I co-built a tool that tracks brand visibility inside AI search. So I have skin in this game, which is exactly why I’ll say the unpopular part out loud: most of what’s being sold as “GEO” is good SEO with the labels changed. Google’s own line in 2026 is that optimizing for AI search is still SEO. They’re mostly right.
The 80% you already know
Clear structure, real authority, content that actually answers the question, clean technical fundamentals. The models pull from the same web the search engines rank. If your SEO is strong, you’re most of the way there, and a lot of “GEO services” are billing you twice for it.
The 20% that’s actually new
Where it does differ: citation behavior. Which sources a model reaches for, how it summarizes you, whether your brand shows up in the answer at all. That’s worth measuring, because you can’t improve what you can’t see. But it’s a layer on top of fundamentals, not a replacement for them, and not a new religion.
The honest move is to do the SEO, then watch the AI-citation layer deliberately. I can set that up with your team.