What Does Google's May 15 AI Search Guide Mean for SaaS Sites?

On May 15, 2026, Google Search Central published its first official guide to optimizing for generative AI features on Google Search. The guide's central position is direct: from Google's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO. The document explicitly rejects the need for llms.txt files, content chunking, special AI markup, or rewriting content for AI. At Phoenix Studio in Bengaluru, my inbox the week of May 11 to 15 fielded multiple founder pings about AEO retainers, llms.txt deployments, and GEO services other agencies were pitching. Google just told us that most of that work is unnecessary. In this piece I walk through what the guide actually says, what it changes for a Webflow build checklist, and what I tell B2B SaaS founders who were quoted high prices for AEO services they did not need.

What exactly did Google publish on May 15, 2026?

Google Search Central published a Search Central Blog post and an accompanying full guide titled "Optimizing your website for generative AI features on Google Search." The blog post sits at developers.google.com/search/blog/2026/05/a-new-resource-for-optimizing and the full guide sits in the Search Central docs fundamentals section. The document is Google's first official position statement on how generative AI features in Search interact with traditional SEO.

The guide's core position is that there is no separate optimization discipline for AI Overviews or AI Mode beyond what already constitutes good SEO. Indexability, crawlability, unique non-commodity content, clean DOM, and structured data where it supports rich results remain the foundation. The document explicitly addresses and rejects several emerging tactics that other vendors had been pitching as necessary for AI search visibility.

Why is Google calling AEO and GEO "still SEO"?

Google's position is that the underlying ranking, retrieval, and citation systems that power AI Overviews and AI Mode are extensions of Google Search's existing index. The retrieval-augmented generation pattern that AI Overviews use pulls from the same index that traditional search results pull from. The query fan-out pattern in AI Mode expands one user query into multiple Search queries against the same index. There is no separate AI-only index that requires separate optimization.

The practical implication is that the work to earn AI Overview citations is the same work that earns traditional Search rankings. The piece I wrote on pre-I/O Gemini Omni AEO prep covered the answer-block discipline that performs against both surfaces. The May 15 guide is Google's confirmation that the discipline is correct and that vendors selling specialized AEO services are mostly selling existing SEO work under a new label.

Does my Webflow site need an llms.txt file?

No, Google's May 15 guide explicitly states that you do not need to create an llms.txt file for AI Overviews or AI Mode visibility. The llms.txt proposal, which gained traction in 2024 and 2025 as an AI-specific robots.txt equivalent, does not affect how Google's AI features discover or cite content. The standard robots.txt and sitemap.xml files that already control Google crawler access continue to be the only required signals.

For Phoenix Studio clients who were pitched llms.txt deployment as part of an AEO retainer, the guide is the simplest possible counter-argument. The file does not need to exist on the site. Time spent deploying it is time that could go to actual content work. The piece on Google's FAQ rich results deprecation covered another tactic that vendors had been over-selling and that Google has now reframed.

Should I be chunking content for AI Overviews?

No, Google's May 15 guide states that content chunking specifically for AI features is not necessary. Google's retrieval systems already segment and extract content from indexed pages based on query relevance. Sites that deliberately fragment content into AI-optimized chunks do not earn additional citation surface and may produce worse user experience for traditional Search visitors.

The structural pattern that does work is the answer-block discipline I have been pushing with retainer clients. A 40 to 60 word answer block at the top of an H2 reads cleanly when extracted by an AI Overview, reads cleanly inside the article body for human readers, and matches the structural pattern that traditional Search has favored for years. The discipline is one structure that serves both surfaces, not two separate optimizations.

How does query fan-out actually work in AI Mode?

Query fan-out is the pattern by which AI Mode expands a single user query into multiple related Search queries to gather context for the AI-generated response. A user query like "best B2B SaaS analytics tools for fintech" might fan out into queries about analytics tool categories, fintech-specific compliance requirements, B2B SaaS evaluation criteria, and recent product reviews. The AI Mode response aggregates content from sites that surface in any of those queries.

The practical implication for B2B SaaS Webflow sites is that ranking for the exact user query is not the only path to AI Mode citation. Sites that rank for any of the fanned-out related queries can earn citation surface. The audit pattern at Phoenix Studio is to identify the three to five most-likely fan-out queries for each high-priority page and verify that the site has answer-block content for each. The work compounds across the broader topic cluster rather than depending on a single keyword.

What does Google's RAG explanation change for B2B SaaS sites?

Google's May 15 guide describes the retrieval-augmented generation pattern that AI Overviews use, which retrieves relevant content from the Search index and generates a synthesized response with citations to the retrieved sources. The explanation matters because it confirms that the path to citation is through indexed content, not through any AI-specific submission process or feed.

For B2B SaaS Webflow sites, the RAG explanation removes ambiguity about what the optimization target is. The target is being indexed, being relevant to the query, being structured for clean extraction, and having unique non-commodity content. Sites that ship that combination earn citation surface. Sites that try to game the AI-specific layer do not. The discipline is straightforward, which is also why some vendors had been pitching it as more complicated than it actually is.

Is schema markup still worth shipping on Webflow CMS templates?

Yes, schema markup is still worth shipping on Webflow CMS templates, but not because AI features require it. Google's May 15 guide states that there is no special schema.org markup you need to add for AI features. Schema markup remains valuable because it powers rich results in traditional Search and provides structured data signals that help with disambiguation.

For Phoenix Studio's standard Webflow build checklist, schema markup ships on Article templates, Product templates, FAQ patterns, and BreadcrumbList navigation. The work is small per template and produces durable benefits across traditional Search and AI Search. The piece on FAQ rich results deprecation covered the one schema type that lost its rich-result surface this year, and what to ship instead.

What about agentic browsers and the Universal Commerce Protocol?

Google's May 15 guide references the Universal Commerce Protocol and agentic browsing patterns as emerging surfaces that may interact with site content in the future. The guide does not specify implementation requirements for these surfaces beyond clean DOM, accessible markup, and the agent-friendly website patterns already documented at web.dev.

For B2B SaaS Webflow sites, the practical pattern is to ensure clean semantic HTML, accessible form labels, and schema where it supports product or commerce features. The Phoenix Studio build checklist already includes these patterns for accessibility and traditional SEO. The agent-readiness work is largely the same work, which is consistent with Google's broader framing that AI search optimization is search optimization.

Which AEO agency pitches should B2B SaaS founders ignore now?

B2B SaaS founders should ignore AEO agency pitches that propose llms.txt deployment, content chunking specifically for AI, AI-specific schema markup beyond what Google's rich-results requirements already demand, and rewriting existing SEO-performant content into AI-optimized formats. Google's May 15 guide explicitly states that these tactics are not necessary for AI features.

The honest read from Phoenix Studio in Bengaluru is that several Indian agencies in the week before May 15 were pitching AEO retainers ranging from INR 50,000 to INR 1.5 lakh per month to B2B SaaS founders. Most of those retainers were rebadged SEO services with AI-flavored framing. The May 15 guide is the simplest possible answer to those pitches: read Google's actual position, then evaluate which work is real and which is rebrand. The piece on AI Mode Reddit citations covered the parallel discipline of evaluating which content actually earns citation.

What should a Webflow Partner change in their build checklist this week?

A Webflow Partner should not change their build checklist this week based on Google's May 15 guide. The guide confirms that existing SEO best practices are the correct discipline for AI Search visibility. The only adjustment for Phoenix Studio is to add a CMS audit pattern that surfaces non-commodity content gaps, which is a refinement of existing work rather than new work.

The structural pattern through the next 30 days is to update service-page copy to reflect Google's explicit position, audit retainer scopes for any AEO-specific work that is no longer necessary, and reallocate time saved toward the answer-block rewrites that actually move citation surface. The piece on pre-I/O AEO prep covers the answer-block discipline that compounds across both Search surfaces and represents the durable work for the next quarter.

For the next layer of nuance on top of the May 15 AI search guide, my piece on why the May 2026 Google Quality Rater update matters for Webflow B2B SaaS sites covers the Synthesized Authority and Compression Fidelity rubrics that change what counts as a defensible page.

If you are a B2B SaaS founder who was quoted high prices for an AEO retainer in the last 30 days and want to talk through which line items are still necessary after Google's May 15 guide, drop me a line and tell me what your current Webflow build checklist looks like. I will share the Phoenix Studio audit pattern I am running on retainer clients this week. Let's chat.