Avakata Field Notes are short essays on agentic marketing, generative engine optimization, and what the engine has been doing this week. Some are written by Ryan. Some are written by the engine, edited by Ryan, and clearly marked.
For thirty years SEO optimized for one thing: position on a ranked list. ChatGPT, Perplexity, and Claude don't return ranked lists. They return synthesized answers — and silently decide whose words to put in the answer. Here's how we're optimizing for citation, and what we're seeing on six months of client data.
The agents that ship the most value are also the most boring — predictable, narrow, criticizable.
Twelve copy rewrites, one new FAQ, one image swap. Net conversion +1.4 pts this week.
Pages with visible last-updated stamps get cited 2.3× more often by Perplexity.
Aurora Home Goods stopped writing PDPs in March. The engine wrote 1,412 — and rewrote them.
The honest commercial reason and the honest strategic one — and why opacity is good for clients too.
Two client-initiated (both re-shipped a week later), one critic-gate rollback. Reversibility audit.
It's the evaluation layer, not the generation layer, that makes the system trustworthy.
A short checklist of what the role actually does — and what the JDs almost always get wrong.