Google publishes a guide to optimizing for generative AI features on Search, and one of its sections is titled "Mythbusting generative AI search: what you don't need to do." It names five practices that are being sold to businesses right now and says, in Google's own words, that they can be ignored for Google Search. If your agency has been asked to buy any of them — or has been quoting them to clients — this is the primary source, and it is worth reading before the next proposal goes out.
One boundary first, because it decides how far the guidance travels: everything below is Google describing Google Search, including its generative AI features. It is not a statement about how other assistants retrieve or cite. Treating it as universal would repeat the mistake it corrects.
What did Google actually publish?
A guide, last updated in July 2026, aimed at site owners who want to understand how generative AI features on Google Search relate to their content. Its framing is that SEO best practices remain relevant because those AI features are rooted in Google's core Search ranking and quality systems. The mythbusting section exists because, in Google's phrasing, terms like Answer Engine Optimization and Generative Engine Optimization are common online while many of the suggested "hacks" are not effective or supported by how Google Search actually works.
That framing is the useful part for an agency. The question it settles is not "does AI search matter" — it does — but "is this a new discipline you have to buy".
You do not need an llms.txt file or special AI markup
Google's wording: you don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search, including its generative AI capabilities, because Google Search itself does not use them. It adds a clarification worth keeping, because it is the source of a lot of confusion: Google may discover, crawl and index many kinds of files beyond HTML, and that does not mean the file is treated in a special way.
So if a vendor demonstrates that Googlebot fetched the file they sold you, that demonstrates fetching. It does not demonstrate use.
You do not need to "chunk" your content
Google's wording: there's no requirement to break your content into tiny pieces for AI to better understand it, because Google systems are able to understand the nuance of multiple topics on a page and show the relevant piece to users.
The guidance it gives instead is deliberately unhelpful to anyone selling a template: sometimes shorter or longer pages work well depending on audience and subject matter, there is no ideal page length, and pages should be made for the audience rather than for generative AI search.
This one is worth flagging to clients who have been handed a "restructure every page into question blocks" proposal. Structuring content in question-and-answer form is often good writing — it is not a technical requirement, and the difference matters when someone is pricing a site-wide rewrite.
You do not need to rewrite content for AI systems
Google's wording: you don't need to write in a specific way just for generative AI search, because AI systems can understand synonyms and general meanings of what someone is seeking in order to connect them with content that might not use the same precise words. Its conclusion is explicit — you don't have to worry that you lack long-tail keywords or have not captured every variation of how someone might search.
For agencies this quietly retires a whole category of upsell: the keyword-variation expansion sold as AI readiness.
You do not need to chase mentions across the web
Google's wording: its generative AI features can show what is being said about products and services across the web, including blogs, videos and forum discussions, but seeking inauthentic "mentions" is not as helpful as it might seem, because core ranking systems focus on high-quality content while other systems block spam, and the AI features depend on both.
Read that as a description of the mechanism rather than a warning. Mentions matter as a signal of the thing; manufacturing them targets the signal.
Structured data is not the key to AI visibility
Google's wording is the most direct of the five: structured data isn't required for generative AI search, and there's no special schema.org markup you need to add. The same passage then says it is still a good idea to continue using structured data as part of an overall SEO strategy, because it helps with rich result eligibility.
Most of the advice circulating quotes only the second sentence. Both belong together: keep doing structured data, and stop selling it as the route into AI answers. What structured data does still earn on a local site is covered in Local Business Schema Markup: What Still Earns Its Keep.
What does Google say to do instead?
Its positive guidance is short enough to fit in a proposal, which is part of why it is hard to sell against:
- Create valuable, non-commodity content. Google's phrasing is that content people find unique, compelling and useful will likely influence a site's presence in generative AI search, and that recycling what others have already said — or what a generative model could easily produce — adds little unique insight.
- Keep a clear technical structure. Meet the Search technical requirements, follow crawling best practices, keep semantic HTML aimed at human readability, follow JavaScript SEO practices where relevant, provide a good page experience, and reduce duplicate content.
- Optimize local business details. For a local client this is the part with the most leverage, and it is ordinary local SEO work rather than anything AI-specific.
- Measure with first-party data. Use the Generative AI performance report in Search Console, and be wary of third-party tools that promise ranking success or claim to use internal Google metrics — Google states that no third-party tool has access to its internal ranking or AI systems.
There is also a hard prerequisite underneath all of it: a page must be indexed and eligible to be shown in Google Search with a snippet, fulfilling the Search technical requirements.
What this changes about how an agency sells
Two things, and neither of them is "stop caring about AI search".
The first is defensive. When a client forwards a pitch promising AI visibility through special files, chunking or schema, you now have a primary source rather than a counter-opinion. Linking a client to the vendor's own platform documentation ends that conversation faster than any argument you could construct.
The second is what remains sellable, and it is narrower and more honest: the ordinary work of being findable, plus a measurement record. Nobody can promise a mention in an AI answer. What can be delivered is a dated baseline, the specific pages published against the gaps it showed, and the next dated reading. The method for that baseline is in How to Test Local Business Visibility in AI Answers.
Frequently Asked Questions
Does this mean AI search does not matter for local businesses?
No. It means the work that makes a business visible in AI answers on Google is largely the work that makes it visible in Google Search, because Google states those features are rooted in its core ranking and quality systems. What is new is the measurement surface, not a separate discipline.
Is llms.txt useless?
For Google Search, Google states it does not use such files. Other AI providers may make their own choices, and any claim about them should come with that provider's documentation attached rather than being generalized from Google's position.
Should we stop writing FAQ sections on client pages?
No. Writing in question-and-answer form is often the clearest way to answer a real question, and clarity is what the positive guidance asks for. What Google says you do not need is chunking content as a technical requirement for AI systems.
Does any of this apply to assistants other than Google's?
Not automatically. This guidance describes Google Search and its generative AI features. Any claim about a different assistant needs its own source, and most of the confident advice circulating about those assistants does not have one.
Where can I read the original?
Google's guide to optimizing for generative AI features on Google Search, linked in the references below. The mythbusting section is a named heading inside it, so it is easy to send to a client directly.
References
- Google Search Central. "Google's Guide to Optimizing for Generative AI Features on Google Search." developers.google.com
- Google Search Central Blog. "A new resource for optimizing for generative AI in Google Search." developers.google.com
- Google Search Central. "AI Features and Your Website." developers.google.com
- Google Search Central Blog. "Introducing Search Generative AI performance reports in Search Console." developers.google.com