The AI Visibility Index is a periodic measurement of which brands AI assistants actually name and recommend, by category. It is sampled on a fixed suite of non-branded buyer questions, and each edition states the engines it covers and how many times each question was asked — because those two numbers decide how much the result is worth. Every figure carries a confidence interval and a measured run behind it.
A table gives you one number per brand. This is that number across six identical runs of the same suite — same questions, same engine, same day. The spread is variance, not a trend, and it is the whole reason every question is asked more than once.
| Brand | Named in | Share of voice | Run-to-run swing |
|---|
Take it apart yourself: every answer verbatim, the numbers as CSV or as JSON. Method: how every number here is produced.
The whole index as a spreadsheet or as JSON — one row per brand per category, carrying the share, both ends of the 95% interval, the sample size, the engine list and the run id. Cite it, re-analyse it, argue with it. Attribution appreciated, permission not required.
Each edition replaces the last at the same URL, so a saved copy is a snapshot and nothing more. Leave an address and we'll tell you when a new edition lands and which way the numbers moved — the part a downloaded file can never tell you.
Want this for your own brand and category? Start with a free page check, or see whether AI crawlers can reach your site at all.