Google updated its Search Central documentation on May 15 with a dedicated guide on how to optimize content for AI search engines — AI Overviews and AI Mode. It’s the first time Google has put official, on-the-record guidance behind a topic that’s spawned an entire cottage industry of “AEO” and “GEO” advice over the past year.
The short version: most of that advice, according to Google, is noise.
SEO is the strategy. There is no separate one.
The guide’s central claim is that generative AI features run on Google’s existing core ranking and quality systems. AI Overviews and AI Mode aren’t a parallel engine — they retrieve from the same index, ranked by the same systems. Google addresses “AEO” (answer engine optimization) and “GEO” (generative engine optimization) directly and says that from its perspective, optimizing for generative AI search is just optimizing for search. Same work, new acronyms.
Two mechanisms are worth knowing because they explain how content surfaces in AI answers:
Retrieval-augmented generation (RAG), also called grounding. Google’s ranking systems retrieve relevant, current pages from the index, and the model generates a response grounded in them — with clickable links back to the source pages.
Query fan-out. The model generates a set of concurrent related sub-queries to gather more sources. For a query like “how to fix a lawn full of weeds,” fan-out might add “best herbicides for lawns” and “how to prevent weeds in lawn.” One user query becomes many retrieval passes.
What Google says to ignore
The most useful part of the guide is the mythbusting section. Google lists specific tactics it says have no effect on visibility in AI search:
- llms.txt files and other special markup. No new machine-readable files needed. Google may crawl many file types, but that doesn’t mean any file gets special treatment.
- “Chunking” content into small pieces. Google’s systems handle multiple topics on one page and surface the relevant part. There’s no ideal page length.
- Rewriting content specifically for AI. AI systems understand synonyms and intent, so chasing every long-tail keyword variation is wasted effort.
- Chasing inauthentic “mentions” across blogs and forums. Core ranking still rewards quality; spam systems still filter the rest.
- Overfocusing on structured data. It’s not required for AI search — though Google still recommends it for rich results in regular Search.
How to optimize content for AI search engines, per Google
This is the part anyone asking how to optimize for AI search engines should read closely, because Google’s ranked priorities are unremarkable — which is the point.
Unique, non-commodity content does the most work. The guide draws a sharp line: commodity content like “7 Tips for First-Time Homebuyers” restates common knowledge anyone could produce, while non-commodity content like “Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line” carries first-hand experience and a real point of view. Content “that could easily be produced by a generative AI model” is the thing to avoid.
After that, the advice on how to optimize for AI search is the same checklist good SEO has always run: a clear technical structure (indexed, crawlable, good page experience, reduced duplicate content), semantic HTML for human readability, and — for commerce and local businesses — Merchant Center feeds and Business Profiles.
There’s also a forward-looking section on agentic experiences: browser agents that read your DOM and accessibility tree to complete tasks, plus emerging standards like the Universal Commerce Protocol. Google frames this as an “if you have extra time” consideration, not a priority.
Our read
For anyone selling AI-specific SEO retainers, this guide is awkward. The tactics Google explicitly dismisses — llms.txt, chunking, AI-specific rewrites — are exactly what a lot of GEO packages are built on. We’ve made a version of this argument before in our work on generative engine optimisation and ranking on ChatGPT: the fundamentals carry most of the weight, and AI search rewards the same depth and first-hand experience that good SEO always has.
One caveat. This guidance covers Google’s surfaces only — AI Overviews and AI Mode. It says nothing about ChatGPT, Perplexity, or other engines that run their own retrieval and citation logic. Optimizing for Google’s AI features is now officially “just SEO.” Optimizing for the wider set of answer engines is a separate, still-unsettled question — and worth tracking. We’ve covered the broader picture in our guide on how to rank on AI Overviews.
The takeaway is almost boring, and Google seems to want it that way: keep doing the fundamentals, ignore the hacks, and write things a machine couldn’t have written.
Source: Google Search Central — Optimizing your website for generative AI features.