Mention rate answers the most basic question in AI visibility: when buyers ask the questions that matter to us, does the engine know we exist? It is computed per engine and per period by dividing the answers that name the brand by the total answers sampled.
Reading the metric correctly
- A mention is not an endorsement: track recommendation rate separately to know whether the engine actually advises choosing the brand.
- Segment by prompt category: a brand may dominate informational prompts and be absent from 'best provider' prompts.
- Watch description accuracy alongside frequency: being mentioned with wrong facts can be worse than absence.
- Compare with competitors to compute share of voice in the category.
Because answers vary between sessions, a single test proves little. Mention rate becomes meaningful with repeated sampling on a schedule, which is how it is tracked in this site's AI Visibility Index.