Is SEO Still Relevant for Generative AI Search?

TL;DR: Yes — according to Google's own official documentation published July 2026, SEO fundamentals remain the foundation for visibility in Google's generative AI features. But that same guidance also states plainly that several tactics common in AEO/GEO advice — llms.txt files, content "chunking," and even structured data — are not required for Google specifically. The honest picture is more nuanced than either "SEO is dead" or "just do SEO": Google's AI features and platforms like ChatGPT and Perplexity retrieve information differently, and treating them as identical is where most AEO/GEO advice goes wrong.

What Google Actually Says

In July 2026, Google Search Central published official guidance titled "Optimizing your website for generative AI features on Google Search" — a direct answer to a question a lot of marketers have been asking. The guidance opens with a clear position: SEO best practices continue to be relevant because Google's generative AI features, including AI Overviews and AI Mode, are built on Google's existing core Search ranking and quality systems, not a separate system with its own rules.

The mechanism Google describes is retrieval-augmented generation, or RAG: Google's AI features pull from the same Search index that powers regular organic results, then generate a response grounded in what that index already ranks well. A second mechanism, query fan-out, has the AI model generate several related searches behind a single user question to gather broader context before answering.

Google addresses AEO and GEO directly, too: from Google's perspective, optimizing for its generative AI search is optimizing for the search experience — meaning it's still SEO, just applied to a newer surface, not a genuinely separate discipline.

Where It Gets Interesting: What Google Says You Can Skip

The more surprising part of Google's guidance is a section explicitly mythbusting common AEO/GEO advice. For Google Search specifically, the documentation states you do not need:

• llms.txt files or similar "AI text files" — Google states plainly that its systems ignore them, and creating one "will neither harm nor help your site's visibility or rankings in Google Search."
• Content "chunking" — breaking content into small fragments specifically for AI parsing. Google says its systems already understand nuance across a page and surface the relevant part without this.
• Rewriting content in a special way for AI — Google's systems understand synonyms and intent without needing exact-match phrasing.
• Heavy structured data investment specifically for AI features — schema isn't required for generative AI search, though Google still recommends it generally for rich-result eligibility.

Google is also direct about one more thing: seeking "inauthentic mentions" across the web — a tactic some AEO/GEO advice explicitly recommends — doesn't work the way it's often pitched, since Google's ranking systems are built to reward genuinely high-quality content, not citation volume for its own sake.

Why This Doesn't Mean "Ignore AEO and GEO"

Here's the part that matters most, and it's easy to miss if you only read the headline: this guidance describes Google's AI features specifically. Google AI Overviews and AI Mode retrieve from Google's own Search index. ChatGPT, Perplexity, and other AI platforms do not.

ChatGPT's browsing and search functions run substantially on Bing's index, not Google's — a distinct system with different technical requirements. Perplexity performs its own independent, real-time web retrieval, with its own source-selection logic. Neither of these platforms is bound by what Google's documentation says about its own systems, and independent citation research across ChatGPT and Perplexity has repeatedly found real, measurable value from exactly some of the signals Google says it personally ignores — llms.txt among them, along with entity schema and structured FAQ content.

So the honest synthesis is this: Google is not wrong about Google. It's simply describing one platform among several, and treating its guidance as universal advice for "AI search" broadly is the same mistake as treating any single platform's behavior as representative of all of them.

What This Means for Your Strategy

A defensible AI visibility strategy in 2026 does two things at once, without confusing them:

• For Google's AI features: keep doing genuine SEO — helpful, non-commodity content with a real point of view, a clean technical foundation, and a good page experience. Skip the llms.txt and content-chunking tactics specifically as a way to influence Google; they won't hurt, but per Google's own word, they won't help there either.
• For ChatGPT, Perplexity, and other AI platforms: the technical layer still matters, because these systems don't share Google's index or its stated indifference to files like llms.txt. Schema markup, entity signals, and Bing indexing remain genuinely useful levers specifically because these platforms retrieve information differently than Google does.

The mistake worth avoiding isn't doing AEO or GEO work — it's assuming one platform's rules apply everywhere. They don't, and Google just said so about itself, in writing.

Frequently Asked Questions