Google's AI Search Guide Didn't Say What Everyone Thinks It Said

In May 2026, Google published an official optimization guide titled "Optimizing your website for generative AI features on Google Search".

The industry's reaction was immediate and predictable. LinkedIn posts, threads, comments — mostly from SEO specialists across Greece and Europe: "We told you. AEO, GEO, llms.txt — all hype. SEO always wins. We knew it."

Nobody ever said SEO is useless. SEO is the baseline — the foundation. The real question is: is it enough on its own today?

The problem with those reactions was simple: they didn't read the article.

The Google guide covers AI Overviews and AI Mode — products inside Google Search. It says nothing about ChatGPT, Claude, or Perplexity. Because those aren't Google products.

What Google's Article Actually Says

It opens with this sentence in the introduction — explicit, unambiguous:

"This guide is for website owners looking for official best practices from Google Search on how to succeed in generative AI features in Google Search (such as AI Overviews and AI Mode)."

This guide covers exclusively AI Overviews and AI Mode — results inside Google Search.

Google says you don't need llms.txt to appear in AI Overviews? True, 100% correct. AI Overviews are built on the same index and ranking systems that classic Google Search has used for decades. There, yes, SEO wins.

But ChatGPT? Claude? Gemini app? Perplexity? Those aren't mentioned anywhere in the guide. Because they're not part of Google Search. Google wrote a guide for its own products — not for the entire AI ecosystem.

Reading that and concluding "therefore llms.txt is useless everywhere" is like reading FIFA's rulebook and concluding that the NBA doesn't need rules. Different game. Different governing body. The rulebook was never about basketball.

The Critical Distinction That Everyone Missed

There are two entirely different ecosystems — and people mixed them up:

What applies to one doesn't automatically apply to the other. And the audience using ChatGPT or Claude to find information isn't opening Google — they're just asking.

Then Came Lighthouse 13.3 — That Same Week

On May 7, 2026 — days before the Google optimization guide dropped — Google released Lighthouse 13.3, the engine behind PageSpeed Insights used by millions of developers worldwide. They added a new category: "Agentic Browsing".

What does it check? Four things:

• WebMCP Integration — tools registered for AI agents
• Accessibility tree — programmatic names for interactive elements
• Cumulative Layout Shift (CLS) — stability for agents
• llms.txt — valid AI discovery file at the domain root

The official documentation states clearly:

"llms.txt: Checks for the presence of a machine-readable summary at the domain root."

If llms.txt were useless, why would Google add it as an official audit check in PageSpeed Insights? The answer: it's not useless — it just doesn't affect AI Overviews. It's about AI agents and LLMs that crawl the web directly. A completely different layer.

The Most Decisive Argument: The SEO Tool Companies Themselves

If SEO alone is all you need — then why are Semrush, SE Ranking, Ahrefs and dozens of other companies built entirely on SEO investing millions to build LLM monitoring features?

• Semrush launched a full AI Visibility Toolkit monitoring ChatGPT, Perplexity and LLMs — tracking over 100 million LLM prompts globally. Their Site Audit tool now explicitly checks for llms.txt and flags it as a technical issue if missing.
• SE Ranking added a dedicated AI Visibility Tracker monitoring brand mentions and citations in ChatGPT, Gemini, Perplexity and other LLMs — separately from classic SEO tracking.
• Ahrefs launched Brand Radar after a major update in March 2025.

These aren't companies chasing trends. They're companies that live and die by data. If LLM visibility weren't a real problem, they wouldn't invest.

Our Data: 172 Prompts, 5 Languages, One Company

This isn't theory. These are real results from an AI visibility audit we recently completed in partnership with Sustainable Growth.

The company: A B2B professional services firm in Greece with excellent SEO — years of investment in articles, backlinks, technical optimization. Doing "everything right" by classic SEO standards.

The methodology: 172 prompts across 5 languages, tested in ChatGPT, Claude, Gemini app, and Google AI Mode / AI Overviews.

This is the gap that forms when you build only for Google. The company is visible where Google looks — and invisible everywhere else.

That's not an SEO problem. It's an AI visibility problem — and it doesn't get solved with backlinks.

Conclusion: What You Actually Need in 2026

Nobody serious in this industry said SEO is dead. SEO is the baseline — the foundation that must be there. That hasn't changed.

But the baseline is no longer enough to cover the entire AI landscape. The extra mile is now required.

• If your audience searches inside Google Search → Focus on SEO. The Google guide is your playbook.
• If your audience asks ChatGPT, Claude, or Perplexity → You need structured data, clean entity presence, llms.txt, JSON-LD, crawlable content for LLM bots.
• If you want both → You need a strategy that covers both ecosystems.

Reading half an article and pronouncing judgment on an entire discipline isn't expertise. It's confirmation bias with SEO branding.

Taxiarchis Vafeas

AEO & AI Visibility Strategy · Digital Marketing Consultant

Helping businesses become visible across both Google and LLM-native AI ecosystems. Based in Greece, working across Greece and the DACH region. In partnership with Sustainable Growth.