corpus of stored AI answers. A citation is the engine choosing your page as evidence.
- AI Overview citations (25). On the queries the site already ranks for, whether Google's AI Overview names it in the generated answer.
- Share of your pages cited (20). Of the pages an engine could cite, how many it actually has. A site cited only on its homepage has a narrower footprint than one cited across its guides.
- Bing indexation (15). This checks whether the site is in Bing's index. ChatGPT's search draws on Bing, so when the site is indexed there it becomes eligible to appear, though that does not make a citation certain.
- Brand mentions (10). How often the brand is named in AI answers, cited or not.
A real answer, and where the citation lands
The score is abstract until you see what it is counting. When a buyer asks an engine a question in your category, the answer is assembled from sources the engine trusts, and each source is a citation. The whole game is being one of those sources.
Why it is not the same as share of voice
Share of voice, how your citations compare with named competitors, is a tempting headline, but a single dominant competitor makes the percentage arithmetically correct and practically meaningless. Auditaar reports share of voice beside the score as a cross-check, never folded into the number itself. The score answers whether you are present; share of voice answers present relative to whom, and blending the two hides both.
How Auditaar differs from an AI-rank tracker
Most tools that advertise AI Visibility track a prompt or two over time and tell you if your name appeared. Auditaar measures citation depth across a large answer corpus: not just whether you showed up once, but how often engines cite you, which of your pages they quote, and who they cite instead. Being present is only a starting point. How deeply a site gets cited tells you more, and deeper citation is associated with continued citation, though it does not guarantee it.
How to move it
- Publish answer-ready content: a direct answer in the first paragraph, question-shaped headings, and a short FAQ. Engines lift passages, not pages. See how to make a page answer-ready.
- Make the content reachable without JavaScript, so a crawler sees the same words a reader does.
- Earn citations from sources the engines already trust, which is the same work that earns backlinks.
- Keep the site in Bing's index, not only Google's.
AI-Visibility Score FAQ
Is a high search ranking enough for AI Visibility?
No. Ranking and citation are separate. Engines assemble answers from passages they trust, and a top-ranking page with no extractable answer is often skipped. The AI-Visibility Score exists precisely because the two can diverge.
Can the score be gamed?
Not usefully. It is built from whether engines actually cite and name the site across a real corpus of answers, not from self-reported metadata. The way to raise it is to become genuinely more citable.
How often does it change?
AI answer corpora update continuously, so the score moves as engines re-answer and re-cite. It is a sample of what engines say, not a fixed certificate.
Auditaar computes an AI-Visibility Score on every audit that reaches the AI-answer layer, with the weight of each signal published. See the full product, or run an audit on your own site.
