The recommendation layer in AI search
How AI answers pick which brands to name, the three tests content has to pass, and the receipts from a fan site that wins.
AI and modern search
Search isn't ten blue links anymore.
Writing on AI Overviews, ChatGPT search, Perplexity, and how content needs to be structured to survive summarization, extraction, and multi-source synthesis across new surfaces.
7 posts, newest
Start here
How AI search understands your content
What AI-driven search needs to understand a page, why clarity beats completeness, and where most content fails.
Read the articleKeep reading
More on AI and modern search, newest first.
-
-
Content structure for answer-first search
When the answer shows up before the click, the page has to be the next step. The four-part structure that does it.
Also in Content strategy
-
Why AI Overviews appear in Google
What triggers an AI Overview, why visibility and visits stopped moving together, and what content still earns the click.
Also in Search behavior and intent
-
Why informational content stalls
Informational pages rank and get read, then leave the reader informed but unanchored. What they need to do instead.
-
How misaligned intent stalls decisions
Most content stops at awareness. How to build the bridges that move a reader from learning to deciding, with real funnel examples.
-
Why transactional pages fail ready buyers
Ready buyers still leave when a page skips the confirmation step. The three questions a transactional page has to answer first.
How I think about AI and modern search
Modern search systems don't send users to your page to read it. They pull pieces out, reassemble them into answers, and cite you if you're lucky.
Pages with original data, explicit structure, and direct answers get cited. Pages that restate what ten other sites already said get compressed into nothing.
From the Modern Search Systems page on the old version of this site, first published January 2026.