What GEO actually is, and what it borrows from SEO
Generative engine optimisation is retrieval plus attribution. Most of classic SEO still applies; the parts that change are passage structure, evidence and off-site consensus.
6 min read
Resources
Field notes on getting cited
What we have learned running these audits. No predictions, no framework diagrams — just the checks that move a score and the ones that do not.
robots.txt for AI crawlers: the seven agents worth naming
GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended and Bingbot each do a different job. Blocking the wrong one removes you from answers rather than from training.
5 min read
Answer-first pages: writing passages that survive being quoted
Retrieval lifts passages, not pages. Put the answer in the first two sentences, make headings question-shaped, and keep every paragraph independently true.
5 min read
llms.txt, honestly: what it does and does not do
llms.txt is a proposed convention, not a standard, and no major assistant guarantees it reads yours. It is still worth 30 minutes of work. Here is why we weight it at 8%.
4 min read
The structured data types that change how models describe you
Organization, FAQPage, Article and Product carry the most weight for AI answers. Missing sameAs and dateModified are the two gaps we find most often.
6 min read
Measuring AI share of voice without fooling yourself
Single readings are noise. Fix the prompt set, run it on a schedule, record mention rate and sentiment, and expect movement that has nothing to do with you.
5 min read