Research that anyone can re-check
Every study we publish carries its full methodology and raw data. Every research crawler we run identifies itself honestly in its user-agent and links back to this page. If one of our bots visited your site, this is who we are and why.
The Machine Time Capsule
100 small-business sites, re-read through machine eyes every 6 months, forever. Snapshot #1 sealed August 2026.
Open the capsule →August 2026Physician, Heal Thyself
The engines’ own doors, measured against their own rules: Bing’s webmaster page is 12 no-JS words, OpenAI and Perplexity 403 plain readers, zero serve llms.txt, one engine practices its preaching.
Read the study → August 2026The Machines Knock on Every Door
Wave 1: the AI accessibility of 43 elite wealth-market firms. Zero block the answer engines, 86% have no written AI policy, elite medicine leads llms.txt. Raw data free.
Read the study → September 2026Dubai Luxury AI Visibility Index 2026
Twelve Dubai luxury developers and brokerages, scored on five machine-checkable signals. Emaar leads 6/6; three big names score zero. Raw data free.
Read the study → August 2026The Accidental AI Lockout
335 small-business sites: 1 in 13 with a robots.txt fully blocks a major AI crawler - most apparently by accident. Full dataset downloadable.
Read the study → July 2026The 20-Agency Re-Test
One month later: llms.txt adoption tripled among 20 leading agencies. Honest FAQ implementation barely moved.
Read the study → June 2026 · Linked by Search Engine RoundtableThe 20-Agency AEO Readiness Study
The original: how technically ready for answer engines are the agencies selling answer-engine services?
Read the study →City-level field data from the Lockout study: Chicago · Phoenix · Columbus · Atlanta · Portland · Nashville · Austin · San Diego
Methods, standards and every claim we make about ourselves: methodology · claims ledger · proof in 90 seconds.