seo for ai search.
ai answers changed one thing about search for certain: a ranking and a visit have come apart, so a page can be read and acted on without anyone arriving on it. they did not change how these systems decide who is credible enough to quote, and that is still the link graph.
the posts here keep those two apart. the strategic question of whether any of this is still worth doing, the page-level habits that make a passage citable, and how to read your own search console data before blaming the robots for a problem your own site is causing.
what changed, and what didn't
whether seo still pays, and the page-level habits that get a passage quoted.
is seo dead? what the link graph actually shows
the honest answer from 3.90 billion links across 121.1 million domains: rankings still work, clicks are worth less, and the signal AI systems read has not changed.
8 min readanswer engine optimization: what actually makes a page citable
aeo, geo, and llm seo mostly describe the same job. here is the part that is genuinely different from ordinary seo, and what we ship on our own site.
9 min read
diagnosing your own site
before blaming ai for lost traffic, rule out the problems your own site causes.
how to find keyword cannibalization in search console (with a real diagnosis)
the aggregate report cannot show it. query+page pairs can. here is the query that finds it, and 44 cannibalized queries found on our own site.
9 min read
- find link building prospects with ai. using an assistant for the research half of outreach, from the competitor analysis topic.