Understanding Google's AI Search Guidelines: What Brands Need to Know

What did Google actually publish in May 2026?

On May 15, 2026, Google Search Central published a new resource titled 'Optimizing your website for generative AI features on Google Search.' Announced by John Mueller on the Search Central Blog, it is the first time Google has directly addressed how its AI-powered search experiences, specifically AI Overviews and AI Mode, choose and present sources.

The document is housed under a new 'Generative AI fundamentals' section in the Search Central documentation, sitting alongside the core indexing and crawling guidance every site owner already follows. That placement is deliberate. Google is not treating this as a specialist addendum for early adopters. It is foundational documentation.

Before this publication, there was no authoritative source to settle the internal debates that had been running for the better part of two years: Do AEO, GEO, and traditional SEO require different playbooks? Should you hire separate practitioners for AI visibility and organic search? Should you rewrite your content in formats AI parsers prefer? Google has now answered these questions on the record.

The guide's headline statement is blunt: 'From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.' That sentence alone should close several internal discussions brands have been having with agencies selling AI-specific retainers disconnected from content fundamentals.

What tactics did Google explicitly debunk?

The guide is as notable for what it rules out as for what it recommends. Several practices that spread across the industry in 2024 and 2025 get direct pushback.

On llms.txt: Google's crawler may discover these files, but they are treated as plain text. There is no special indexing pathway, no preference signal, and no evidence that including one improves inclusion in AI Overviews or AI Mode. Brands that hired consultants to implement llms.txt strategies received nothing in return from Google's systems.

On content chunking: A significant body of advice circulated suggesting that AI systems prefer short, discrete content fragments and that editors should restructure articles accordingly. Google says its systems can understand multi-topic pages and extract the relevant passage without any author pre-fragmentation. Long-form, deeply sourced content written for a human reader is not at a disadvantage.

On special schema for AI: The guide states that no additional schema requirements exist for appearing in generative AI features beyond those that support traditional search results. Implementing proprietary or experimental markup on the assumption that it signals AI-friendliness is unsupported.

On AI-specific rewrites: The guide makes no mention of rewriting existing content in AI-preferred syntax, Markdown formatting, condensed summaries, or bullet-dense structures. Quality and uniqueness are the stated levers. Syntax is not.

The cumulative effect of these clarifications is to redirect attention away from technical workarounds and back to the harder, more durable work of producing content that earns trust. Brands that invested in tactic-first AI visibility programs built on these assumptions should now reassess what they were actually buying.

Who does this affect, and how urgently?

Any brand that relies on organic search traffic needs to take this seriously. The shift is not speculative. It is already visible in click data.

Research from Ahrefs, covering 300,000 keywords split evenly between queries with and without AI Overviews, found that the presence of an AI Overview correlates with a 58% lower average clickthrough rate for the top-ranking organic page. That figure was 34.5% in a prior study conducted eight months earlier. The suppression is not a plateau. It is accelerating.

The brands most exposed are those carrying large volumes of commodity content: summaries of widely available information, listicles that cover the same ground as a hundred competitor pages, and how-to guides that offer no original data or firsthand experience. Google's guide specifically calls out non-commodity content as the standard for generative AI feature inclusion. If your content does not offer unique insight beyond common knowledge, it is a candidate for displacement, not citation.

B2B SaaS teams running comparison-heavy content are directly in scope. Local service businesses trying to appear for prompts like 'best accountant in Denver' are directly in scope. Media brands and publishers with large article archives built on aggregated reporting are directly in scope. For all of them, the question is no longer whether they rank. It is whether they are cited inside the answer where the reader's attention is concentrated.

What does Google actually want from content?

The guide returns repeatedly to a single concept: non-commodity content. Google defines commodity content as material that provides nothing beyond what is already widely available. The standard set for generative AI feature inclusion is content with original research, firsthand experience, expert synthesis, or data that no competitor has sourced.

E-E-A-T, standing for Experience, Expertise, Authoritativeness, and Trustworthiness, remains the governing framework. The guide reinforces that experience-led content, specifically content that demonstrates real-world knowledge rather than summarizing others, carries more weight. The 'Experience' component added in 2022 is not a checkbox. It is Google's mechanism for distinguishing primary sources from derivative ones inside AI features.

Technical requirements remain minimal but non-negotiable. Pages must be indexed and eligible for snippets to appear in AI Overviews or AI Mode at all. Google recommends semantic HTML, standard JavaScript SEO practices, good page experience signals, and reduced duplicate content. None of this is new. The difference is that these requirements now act as the entry ticket into AI features, not just traditional results. A brand doing sophisticated content work on pages Google cannot properly crawl is invisible regardless of how good the writing is.

Local, shopping, image, and video content receive specific callouts in the guide as areas where completeness of information directly influences AI feature inclusion. For local businesses, this means current and consistent structured citations, hours, and service details across every source Google can reach.

How should brands respond to these guidelines?

The guide clarifies what to stop doing. It is less prescriptive about sequencing the response. Based on what Google has confirmed, three moves are worth prioritizing above everything else.

First, audit for commodity content. Run your highest-traffic pages against Google's stated criterion: does this page offer unique insight beyond what is commonly available? If the honest answer is no, the page is exposed. Refreshing it with original data, firsthand perspective, or expert analysis is not optional if you want generative AI feature inclusion. A page that summarizes what three other pages already say is a candidate for omission, not citation.

Second, confirm your technical foundations. Indexed pages, snippet eligibility, semantic HTML, and clean crawl paths are the entry requirement, not the differentiator. Many brands discover during audits that high-value pages have crawl issues, canonical problems, or snippet-blocking that disqualifies them from AI features entirely. Fix those before investing further in content quality work.

Third, start tracking citation share, not just rankings. Traditional rank tracking tells you where a page sits in a list. It does not tell you whether you are being cited inside the AI answer at the top of the page, where the reader's attention lands first. On queries where an AI Overview appears, the cited source captures a meaningfully different level of attention than the organic results below it. Measuring that gap, across the range of queries relevant to your category, is how you find out whether your content is working or just ranking.

Sources

Google published its first official AI search optimization guide on May 15, 2026 Google Search Central Blog, 2026-05-15

Google's full guide to optimizing for generative AI features on Google Search Google Search Central, 2026-05-15

Google's new AI search guide calls AEO and GEO 'still SEO' and debunks llms.txt and content chunking Search Engine Journal, 2026-05-16

AI Overviews correlate with a 58% lower clickthrough rate for top-ranking organic pages, up from 34.5% eight months prior Ahrefs, 2026-02-01

Google publishes guide on optimizing for generative AI features in Search Search Engine Land, 2026-05-16