Definition
An AI visibility score is a single number that summarizes how visible a brand is in AI-generated answers across a set of tracked prompts. Vendors calculate it differently. In Promptwatch, the AI visibility score is a 0 to 100 measure of how prominently a brand appears: every analyzed response gets its own score based on placement, share of attention and depth, and the headline number is the average across all responses, with responses that do not mention the brand counted as zero.
The per-response scoring follows a fixed logic. Placement sets the baseline: being the sole recommendation scores 100, a dominant pick among several 95 to 99, first in a ranked list 90 to 94, second place 65 to 80, third 50 to 64, fourth or later 25 to 49, and a brief passing mention 10 to 24. Attention and depth then adjust it. A dedicated heading, repeated mentions or detailed comparison push the score up; being merely listed, hedged or overshadowed by a competitor pulls it down. A response that does not substantively mention the brand scores 0, and so does a line like 'I don't have information about that company'. A name that appears only in a cited URL or source title also counts as 0, which keeps the score about the answer text rather than the sources.
Two design choices make the number readable:
- Absences count. Dividing by all responses, not just the ones that mention you, means there are exactly two ways to raise the score: appear in more answers, or appear more prominently in the answers you are already in.
- Prominence is separate from tone. An answer that spends three paragraphs on problems with your product scores high on visibility, because you dominate it. How favorably you are described is the job of a separate sentiment score, so a high visibility score combined with low sentiment flags exactly the answers worth fixing.
Other tools use the same label for simpler measures, most often the percentage of answers that mention the brand, which this glossary calls mention rate. Scores from different vendors are therefore not comparable, and a score from one prompt set is not comparable with another set either. Compare like with like: the same tool, prompts, platforms and competitor list over time.
The score is most useful next to its neighbors. AI share of voice measures how often you appear relative to competitors, presence only. Read together, low share of voice with a high visibility score means good treatment but too little coverage, while high share of voice with a low score means you appear often as an afterthought. Position in the answer, sentiment and citation share fill in the rest. Because analysis scores every brand in every answer, competitors' scores sit on the same scale, and the gap to the leader shows whether you have a coverage or a prominence problem.
Slice it by platform before acting. A healthy blended score can hide one engine where you barely exist, and that engine's cited sources usually show why. Our deep dive on the AI visibility score explains the calculation with examples.
Examples of AI Visibility Score
- A brand appears in 40 of 100 tracked answers, scoring about 70 on average where it appears; with the 60 absent answers counted as zero, its visibility score is roughly 28.
- A CRM vendor is listed first in a ChatGPT answer but the response then spends most of its length on a rival's features, so that answer scores in the 70s rather than the low 90s its list position would suggest.
- A meal-kit company sees a visibility score of 85 on prompts about delivery problems alongside a sentiment score in the 30s, revealing that AI answers discuss it prominently but mostly in the context of complaints.
- An agency notices a client's blended score of 42 hides a split: 61 on Perplexity and 9 on Google AI Mode, which leads it to investigate the client's Google rankings for the cited sub-queries.
