On this page
Many teams ask for one number that rates their brand’s AI visibility. A single score is easy to report but hard to act on: it hides which questions, which models and which competitors are behind the result. ShareOfAsk shows the percentages together with their denominators and the answers behind them.
AI Presence
How often does your brand appear in the measured answers?
AI Presence = answers mentioning your brand ÷ measured answers in scope × 100
Failed requests are excluded. A brand counts once per answer, even if its name appears several times. Because an answer can mention several brands, brands’ AI Presence percentages do not need to add up to 100%.
Visibility · AI Presence
Example data
- Answer 1
- Answer 2
- Answer 3
- Answer 4
- Answer 5
- Answer 6
3 answers mention your brand ÷ 6 measured answers × 100
AI Presence = 50%
Question coverage and average first position
- Question coverage
Question coverage = questions with a brand mention ÷ measured questions in scope × 100
A question is covered if at least one included answer mentions your brand. Each question counts once across models and checks. Read how question presence is measured.
- Average first position
Average first position = the average place where your brand first appears in an answer's list of recommendations. Only answers with a ranked list count.
Lower values mean earlier placement. Read it alongside AI Presence, because an early position in a few answers can coexist with low overall visibility.
On the Overview, AI Presence, competitive share of voice and average first position use the latest snapshot. Question coverage uses the available answers from the last 3 months.
Why we break results down by question and model
Models can describe and rank the same brand differently — see when ChatGPT and Claude disagree on your brand. In the Visibility view, you can break down the result by question and model, then use repeated snapshots to see how the pattern changes. That is how you find the specific questions where a closer look could make a difference.
ChatGPT
Claude
Gemini
Grok
Perplexity
Google AI Overviews
Visibility · By question and model
Example data
Tracked question
Which CRM migration partners have strong security practices?
Who can help us migrate our CRM?
–Best CRM consultancy for a mid-size B2B company?
How do I choose a CRM implementation partner?
What the metrics do not claim
- Results represent your configured question set and selected models. They are not weighted by real buyer traffic or question search volume.
- Both Presence and share of voice measure mentions, which can be positive, neutral or negative. A mention is not necessarily a recommendation.
- A failed request is missing data, not evidence that your brand was absent.
- An observed improvement is useful evidence to investigate. It does not establish that one content change caused it.
Full definitions are in our methodology.
See where these metrics appear in the app on the ShareOfAsk product page.
Illustration, not data
FAQ
Does ShareOfAsk give a single AI visibility score?
No. The app does not show one combined visibility score. ShareOfAsk reports documented metrics — AI Presence, competitive share of voice, question coverage and average first position — broken down by question and model, rather than relying on one opaque score.
How is AI share of voice calculated?
Your brand’s counted mentions divided by all counted brand mentions, times 100. Each brand counts once per answer, and the denominator includes tracked competitors and discovered brands admitted to the analysis.
What is the difference between AI Presence and share of voice?
AI Presence shows how often your brand appears in measured answers. Share of voice shows your share of all counted brand mentions, which adds the competitive context.
Does high AI Presence or share of voice mean AI recommends my brand?
Not necessarily. The metrics measure mentions, which can be positive, neutral or negative. Read the original answers before treating a mention as a recommendation.
How can I see my own metrics?
Run a Free Scan: 10 questions across 3 AI models and one action analysis, with results available for 7 days.


