Anthony Lee

About

I'm Anthony Lee — an AI researcher, entrepreneur, and the founder of AI+Automation, an AI SEO consultancy built on original research.

I run the GEO Tests research program — the most comprehensive empirical investigation into how AI platforms (ChatGPT, Claude, Perplexity, Google AI Mode) cite web pages. Across six published papers, the program has analyzed over 165,000 unique URLs, 100,000+ citation events, and 19,000+ queries across 14 verticals.

Research

Key findings from the GEO Tests program include:

  • The SEO Floor (pre-registered, April 2026): Only 7.4% of AI citations come from Google's top-3 results. 75.4% come from pages ranking outside Google's top-30. AI citation does not track Google rank.
  • Schema operates at the retrieval layer: AI engines don't parse JSON-LD at fetch time — but the retrieval indexes feeding them do. Schema gates entry into the citation pool, not amplification within it.
  • Answer capsules don't work: Only 43.6% of cited pages have "answer capsules," versus the 72.4% industry claim. The capsule format added zero predictive power in controlled models.
  • Claim-led first sentences get +26.7% within-page citation lift — the strongest single chunk-level signal found. Bullet points, FAQ formatting, and question headings are all null.

All research is published on Zenodo and OSF with open data, pre-registered protocols, and replicable methodology.

AI+Automation

I founded AI+Automation to give agencies the data and tools to sell, deliver, and report on AI SEO. The firm provides AI visibility monitoring, citation audits, content strategy, and competitive intelligence — all grounded in published research, not guesswork.

This Blog

anthonylee991.com is my personal space — separate from the agency. Here I share my inner thoughts on AI platforms, behavioral economics, the philosophy of search, and whatever else I'm thinking about. It's also a living testbed for GEO experiments, informed by the research program and the geo-kb knowledge base.

Connect