SEO vs AI Search: What Actually Changes

There's a version of this conversation that's mostly panic. SEO is dead. AI is eating search. Everything you've built for organic discovery is about to stop working.

That version isn't useful. Here's the more honest one.

What actually changes

AI search — Perplexity, ChatGPT search, Google's AI Overviews, Bing Copilot — changes the interface between a user and an answer. Instead of a list of ten blue links, the user gets a synthesized response. The links still exist. The source still matters. But the journey from query to answer gets shorter.

That compression hurts sites that were surviving on click-through from ranking. If your content's entire value was getting someone to your page so they could read the answer, AI search can now deliver the answer without the visit. Traffic drops. Revenue tied to that traffic drops.

What it doesn't hurt: content that earns citation. When an AI system synthesizes an answer, it pulls from sources it trusts. That trust is built the same way traditional search trust was built — through relevance, authority, and specificity. The signal changes slightly in form. The substance doesn't.

What stays the same

A site nobody links to doesn't rank in AI search either. A page with thin content doesn't get cited. A brand with no presence across the web doesn't show up as a trusted source.

I run SmartGringo, a Mexico insurance platform, and Bodegas de Mexico, a wine tourism directory. Both operate in categories where AI search is increasingly common — travel planning, insurance research. What's moved the needle for both isn't gaming new signals. It's the same thing that always mattered: specific, accurate content covering real questions real users have, built on a technically clean foundation.

The sites that are struggling in AI search aren't struggling because the algorithm changed. They're struggling because they were over-optimized for ranking mechanics and under-invested in actual utility.

What you actually need to audit

Entity presence. AI models have an opinion about who the authoritative sources are in any given category. That opinion is formed from training data and reinforced by structured signals — schema, consistent NAP data, presence on trusted directories. If your brand isn't clearly defined as an entity, you're invisible in AI-generated answers even if you rank on page one.

Citation worthiness. Read your content and ask honestly: if an AI were composing an answer to the question this page targets, would it reach for this page? Citation-worthy content is specific, factually grounded, and adds something beyond what's already everywhere.

Technical foundation. Page speed, structured data, clean indexing — none of this is new. AI search doesn't require new infrastructure. It requires infrastructure that actually works, which many sites have been deferring for years.

Build something worth citing

The operators panicking about AI search are usually the ones who built for clicks rather than utility. The operators who built something genuinely useful are watching their traffic mix shift — more AI citation, fewer direct clicks — while overall brand discovery holds steady or grows.

Build something worth citing. The rest is noise.