This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
Google Search Central published a guide today on optimizing websites for generative AI features in Search, meaning AI Overviews and AI Mode. In the announcement, John Mueller lists what it covers: the importance of “valuable, unique, non-commodity content,” tips on local, shopping, image and video content, initial guidance on AI agents, why SEO best practices remain foundational, and “mythbusting common ‘AEO/GEO’ misconceptions.”
The stated audience is website owners, SEOs and developers. Communications and PR teams should read it too, mostly for the mythbusting section, because they are increasingly the ones being sold answer engine optimization.
What the guide says you can ignore
The guide is direct. From Google Search’s perspective, it says, “optimizing for generative AI search is optimizing for the search experience, and thus still SEO.” It then lists things site owners can ignore for Google Search. There is no need for llms.txt files or other special markup, because Google Search does not use them. There is no need to “chunk” content into small pieces. There is no need to rewrite content in a special way for AI. Seeking “inauthentic mentions” across the web “isn’t as helpful as it might seem,” since core ranking systems focus on quality and other systems block spam. And structured data is not required for these features, though it remains useful for rich results.
Search Engine Journal points out that this is more explicit than Google’s earlier documentation on AI features, particularly in naming llms.txt, chunking and AEO/GEO directly.
Why this lands on PR desks
Several of the dismissed tactics map closely onto services pitched to communications teams. Seeding mentions across forums, listicles and placed articles so that AI answers pick up a brand is often a familiar PR deliverable under a new name. Restructuring a newsroom into short question-and-answer fragments is frequently sold as AI readiness. Google is saying, for its own products, that the first is unlikely to help and may be filtered, and that the second is unnecessary.
That does not make every AEO recommendation wrong, and it does not mean other AI systems behave like Google. The guide is explicit that its advice concerns Google Search, and Google is describing its own systems as an interested party. Still, it is the clearest written statement Google has made on the subject, and it gives communications leaders a reference point.
A short checklist for comms teams
Ask any agency proposing AEO or GEO work which recommendations apply to Google specifically, and how they square with this guide. If a proposal is built on mentions, ask how many of those mentions would be authentic by any reasonable standard.
Redirect effort toward what the guide calls non-commodity content. Its example contrasts a generic “7 Tips for First-Time Homebuyers” with a piece built on specific first-hand experience. For a company, the equivalent is original research, specific expertise, and primary facts it is uniquely placed to publish. For a corporate newsroom, that usually means dated, sourced statements that other writers can cite, rather than another overview of the industry.
Keep the basics in order. The guide repeats that a page must be indexed and eligible to show a snippet in Search to appear in these features. A statement page that is blocked from crawling or marked noindex is invisible to the AI answer as well.
And measure the outcome you care about. That is not whether the brand is mentioned in an answer. It is whether the answer about the company is accurate, and which sources it relies on.
In 2023 I wrote that “helpful content” isn’t a reputation strategy and that Google keeps punishing performance theater. This guide is the generative-search version of the same message. The acronyms are new. What works is still being the most credible source for the facts about yourself.