Google's Guide to Optimizing for Generative AI Search: SEO Still Applies, llms.txt Not Needed

Business owners are being offered a growing list of services with new names: answer engine optimization, generative engine optimization, llms.txt files, content "chunking" for AI. The pitch is that ranking in Google's AI answers needs a different playbook from ordinary search. Google has now published its own position, and it is more conservative than much of that market suggests.

On May 15, 2026, Google added a new document to its Search documentation: a guide to optimizing for generative AI features on Google Search. The documentation changelog describes it as covering non-commodity content, local, shopping, image and video content, mythbusting of common "AEO/GEO" misconceptions, and initial guidance on AI agents. Google has revised it since; the page showed a last-updated date of July 10, 2026 when we read it.

This article summarizes what the guide says and what it means if you are budgeting for a new website or for SEO services. It reports Google's statements about Google Search only. Other AI products run by other companies work in their own ways and are outside this guide.

The headline: Google treats this as SEO

The guide opens by asking whether SEO is still relevant for generative AI search, and answers that it is. It acknowledges the terms AEO and GEO, then states that from Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, "and thus still SEO."

It gives a short technical reason. Google says its AI features rely on retrieval-augmented generation, also called grounding, in which relevant pages are retrieved from the search index and the response is generated with links to those sources. It also describes query fan-out, where the system issues several related searches at once to gather more results.

What Google says does help

The guide groups its positive advice into three areas.

Content that is not a commodity

Google draws a line between commodity content, which repeats common knowledge and could have been written by anyone, and content based on first-hand experience and expertise. Its example of the commodity kind is a generic list of tips for first-time homebuyers. It recommends unique points of view, clear organization with headings and sections, and supporting images and video. It also warns against producing many variations of a page mainly to capture rankings, which it ties to its scaled content abuse policy.

A sound technical base

To appear in generative AI features, the guide says a page must be indexed and eligible to be shown with a snippet. Beyond that it lists familiar items: crawlable, publicly accessible content, semantic HTML, following JavaScript SEO practices if the site depends on a JavaScript framework, good page experience on mobile, and fewer duplicate pages.

Accurate business and product data

For local businesses and online stores, Google points to Business Profiles and Merchant Center feeds as the way details and products reach its AI responses.

What Google says you do not need to do

The mythbusting section is the part most likely to save a business money. According to the guide:

• llms.txt and AI-specific files: you do not need new machine-readable files, AI text files, special markup or Markdown versions of pages to appear in Google Search.
• Chunking: there is no need to break content into small pieces for AI. Google says its systems can understand several topics on one page and show the relevant part.
• Rewriting for AI: exact phrasing is not required, because the systems understand synonyms and general meaning.
• Manufactured mentions: chasing inauthentic mentions across blogs, forums and videos is discouraged.
• Extra structured data: Google states that structured data "isn't required for generative AI search" and that there is no special schema.org markup to add, while noting it remains useful for rich results.

On llms.txt in particular, Google added a clarifying note on June 15, 2026. The changelog entry says these files are not needed for Google Search and will not affect visibility or rankings positively or negatively, and that it is fine to maintain them for other services that use them.

Measurement and third-party tools

The guide tells site owners to track visibility with the Generative AI performance report in Search Console. It also contains a blunt sentence about the tools market: "No third-party tool has access to our internal ranking or AI systems." It advises caution about tools or services that claim otherwise or that guarantee rankings.

AI agents: early guidance

The guide includes a section on what it calls agentic experiences, where an AI agent carries out a task such as making a booking or comparing products. Google says such agents may work with a site through its visual rendering, its page structure and its accessibility tree, and it refers readers to separate best practices for agent-friendly websites and to emerging protocols such as the Universal Commerce Protocol. The changelog calls this "initial guidance", so treat it as a direction of travel, not a settled set of requirements.

A related update on AI-written content

On October 1, 2026, Google also updated its separate page on using generative AI content on your website; the changelog says the revision adds information from its Search Quality Raters guidelines. That page says generative AI can be useful for research and for structuring original content, but that using it to produce many pages without adding value for users may violate the scaled content abuse policy. It tells site owners to fact-check AI output, including titles, descriptions and structured data.

What it means for your business

Our reading is that this guide shifts where the money should go. If Google's AI features draw on the same index and the same eligibility rules as ordinary results, then a site that is technically sound and publishes material competitors cannot easily copy is already doing most of the work. A separate budget line for Google-specific "GEO" tactics is hard to justify on Google's own description.

It also raises the bar for content. Generic articles are exactly what Google labels commodity content. A business has things a language model does not: its own project data, pricing logic, process, photographs, failures and customer questions. Content built from those is harder to replace with a summary.

One caution: this is guidance about Google Search and not a promise of results for any individual site. If visibility in other AI assistants matters to you, their requirements may differ.

What to do next

Start with eligibility. Confirm in Search Console that your important pages are indexed, and check that no robots rules or snippet restrictions are blocking pages you want shown. If your site is a JavaScript application, have a developer verify that the main content is present in what Google renders.

Then review any proposals you have received. Ask suppliers which items in an "AI optimization" package are for Google and which are for other platforms, and compare the Google items against the mythbusting list above. Work such as fixing indexing, improving page speed, consolidating duplicates and producing expert content is supported by the guide. An llms.txt file or a re-chunking project sold as a Google ranking measure is not.

Finally, pick a handful of pages that matter commercially and rewrite them around what only your business knows, with every fact checked by a person.

Conclusion

Google's guide to optimizing for generative AI features, added on May 15, 2026 and clarified on June 15, 2026, says that appearing in AI Overviews and AI Mode rests on the same foundations as ordinary Search: indexable pages, a clean technical structure and content with something original to say. It says llms.txt files, chunking, AI-specific markup and manufactured mentions are not needed for Google Search.

For most businesses that is good news, because the investment that helps with AI features is the same investment that makes a website work for customers. If your site needs technical work to meet that baseline, Entrant Technologies builds websites, web applications and custom software, and you can request a quote to discuss the scope.