// Journal · agent-readable
Practical writing on shipping agent-native websites — structured for humans, parsed by ChatGPT, Perplexity, and Claude.
A data-backed rebuttal to the GEO/AEO playbook. Schema barely moves citations. Word count doesn't correlate. 44% of citations come from the first 30% of the page. Each engine has a different citation graph. Here's what to do about it.
Every agent architecture reads a different discovery signal. A single agent-card.json isn't enough. Here's the 5-signal matrix — one per architecture — that makes a site legible to every agent shipping in 2026.
Neil Patel is right that prompt volume shouldn't drive GEO strategy — but the alternative isn't 'buyer intent', it's citation probability. On our 22-site cohort, optimising for citation probability drove 4× the conversions.
A 2026 methodology for tracking brand and competitor citations across ChatGPT, Perplexity, and Google AI Overviews — what to measure, how to benchmark, and which tools earn their keep.
How AI-native founders should think about being cited by ChatGPT, Perplexity, and Claude — the channel-by-channel playbook for earning LLM citations.