Google says GEO is still SEO. Here is the half that guidance leaves out

In 2026 Google did something the GEO industry had been waiting for: it published official guidance on optimizing for its generative AI features, and the headline was blunt. From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and is therefore still SEO. The guide, published in Google Search Central in May and reinforced in a June documentation update, says the AI features are rooted in Google's core ranking and quality systems and use retrieval to pull in relevant, current pages. It also tells you what you do not need: llms.txt, content chunking, AI-specific rewriting, special schema.

A lot of practitioners read that as "GEO was hype, it is just SEO, carry on." That reading is half right, and the half it skips is the half you get paid for.

AI visibility, share of voice and sentiment across engines. Google's guidance tells you how to optimize one of these engines. It does not tell you where you stand on any of them. Illustrative data.

Where Google is right

Give Google its due, because this part is genuinely useful. If its generative features retrieve from the same index and lean on the same ranking and quality signals as classic Search, then the fundamentals really do carry over. Clear, accurate, well-structured content that earns trust is what gets retrieved, whether the output is a blue link or a sentence in an AI Overview. And the guidance that you do not need a pile of AI-specific tricks is a real service to anyone being sold llms.txt files and "AI schema" as a silver bullet. For Google's own surface, good SEO is most of GEO.

If a vendor's entire pitch is a special file or a rewriting gimmick, Google just told you to be skeptical. Fair enough.

The half it leaves out

Here is what "it is still SEO" quietly collapses.

1. It is about optimizing, not measuring

Google's guide tells you what to do. It does not tell you whether it worked. SEO taught everyone to measure the outcome: rank, impressions, clicks. The generative layer has its own outcome, and it is not rank. It is whether the answer cites you, in words the buyer reads instead of scrolling. You can follow every piece of Google's advice and still not know if you are the brand the answer names or the one it leaves out. "Do good SEO" is not a measurement. Citation is, and it is a different number.

2. Ranking is not citation

This is the trap inside "it is still SEO." The signals overlap, but they are not the same, and the gap between them is real and common: pages that rank well on Google and are still absent from the AI answer for the same question. You earned the authority and the position, and the engine names someone else. If you assume ranking equals citation because "it is still SEO," you never look for that gap, and it is usually the fastest win you have. We wrote up how to find and close it in the AI Citation Gap.

3. It only speaks for Google

This is the big one. Google's guidance describes Google's features. It says nothing about ChatGPT, Claude or Perplexity, and it cannot, because they are separate systems with their own retrieval and their own favored sources. The evidence that they differ is not subtle: independent reporting through 2026 shows a source like Reddit cited heavily by some engines and barely by others, with the share swinging from a large fraction of citations on one engine to a rounding error on another. A page tuned perfectly to Google's advice can be invisible in a Perplexity answer, and Perplexity might be the tool your customer opens first. "It is still SEO" is true for one engine and untested for the rest.

What to actually do with this

Do not throw out Google's guidance. Use it correctly.

• Treat it as the optimization baseline. The fundamentals it describes are the right way to make a page worth citing. Do them, and skip the gimmicks it debunks.
• Add the measurement it omits. Track whether AI answers actually cite you, as a rate over a rolling window, not a one-time check, because answers are non-deterministic.
• Measure every engine, not just Google. Your visibility on ChatGPT, Claude, Perplexity and Gemini are four different numbers. Watching only Google is watching one screen of four.
• Watch citation, not just rank. Keep your Search Console rankings, and put the AI citation rate next to them. The disagreements are your work list.

The short version: Google told you how to optimize. It did not tell you whether you won, or how you are doing anywhere it does not control. That is not a knock on the guidance. It is just the part that was never Google's job to answer.

How llemmy helps

llemmy measures the layer Google's guidance stops at. It tracks whether AI engines mention and cite your brand across ChatGPT, Claude, Gemini, Perplexity and Google AI, shows the sources each engine leans on, and reports every number with its sample size and a 95% confidence interval so you are reading signal, not a single lucky answer. It reads Google's AI Overviews off the search results and queries the other engines through their official APIs, refreshed daily, and it lines your AI citation up against your Google Search Console rankings so the ranking-versus-citation gaps surface on their own. Google will tell you how to optimize. llemmy tells you whether it worked, on every engine, not just one. Run a free GEO audit or start tracking free.

FAQ

Is GEO the same as SEO?

For optimizing your own pages, largely yes: Google's 2026 guidance says the fundamentals that earn ranking also earn generative citations, because its AI features draw on the same core ranking systems. But GEO and SEO diverge on two things that guidance does not cover: measuring whether you are actually cited rather than just ranked, and every AI engine that is not Google, which retrieves and cites differently.

Does Google say I need llms.txt or special AI schema?

No. Google's 2026 Search Central guidance states that llms.txt, content chunking, AI-specific rewriting and special schema are not needed for its generative AI features, which are rooted in its core Search systems. Treat vendors selling those as the whole answer with caution. The real gap is not special files, it is measuring citation across engines.

If GEO is just SEO, do I still need GEO tools?

You still need to measure the thing SEO tools do not show: whether AI engines mention and cite your brand in their answers, and which sources they use. Google's guidance tells you how to optimize; it does not tell you whether it worked, and it only speaks for Google's own AI features. Measuring citation across ChatGPT, Claude, Gemini, Perplexity and Google AI is the part good SEO alone leaves blank.

Does Google's guidance cover ChatGPT and Perplexity?

No. Google's guidance describes optimizing for Google's own generative features. ChatGPT, Claude and Perplexity are separate systems with their own retrieval and their own favored sources, so a page tuned to Google's advice can still be absent from their answers. That is why cross-engine measurement matters: the engines disagree, and one of them is your customer's default.

By the llemmy team, July 2026. Grounded in Google's 2026 Search Central guidance on optimizing for generative AI features (published May, reinforced in a June documentation update), which frames the work as still SEO and lists llms.txt, chunking, AI rewriting and special schema as not needed. Related reading: The AI Citation Gap, What makes a page AI-readable, and Reddit and the AI answer.