Search is splitting in two. The old half is the ranked page, where organic click-through has already fallen hard as AI summaries absorb the easy questions. The new half is the answer itself, written by ChatGPT, Gemini, Claude, and Perplexity, with a handful of citations behind it. Winning the first half is SEO. Winning the second is generative engine optimization, and I was building GEO test protocols for a biotech client's physician and patient queries before the discipline had a name.
The work divides cleanly. On the SEO side: technical hygiene, intent-mapped content, and the unglamorous structural discipline that still decides who ranks, because the classic index remains the raw material the answer engines read. On the GEO side: making your brand quotable by machines. That means clear claims a model can lift with confidence, structured data that actually parses, entity consistency across every surface a crawler touches, machine-readable editions of your content, and measurement of the referral traffic AI assistants are already sending, which most analytics setups mislabel or miss entirely.
I hold myself to the standard I sell. This site ships an llms.txt edition for AI crawlers, valid structured data on every page, question-and-answer blocks only where a human can see them, and a validation script that refuses to deploy anything broken. The library you see under The Signal is a working demonstration: one hundred posts built to be cited, interlinked, and lifted into answers. When I tell a client what the answer layer rewards, I am describing what I have already built, not what I read in a newsletter.
Who this is for: brands whose customers ask questions before they buy, category leaders who keep losing the AI answer to a wrong or stale source, and regulated advertisers who need visibility without compliance risk. The engagement starts with an audit of how the engines currently describe you, which is frequently the most clarifying document a marketing team has seen in years, and it builds from there into the structural and editorial work that changes the answer.
Ten blue links are not coming back. The brands that get quoted next were structured for it this year.
Ask how the engines describe you
The audit of your current AI answers is the most clarifying document most teams have seen.
Start the conversation