Counted, not claimed. Every number — on a client report or the public board — is what the AI engines actually said to a real buyer question: a timestamped measurement, not an opinion, never a paid placement.
We build ~40 questions your buyers actually ask, tailored to your category, city and segments. The focus is discovery — the moment a buyer asks who to hire, with no brand in the question — because that’s the visibility that wins new customers:
Each question runs repeatedly on every engine your buyers use, because a single AI answer is noise; a pattern is a signal.
Raw, verbatim, timestamped. Every number in your report traces to a stored answer you can open and read.
Your score is the share of discovery answers that name you, with a 95% confidence range. And because AI answers are assembled from sources, the scan also produces the exact domains the engines cite in your market — ranked by frequency, marked where you’re missing:
The cited-source list is the roadmap: listings claimed, real reviews generated, editors pitched, answer-shaped pages published — work aimed at the exact places AI looks.
The identical prompt set re-measured monthly. Movement (or its absence) is visible, attributable, honest.
For each category and city, the questions a prospect actually types. Never “what is [brand]” — a question that already contains a name can’t measure discovery, so it never touches the board:
Every question runs five times per engine — a single AI answer is noise; a pattern is a signal.
Timestamped, verbatim, auditable — every rank traces to receipts.
A brand’s visibility is the share of answers that name it, with a 95% Wilson range. Two brands whose ranges overlap are a statistical tie — shown as overlapping bands, not a false gap.
Ranked by measured visibility. Claiming never changes a number, and the live prompt list is never published or sold — the method is public, the yardstick isn’t gameable.