How we measure
Every month we ask the AI assistants your customers use the questions they actually ask about businesses like yours. We use each platform’s official tools with web search turned on.
- Many answers, not one. AI gives a different answer almost every time, so we ask every question several times on every platform. This dashboard is built from answers so far.
- From across your service area. Questions are asked as if the customer is in each town we test, because answers change from place to place.
- Ranges, not guesses. Each rate comes with a likely range (a 95% confidence interval). We only call something a real change, or say you lead or trail a competitor, when the ranges don’t overlap.
- The Visibility Score (0–100) adds up six measured parts: how often AI names you (40 points, full marks at the rate the most-named local businesses reach), how often you’re named first (15), how you compare with your strongest competitor (15), whether AI gets your details right (10), whether AI can read your website and your listings are in order (12), and your reviews (8). Its likely range comes from the uncertainty in how often you’re named.
- The plan simulator projects what unfinished steps could add. Website, listing, and review points are recalculated with the score’s own rules; visibility gains are estimates based on your gap to the leaders. They’re projections, and the next measurement shows what really happened.
- Nothing edited. Every number comes from the stored answers, and you can read them yourself on the Overview and Every question tabs.
We ask through each AI company’s official developer access. Results closely track, but aren’t identical to, what people see in each app, which can also depend on their history and exact location. No one can guarantee an AI recommendation; we measure honestly and work to improve your odds.
