What is an AI visibility score?
An AI visibility score is one number that sums up how often, and how prominently, AI engines name your brand across a fixed set of questions. Most tools roll mention rate, position and share of voice into one figure. That is convenient for a dashboard and easy to misread, because two scores built from the same answers can differ by a wide margin depending on which questions, brands and engines went in.
What goes into an AI visibility score?
A useful score is built from measures you can see separately, on a fixed set of buyer questions and engines:
- Mention rate: the share of answers that name you, with an interval.
- Position: where you sit on the shortlist when named.
- Share of voice: your share of all brand mentions, against a stated set of brands.
- Citation share: how many of the engines' sources are your pages.
- Sentiment: whether the answer speaks well of you.
Why can two scores for the same brand disagree?
Because the choices behind a score move it more than the answers do. In one benchmark, the category leader's mention rate was 59% across all 220 answers and 40% on the 16 questions that named no brand (the post). Its share of voice moved from 22.8% to 34% when two broad SEO suites were dropped from the list of brands counted. Not one answer changed.
Missing data moves it too. A tracker that never loaded Google's AI Overviews recorded no Overview on 96% of searches; counted as misses, those would have read as a brand invisible on Google (the post).
What does a visibility score look like in a real example?
In the public Tesla example (Oct 4, 2026), Tesla's mention rate on 25 buyer questions was 53% across four engines, with a 95% range of about 43% to 62%. A single blended score would hide that ChatGPT and Google AI Overviews named Tesla in 64% of answers and Perplexity in 40%.
It would also hide the topic split: Google AI Overviews named Tesla in 4 of 5 solar panel installation answers, Claude in none. And had the five brand questions been pooled in, 19 of 20 answers naming Tesla would have lifted the score for no reason a buyer would care about.
What should you ask before trusting a score?
Ask any vendor, including us, five questions:
- Which engines are asked, and are they the same questions in the same words?
- How many answers is each figure based on, and is there an interval beside it?
- Are questions that name you or a competitor kept out of the headline?
- Is a search with no answer left out, or counted as a miss?
- Can you open every answer and cited page behind the number?
How does Proofsource score AI visibility?
Proofsource does not hide its parts inside one index: it shows mention rate, average position, share of voice and citations per engine, each with a 95% Wilson interval, on buyer questions kept apart from brand questions. See how an AI visibility score is built and the AI Visibility feature page.
Common questions about AI visibility score
What is a good AI visibility score?
Only a comparison makes it good or bad: your score against the brand named most often, on the same questions, engines and period, with intervals that do not overlap. A score with nothing beside it cannot be judged.
Can I compare scores from two different tools?
Rarely. Tools differ in engines, questions, brand sets and how they count missing answers, so the same brand can score very differently. Compare within one tool over time.
How often does an AI visibility score change?
Single answers churn daily, but a fixed set of questions moves much less. Judge a change by whether it holds across runs and clears the interval, not by one reading.
Should brand questions be part of the score?
No. They name you almost every time and inflate the score. Read them for accuracy instead.
Related terms
- AI visibility
AI visibility is how often AI assistants such as ChatGPT, Perplexity, Claude and Google AI Overviews name your brand when people ask about your category.
- Mention rate
Mention rate is the share of AI answers to a question that name your brand.
- Position in AI answers
Position in an AI answer is where your brand sits among the brands the answer names: first, third or last on its shortlist.
- AI share of voice
AI share of voice is your brand's share of all the brand mentions in AI answers to a set of questions.
Worked example: Tesla, public AI answers, Oct 4, 2026: Proofsource's first scan of Tesla, 25 buyer questions and 5 brand questions, one answer per question on each of four engines. A small sample from one brand, shown as an example, not a benchmark.
See where AI leaves you out, and what to fix.
See which buyer questions name your brand, who gets named instead, and what to fix. 200 answers. 25 Prompts/Question. Top AI engines. 2 days. Free. No card.
Google AI OverviewsGoogle AI Mode
