RESEARCH
WHAT WE MEASURE,
AND WHAT WE FOUND.
Research here means three kinds of work: how we measure whether AI search systems retrieve, cite and recommend a company, what the platforms actually document about their own crawlers, and what those crawlers actually do on this site, read from the logs. Each piece is dated, carries its sources, labels every claim, and states what it does not claim.
Newest first. A piece that goes out of date is re-verified and re-dated, never quietly edited.
-
Field data · edition 01 · 16 September – 13 October 2026 · published
Crawler Log 01
Every request a verified bot made to this site in 28 days, read from the CDN’s edge logs: which crawlers came, how much of the site each one read, who asked for llms.txt, and how many requests only claimed a crawler’s name.
Open the log -
Reference · version 1.1 · sources read 27 September 2026 · published
AI Search Crawler Reference
Which crawlers and fetchers each AI search provider documents, what each one is for in the provider’s own words, how to verify that a request is genuine — and what is not documented at all. Every row carries a claim class, a source link and a verified date.
Open the reference -
Methodology · version 1.0 · published
AI Visibility Methodology
How we measure whether AI search engines retrieve, cite and recommend a company: a seven-step chain, a seven-level ladder, frozen prompts, five claim classes — and the result the method returned when we pointed it at ourselves.
Open the method
