Google’s John Mueller was asked a direct question on Bluesky: are there entire industries where Generative Engine Optimization simply doesn’t matter yet? His answer cuts through a lot of the noise that has built up around GEO as a discipline — and it’s worth unpacking what it actually means for how sites should be prioritizing their work.
The question that started it
Bluesky user @galloni.net asked Mueller a pointed question, wondering aloud whether categories like adult content, gambling, or other verticals where Google still seems to dominate discovery might be exempt from needing any GEO-specific strategy. It’s a reasonable thing to wonder, given how much marketing language around GEO implies that AI search requires an entirely separate playbook from traditional SEO.
Mueller’s reply was almost anticlimactic in its simplicity: “I’m not quite sure what you’re asking; from our POV there’s nothing really special you need to do for generative AI responses in search.” He wasn’t dodging the question — he was rejecting its premise. There isn’t a subset of industries where GEO doesn’t matter, in his framing, because there isn’t a version of “GEO” that exists as a separate discipline from Google’s perspective at all.
Why Google’s AI answers aren’t a separate system
The reasoning behind Mueller’s answer comes down to how Google actually builds its AI-generated search results. Those results are pulled from the same crawled index that powers regular organic search. When a query triggers an AI Overview or a similar generative response, Google draws from the top-ranking pages for that query, plus results gathered through what’s known as query fan-out — a process where the system generates a batch of related sub-queries to fill in context, cover related angles, and surface information a searcher is likely to want next.
None of that machinery is separate infrastructure sitting apart from the main index. It’s the same crawl, the same ranking systems, and the same signals that have always determined organic visibility, just assembled and summarized differently at the point of display. If a page doesn’t rank well in traditional search for a given topic, it has no special back door into being cited in an AI-generated answer either. The inputs are shared.
That’s the core of why Mueller could answer a question about specific industries with a blanket statement instead of a nuanced, sector-by-sector breakdown. If AI search results are just a repackaging of the existing index plus fan-out queries, then the entire premise of certain industries being “exempt” from GEO doesn’t hold up — because there was never a separate GEO system to be exempt from in the first place. It’s not that adult content or gambling sites don’t need optimization; it’s that the optimization they need is the same organic SEO work every other industry needs.
The gap between market demand and technical reality
What makes this exchange notable isn’t just the technical answer — it’s the contrast between that answer and how aggressively GEO has been marketed as a distinct service category over the past couple of years. Businesses have been actively asking for help ranking in AI-based search results, and where there’s demand with budget attached, an industry tends to spring up to meet it. Consultants who specifically brand themselves around AI search optimization have multiplied, and even SEO professionals who understand the underlying mechanics have started offering “GEO services” as a line item, often while privately acknowledging that the work is just SEO wearing a new label.
One visible symptom of this trend is the sudden popularity of LLMs.txt, a proposed file format meant to point large language models toward markdown-formatted, “LLM-optimized” content. Most major SEO plugins now include a feature to generate one automatically. The catch is that no search engine and no major AI chat product actually reads or uses LLMs.txt as an input for indexing or answer generation. It’s a file that satisfies a demand for something tangible to point to as “AI optimization,” even though it does not currently affect anything about how a site is crawled, indexed, or cited. Mueller and other Google representatives have made similar points about llms.txt in the past — the format doesn’t help large language models differentiate one site from another in any way that changes outcomes.
What this means for prioritizing your work
The practical takeaway isn’t that AI search visibility doesn’t matter — it clearly does, and getting cited in an AI Overview or a chatbot’s response can drive meaningful traffic and brand exposure. The takeaway is about where to spend effort. If AI-generated answers are sourced from the same index that ranks pages in regular search, then the fastest way to improve AI search visibility is to improve the fundamentals that were always going to matter: technical crawlability, content that directly and clearly answers real user questions, credible signals of expertise and trustworthiness, and a site structure that makes it easy for both crawlers and readers to find the right page.
That doesn’t mean every AI-specific tactic is worthless — testing how your content gets summarized, understanding which pages tend to get pulled into fan-out queries, and paying attention to how your brand shows up in AI-generated citations can all be useful diagnostic work. But treating GEO as a separate strategic track, with its own budget line and its own specialist team disconnected from core SEO, misreads how the underlying system actually works, at least according to the engineer who helps build it.
This also has implications for how agencies and in-house teams should evaluate vendor pitches. A consultant or tool promising GEO-specific ranking factors, GEO-exclusive audits, or GEO deliverables that don’t map back to recognizable SEO fundamentals is selling a narrative more than a mechanism. The safer bet, based on how Google itself describes the pipeline, is to treat strong SEO as the whole strategy and treat AI search visibility as one of its downstream outcomes, not a parallel discipline requiring its own separate investment.
Frequently asked questions
Does GEO exist as a separate discipline from SEO?
According to Google’s John Mueller, no — at least not from Google’s perspective. AI-generated search results are built from the same crawled index and ranking systems used for regular organic search, so there isn’t a distinct set of “GEO” ranking factors sitting apart from standard SEO fundamentals.
Are there industries where AI search optimization doesn’t apply?
Mueller’s answer implies no. Since AI answers are sourced from the same index across every vertical, including sectors like gambling or adult content where Google still drives most discovery, there’s no special category of exemption — the same SEO fundamentals apply everywhere.
What is query fan-out?
Query fan-out is the process Google’s AI search systems use to generate a batch of related sub-queries around an original search, pulling in context and related information from across the index to build a more complete generative response.
Does LLMs.txt actually help AI search rankings?
No. LLMs.txt is a proposed file format meant to guide large language models to markdown content, and many SEO plugins now generate it automatically. However, no major search engine or AI chat product currently uses it as an input, meaning it has no measurable effect on visibility despite its popularity.
Where should sites focus their effort instead of separate GEO tactics?
On the same fundamentals that drive organic search performance: technical crawlability, clear and directly useful content, credible expertise signals, and a site structure that surfaces the right pages, since AI-generated answers draw from that same underlying index.
For related coverage on how Google’s crawling and indexing decisions affect both organic and AI-driven visibility, see how Google made it easier for SEOs to remove spammy sites from competing indexes, why some SEOs split their sitemaps into multiple files, and how Google may ignore robots.txt in ways that affect SEO. For more on the opt-out mechanics shaping AI search specifically, read about how Google gives sites an AI search opt-out without the data to use it and why Mueller says llms.txt can’t help LLMs differentiate between sites.