Quick Summary
On May 15, 2026, Google published its first official guide on optimizing for generative AI features in Search. The headline: AEO and GEO are still SEO from Google’s perspective. The guide also lists five popular tactics site owners can stop worrying about, including llms.txt files and content chunking.
Key Takeaways
- Google’s position: AEO and GEO fall under SEO. Optimizing for generative AI search is optimizing for the search experience.
- Two AI mechanics power AI Overviews and AI Mode: retrieval-augmented generation (grounding) and query fan-out.
- Five tactics Google says to ignore: llms.txt files, content chunking, rewriting for AI, chasing inauthentic mentions, and overfocusing on structured data.
- Foundational SEO still applies: non-commodity content, clear technical structure, indexable pages with snippets, and Merchant Center or Business Profile data for commerce and local.
- Scaled content warning: creating separate pages for every fan-out query variation violates Google’s scaled content abuse policy.
- Scope note: this guidance applies to Google’s AI features only, not ChatGPT, Perplexity, or Claude.
Google has released its first official guide on optimizing for generative AI features in Google Search. The document, titled Optimizing your website for generative AI features on Google Search, was published in the Search Central documentation on May 15, 2026, and addresses AEO and GEO by name for the first time in official Google guidance.
Google’s Position on AEO and GEO
In one line: Google says AEO and GEO are not separate disciplines from SEO.
The guide directly addresses the AEO and GEO terminology that has become common across the SEO industry over the past two years. Google defines AEO as “answer engine optimization” and GEO as “generative engine optimization,” acknowledging both as terms used to describe work focused on improving visibility in AI search experiences.
From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.
The guide does not say AEO and GEO are wrong or harmful as concepts. It frames them as overlapping terminology for the same underlying practice when applied to Google’s generative AI features specifically.
The Two AI Mechanics Behind AI Overviews
In one line: Google’s generative AI features run on retrieval-augmented generation and query fan-out.
The guide identifies two specific AI techniques that power generative AI features in Google Search. These are the mechanisms behind AI Overviews and AI Mode.
Retrieval-Augmented Generation (RAG)
Google describes RAG as a technique, also known as grounding, used to improve the quality, accuracy, and freshness of AI responses. The system relies on core Search ranking to retrieve relevant, up-to-date web pages from the Search index, then reviews the specific information from those pages to generate a response. Clickable links to the supporting pages appear alongside the response.
Query Fan-Out
Query fan-out is described as a set of concurrent, related queries generated by the model to fetch additional relevant search results.
Google’s Example
Original query: “how to fix a lawn that’s full of weeds”
Fan-out queries: “best herbicides for lawns,” “remove weeds without chemicals,” “how to prevent weeds in lawn”
What Google Says To Do
In one line: Apply foundational SEO. Nothing new is required.
The guide outlines foundational practices Google recommends for visibility in generative AI search. These map to existing SEO fundamentals rather than introducing new requirements.
| Recommendation | What Google Says |
|---|---|
| Unique point of view | Create first-hand content rather than recycling what others have published or what an AI model could easily produce. |
| Non-commodity content | Content based on common knowledge (“7 Tips for First-Time Homebuyers”) adds little. Expert or experienced takes that go beyond common knowledge perform better. |
| Clear organization | Use paragraphs, sections, and headings that help human readers navigate. |
| Images and video | Generative AI features can surface relevant images and videos. Following existing image and video SEO best practices is sufficient. |
| Technical structure | Pages must be indexed and eligible to appear in Search with a snippet. Crawling, JavaScript handling, and page experience requirements all apply. |
| Local and ecommerce details | Merchant Center feeds and Google Business Profiles help products and local information appear in AI responses. |
What Google Says To Ignore
In one line: Five tactics the AEO community debates that Google says are unnecessary for its AI features.
The mythbusting section is the most directly newsworthy part of the document. Google explicitly names five tactics and tells site owners they do not need to implement them for visibility in Google’s generative AI features.
1. llms.txt Files and Special Markup
Google states that site owners do not need to create new machine-readable files, AI text files, or markdown to appear in generative AI search. The guide notes that Google may discover and index many file types beyond HTML, but indexing a file does not mean Google treats it in any special way. This echoes earlier analysis from Ahrefs showing major AI providers have not confirmed they parse llms.txt files.
2. Content Chunking
According to the guide, breaking content into small pieces is not required for AI to understand it. Google’s systems can understand multiple topics on a single page and surface the relevant section to users. The guide explicitly states there is no ideal page length.
3. Rewriting Content For AI Systems
Google says writing in a specific style for generative AI search is unnecessary. AI systems understand synonyms and general meaning, so site owners do not need to capture every variation of how someone might phrase a query.
4. Seeking Inauthentic Mentions
The guide directly addresses the practice of pursuing brand mentions across blogs, forums, and videos for AI visibility. Google states that while its generative AI features can surface what is being said about products and services across the web, seeking inauthentic mentions is not as helpful as it might seem. Spam systems and quality ranking systems both apply to generative AI features.
5. Overfocusing On Structured Data
Structured data is not required for generative AI search, and the guide states there is no special schema.org markup needed. Google still recommends structured data as part of an overall SEO strategy because it helps eligibility for rich results, but it is not a generative AI ranking signal.
Agentic Experiences and Universal Commerce Protocol
In one line: Google is signaling that browser agents and commerce protocols are the next frontier.
The guide includes a forward-looking section on AI agents. Google describes agents as autonomous systems that can perform tasks like booking reservations or comparing product specifications. Browser agents may access a website by analyzing visual renderings, inspecting the DOM, and interpreting the accessibility tree.
Google names Universal Commerce Protocol (UCP) as an emerging standard that will let Search agents do more on behalf of users. Site owners interested in agentic readiness are pointed to the agent-friendly website best practices on web.dev.
The Scaled Content Warning
In one line: Creating a separate page for every fan-out query is a spam policy violation.
The guide includes a notable warning about creating separate content for every possible query variation, including fan-out queries. Google states that doing so primarily to manipulate rankings or generative AI responses violates the scaled content abuse spam policy. The guide frames this as both a policy issue and an ineffective long-term strategy, citing improvements in Google’s ability to understand page relevance without exact keyword matches.
What This Means For Site Owners
In one line: No new ranking signals, no new schema, no new files. Same fundamentals, reframed for AI Overviews.
The guide’s recap section lists four takeaways: apply SEO best practices to generative AI search, create non-commodity people-first content, prioritize effective SEO over AEO and GEO tactics, and explore agentic experiences as they emerge.
The document does not introduce new ranking signals, new schema requirements, or new technical files. It reframes existing SEO fundamentals in the context of AI Overviews and AI Mode, and uses the mythbusting section to push back on tactics Google views as ineffective or off-policy for its own generative AI features. For teams looking to measure and improve AI visibility across multiple platforms, dedicated AEO tracking software remains the standard approach for tracking citations across ChatGPT, Perplexity, Claude, and Google’s AI features.
Scope Note
This guidance applies specifically to Google Search’s generative AI features. It does not describe how ChatGPT, Perplexity, Claude, or other AI search products surface or cite content, and Google does not claim authority over those systems.
Frequently Asked Questions
Does Google recognize AEO and GEO as separate from SEO?
No. Google’s official position, as stated in the May 15, 2026 guide, is that optimizing for generative AI search is optimizing for the search experience and therefore still SEO.
Do I need an llms.txt file to appear in AI Overviews?
No. Google explicitly states that llms.txt files and similar AI-specific markup are not required to appear in generative AI search features.
What AI techniques power Google’s AI Overviews and AI Mode?
Google identifies two: retrieval-augmented generation (RAG), also called grounding, which retrieves relevant pages from the Search index to generate responses, and query fan-out, which generates concurrent related queries to fetch additional results.
Is structured data required for generative AI search?
No. Structured data is not required for generative AI search and there is no specific schema.org markup needed. Google still recommends structured data for traditional rich results eligibility.
What is query fan-out?
Query fan-out is a technique where the AI model generates multiple related queries from a single user prompt and runs them concurrently to gather additional relevant search results before generating a response.
What is Universal Commerce Protocol?
Universal Commerce Protocol (UCP) is an emerging protocol Google names in the guide as a standard that will allow Search agents to take more actions on behalf of users, such as completing transactions or comparing products across sites.
