AI Search
Google published a new guide on May 15 to help website owners optimize for generative AI in Search, addressing AEO and GEO misconceptions directly for SEO.
Google published a new official guide on May 15, 2026, aimed at helping website owners, SEOs, and developers understand how to optimize their content for generative AI features within Google Search – and in doing so, challenged a wave of industry terminology that the company’s own representatives had been criticizing for months.
The guide, titled “Optimizing your website for generative AI features on Google Search,” was announced by John Mueller, a member of Google’s Search Relations team, through the Google Search Central Blog. It is the first consolidated resource the company has produced specifically addressing how content surfaces – or fails to surface – inside its generative AI products, including AI Overviews and AI Mode. The announcement arrived at a moment when confusion about so-called AEO (Answer Engine Optimization), GEO (Generative Engine Optimization), and a proliferating alphabet of other acronyms had reached notable intensity across the digital marketing industry.
According to Google, the new resource is structured around five areas. The first is guidance on the importance of providing valuable, unique, non-commodity content – a principle that has sat at the center of Google’s quality guidance for years but takes on fresh urgency in an environment where AI systems synthesize answers from across the web. The second covers practical tips about local, shopping, image, and video content, acknowledging that generative AI features do not operate identically across these formats.
The third area takes direct aim at widespread misinformation. The guide includes what Google describes as mythbusting of common “AEO/GEO” misconceptions – signaling that the company views much of the advice circulating under those acronyms as inaccurate or misleading. This is not the first time Google has addressed the acronym problem. On August 14, 2025, Mueller warned that aggressive promotion of GEO, AIO, and AEO terminology may indicate spam tactics targeting marketing professionals. A December 17, 2025 podcast episode from the Search Relations team concluded that optimizing for AI-powered search requires no fundamental changes from traditional SEO practices. The May 15 guide is, in effect, a permanent written version of that recurring message.
The fourth section provides initial guidance on AI agents, which Google acknowledges is a “quickly emerging and evolving space.” This is new territory for Google’s documentation. The rise of AI agents capable of completing tasks on behalf of users – not merely retrieving information – creates content accessibility questions that standard SEO guidance has not addressed. How Google’s search infrastructure is evolving toward agentic behavior has been a recurring topic in the industry, and including agents in the new guide marks the first time Google has offered official documentation touching on this dimension.
The fifth and perhaps most strategically significant component is an explanation of why SEO best practices remain foundational to success with generative AI features. According to Google, the same principles that govern visibility in traditional blue-link results – structured, accessible, high-quality content – also determine how content is selected for AI-generated responses.
The new generative AI guide sits alongside, rather than replacing, Google’s existing SEO Starter Guide, which covers fundamental mechanics such as crawling, indexing, URL structure, duplicate content, meta descriptions, and link strategy. According to Google’s SEO Starter Guide documentation, last updated December 10, 2025, Google is “a fully automated search engine that uses programs called crawlers to explore the web constantly, looking for pages to add to our index.” Content visibility depends on that crawler being able to access, parse, and understand a page.
Those mechanics have not changed. What the new guide adds is a layer specific to how content gets selected for AI-generated features sitting on top of that index. The interplay between the two documents matters. A site that fails the foundational requirements of the SEO Starter Guide – poor crawlability, blocked resources, thin or duplicate content – will not suddenly become eligible for AI Overviews through any specialized optimization strategy.
The SEO Starter Guide explicitly dismisses several persistent myths: that meta keywords carry ranking weight (Google states they do not), that keyword density strategies help (Google classifies keyword stuffing as a spam policy violation), that exact-match domains confer significant ranking advantages, that content length alone determines ranking, and that E-E-A-T functions as a direct ranking factor. According to the documentation, “Thinking E-E-A-T is a ranking factor – No, it’s not.” These clarifications are relevant to the broader AI search conversation. Some practitioners have promoted content lengthening and E-E-A-T restructuring as AI optimization strategies, despite Google’s existing documentation contradicting those premises.
The guide also covers the mechanics of title links and snippets. According to Google, the title link in search results can draw from multiple sources including the page’s title element and other headings. Snippets – the descriptive text below the title – are sourced from actual page content, though they may also draw from a well-written meta description. The practical guidance here is unchanged from the pre-AI era: clear, accurate, unique titles; succinct meta descriptions; high-quality images with descriptive alt text placed near relevant text.
The context in which this guide arrives is anything but calm for publishers. SISTRIX’s March 2026 data showed click-through rates at position one in Google Search collapsing from 27% to 11%, driven by the expansion of AI features that answer queries directly on the results page. AI Overviews now reach over 2 billion users globally, a figure Google disclosed during Alphabet’s second quarter 2025 earnings on July 23, 2025. Research published in April 2025 documented that AI Overview presence correlates with reductions in organic click-through rates of 34.5% to 54.6% for top-ranking pages.
Against that backdrop, the industry developed a cottage economy of consultants and frameworks promising specialized AI search visibility techniques. On January 8, 2026, Danny Sullivan explicitly warned against fragmenting content into bite-sized chunks for LLM optimization, saying such tactics focus on manipulating ranking systems rather than serving readers. On August 27, 2025, Mueller issued a direct warning against using large language models to build topic clusters, stating such practices create site “liability.” The May 15 guide consolidates all of this guidance into a single, navigable document.
The guide’s emphasis on “valuable, unique, non-commodity content” reflects a sustained strategic position. Research from Semrush published in June 2025 found that AI search visitors are worth 4.4 times more in economic terms than traditional organic search visitors, which suggests that while AI features may reduce total click volume, the clicks that do occur carry higher commercial value. For content producers capable of providing genuinely differentiated information, the shift may prove less damaging than headline traffic numbers suggest.
What the guide does not provide is a mechanism for controlling whether content appears in AI features specifically. Google’s existing documentation allows website owners to block crawling, noindex pages, and use snippet controls. But granular controls distinguishing traditional search results from AI Overviews or AI Mode have not been introduced, despite repeated requests from publishers over the course of 2025. The new guide does not change that.
The inclusion of guidance on AI agents is notable precisely because Google acknowledges it is preliminary. The agent ecosystem – tools that perform bookings, purchases, and complex research tasks autonomously – creates a content accessibility challenge distinct from traditional search ranking. Google has been testing agentic search features since at least late 2025, when Mueller confirmed that AI Mode could “reserve a table” automatically. Whether an AI agent completing a task on behalf of a user cites or credits a source website is a different question from whether a user visiting that site found the content useful.
Tests conducted by the DEJAN agency and published on May 30, 2025, found that Google’s AI Mode draws from what appears to be a proprietary content store separate from the live search index, a finding that intersected with antitrust court documents describing a technology called FastSearch. According to those documents, FastSearch “delivers results more quickly than Search because it retrieves fewer documents, but the resulting quality is lower than Search’s fully ranked web results.” If AI agents rely on similar cached or pre-indexed infrastructure rather than live web retrieval, then real-time content freshness – a core property of local and news content – may not be sufficient for inclusion in agent-generated responses.
The new guide’s guidance on local and shopping content is therefore particularly relevant for businesses whose product prices, availability, and store hours change frequently. Any gap between the live web and the infrastructure AI features draw from has direct commercial consequences for those categories. Google’s guide acknowledges this reality by addressing them specifically, though the technical mechanisms that govern freshness within AI features remain undisclosed.
The “mythbusting” framing in the guide’s announcement is deliberate. According to Google, common AEO and GEO misconceptions – that content needs to be reformatted into FAQ structures, that headers should be rewritten as questions, that content chunks need to be artificially shortened for LLM consumption – are not supported by how the underlying systems work. This position has been consistent across Google’s 2025 and 2026 communications.
An analysis published in July 2025 examined the history of optimization panics in SEO, noting that mobile-first indexing, voice search, and social signals each generated waves of specialist advice that largely failed to materialize into mandatory practice changes. The pattern repeats: a format change in search results prompts predictions of a new optimization discipline, consultants develop frameworks and acronyms, and Google eventually clarifies that the existing principles still apply. The May 15 guide is that clarification for the generative AI era.
What distinguishes the current moment is scale. AI Overviews serve billions of users. AI Mode is no longer experimental for most US users – Google removed the waitlist on May 20, 2025. The traffic stakes are real: Google’s strategic shift away from routing users to external websites is a documented direction acknowledged by company executives. Google’s Head of Search Elizabeth Reid acknowledged in March 2025 that the search bar would become “less prominent over time.” In that context, the new guide serves both as practical documentation and as a signal of Google’s preferred narrative: that its generative AI features and traditional SEO exist on the same optimization continuum, not separate tracks.
A Brainlabs study released in July 2025 found that 96% of links appearing in AI Overviews came from websites already ranking in the top 10 organic results. This single data point is among the most consequential for practitioners navigating the AI search landscape. It means that no shortcut exists from outside the top of the organic index into AI feature visibility. The path runs through conventional ranking – which is precisely what the SEO Starter Guide addresses and what the new generative AI guide reinforces.
Mueller’s post on May 15 notes that the guide sits within the Google Search Central documentation under a “Generative AI fundamentals” navigation section – a section that did not exist in earlier versions of the documentation. Its placement alongside the existing SEO fundamentals, crawling and indexing, and ranking and search appearance sections reflects a deliberate decision to frame AI optimization as an extension of existing practice rather than a separate discipline.
According to Google, the purpose is to “help you continue to create great content that reaches your audience effectively through Google Search.” The guide is available through Google Search Central, and Mueller invited comment through LinkedIn and the Google Search Central Community.
The following is drawn directly from the two documents Google published: the May 15, 2026 announcement and the SEO Starter Guide (last updated December 10, 2025). Each point reflects explicit guidance from those sources.
Who: John Mueller, a member of Google’s Search Relations team, published the announcement on behalf of Google Search Central.
What: Google released a new official guide titled “Optimizing your website for generative AI features on Google Search,” covering five areas: the importance of unique and non-commodity content, guidance on local, shopping, image, and video content, mythbusting of common AEO and GEO misconceptions, initial guidance on AI agents, and an explanation of why traditional SEO best practices remain foundational for AI search visibility.
When: The guide was published on May 15, 2026, through the Google Search Central Blog.
Where: The guide is available through Google’s Search Central documentation, listed under a new “Generative AI fundamentals” section that was not present in earlier versions of the documentation navigation.
Why: Google published the guide in response to widespread confusion within the digital marketing industry about whether generative AI features require fundamentally different optimization approaches. The company has consistently argued across multiple official communications since mid-2025 that traditional SEO practices remain effective – and the guide converts that repeated verbal guidance into permanent documentation as AI Overviews and AI Mode serve a combined user base exceeding 2 billion people monthly.
Every morning, get a curated digest of the last 24 hours in programmatic advertising
