An AI Visibility Score is a 0-100 measure of how prominently a single AI model recommends a brand in a category, based on whether the brand is mentioned, how early it appears, and how consistently it shows up across queries.
A score of 100 means the model names the brand first, essentially every time the category question is asked. A score of 0 means the model never names it. Everything in between reflects the share and prominence of the brand across the period's recorded answers.
Scores are computed per model, so a brand carries four scores, one each for ChatGPT, Claude, Gemini, and Perplexity, plus a consensus score that blends them. The gap between a brand's best and worst model score is its spread, and a wide spread is a finding in itself: the models disagree about you.
Why it matters now
A rank tells you the order; a score tells you the distance. Two brands can sit at #1 and #2 and be nearly tied, or separated by a chasm. Without a score, that difference is invisible, and it is exactly the difference between a race you can win this quarter and one you cannot.
As buying moves into AI answers, the score becomes the honest read on how entrenched a position is. A brand named first almost every time carries a very different score from one that squeaks into first place on a coin-flip, even though both show as #1.
How WhereDoIRank measures it
The score runs 0-100 and is computed per model from that model's recorded answers over the period. Three things move it: whether the brand is mentioned at all, how early it appears when it is, and how consistently it shows up across the category's questions. A brand named first in nearly every answer approaches 100; one the model never names sits at 0.
Each brand therefore carries four model scores, one each for ChatGPT, Claude, Gemini, and Perplexity, plus a consensus score that blends them. We recompute at every close against a fresh run, so the score is a moving measurement, not a fixed grade.
A worked example
Imagine a CRM category. On ChatGPT, a brand is named first in most answers and scores in the high 80s. On Gemini it is named late, or skipped, and scores in the 30s. Its consensus score lands somewhere in between, and its spread is wide.
Read alone, a mid-range consensus score looks like “doing okay.” Read per model, it is a precise instruction: the ChatGPT position is already won and needs defending, while the Gemini gap is where the work is. The single number hid a story the four scores tell plainly.
Common misconceptions
The score is not a quality rating and not our opinion of the brand. It is a measurement of how the models actually answer, nothing more. A score of 100 does not mean a product is best; it means a model recommends it first, essentially every time the category is asked. We are the messenger, not the judge, so the score reports the answers and never grades the brand.
It is also not a fixed grade. Because it is recomputed from a new run at every close, a score is a snapshot of one close, and it is meant to be read as a trend across closes rather than a verdict from any single one.
How it connects
The AI Visibility Score is the per-model number; the consensus rank is the blended ordering those scores produce; the spread is the gap between a brand’s best and worst model score, and a wide spread is a finding in itself. Each close locks the scores in as the reference every movement is then measured against.