Google Just Told You How to Rank in AI Search. Most "GEO Experts" Won't Like It.

I've been saying this for a while now. The whole GEO thing felt off. Too many consultants selling "AI search optimization" packages that sounded suspiciously like regular SEO with a new price tag.

Turns out Google agrees with me.

They just published their first official guide on optimizing for generative AI features in Google Search. And the core message is almost comically simple: "Optimizing for generative AI search is optimizing for the search experience, and thus still SEO."

That's it. That's the guide. I mean, there's more to it, but that one sentence should be printed on a t-shirt and mailed to every GEO consultant who charged someone $5,000 for a "generative engine audit."

Google Finally Said the Quiet Part Out Loud

Let me set the scene. Over the past year, an entire cottage industry popped up around terms like AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization). Consultants, courses, tools, agencies, all promising to crack the code of how to show up in AI-generated search results.

Google's guide acknowledges these terms exist. Then it politely says: from our perspective, this is still SEO.

Not a new discipline. Not a paradigm shift. Not something that requires a separate budget line or a new agency retainer. Just SEO. The same SEO that's been working for the past two decades, with some technical context about how their AI features actually pull information.

The Reddit community caught on immediately. One thread summed it up: "Google have released the first, official AI SEO guide and the ONLY guide for AI SEO, GEO, AEO of any of the LLM/AI vendors." And the reaction was a mix of validation and schadenfreude.

What the Guide Actually Says (The Short Version)

Google explains two key technical concepts that power their AI features:

Retrieval-Augmented Generation (RAG): This is how AI Overviews and AI Mode find information. The AI doesn't have its own separate index. It pulls from Google's existing search index, the same one that powers regular search results. So if your page ranks well in traditional search, it's already in the pool for AI features.

If you want to understand RAG in depth, I've written about it before. The short version: the AI retrieves relevant pages, then generates a response based on what it found. No separate database. No special AI index. Your regular rankings feed directly into this.

Query Fan-Out: When someone asks something like "how to fix a lawn full of weeds," the AI doesn't just search that one query. It fires off multiple related searches simultaneously: "best herbicides for lawns," "remove weeds without chemicals," "how to prevent weeds in lawn." This means your page can get pulled into an AI response even if it doesn't directly match the original query, as long as it's relevant to one of the fan-out queries.

The implication is straightforward. If you're doing solid SEO and ranking for relevant terms, you're already optimized for Google's AI features. There's no separate game to play.

The Mythbusting Section (This Is Where It Gets Fun)

This is the part that should make a few people uncomfortable. Google included a section explicitly titled "Mythbusting generative AI search: what you don't need to do." Here's their list:

You don't need llms.txt files. Google says you don't need to create "new machine readable files, AI text files, markup, or Markdown" to appear in generative AI search. I wrote a whole post about llms.txt when it was gaining traction. Google just said it doesn't matter for their AI features. That stings a little, honestly.

You don't need to "chunk" your content. No requirement to break content into tiny pieces for AI comprehension. Google's systems can understand multiple topics on a single page and surface the relevant piece. The entire "you need to structure your content in AI-friendly chunks" narrative? Not a thing, according to Google.

You don't need to rewrite content for AI systems. This one's important. AI systems understand synonyms and general meaning. You don't need to stuff every long-tail keyword variation into your content. You don't need to "write for AI." Write for humans.

You don't need inauthentic mentions. Some GEO services promise to get your brand "mentioned" across the web to boost AI visibility. Google's core ranking systems focus on high-quality content, and their spam systems filter the rest.

You don't need excessive structured data. No special schema.org markup for AI search. Structured data is still useful for rich results, but it's not a magic ticket into AI Overviews.

Every single item on this list is something that someone, somewhere, has been selling as a GEO service. Every. Single. One.

Wait, No EEAT?

Here's the thing the Reddit community noticed that I think is the most interesting detail: Google's guide doesn't mention E-E-A-T once.

Not once.

For the past few years, EEAT (Experience, Expertise, Authoritativeness, Trustworthiness) has been the SEO industry's favorite framework. Consultants built entire strategies around it. "You need to demonstrate EEAT for AI search" became a standard pitch.

But Google's official AI optimization guide doesn't bring it up at all. Why?

My take: EEAT is a quality rater concept. It's how Google trains human evaluators to assess search results. It's not a direct ranking signal, and it never was. The guide focuses on actionable technical recommendations, not abstract quality frameworks. Google didn't need to say "make good content" because the entire guide is built on that assumption.

But the absence is telling. It undermines the narrative that EEAT is some special lever you need to pull for AI visibility. If it were critical, you'd think Google would mention it in their first official AI optimization guide.

What GEO Experts Have Been Selling You

Let me be direct about what's been happening in this space.

Over the past year, I've seen GEO services offering:

• Special llms.txt file creation and optimization (Google says: not needed)
• "Entity optimization" and knowledge graph manipulation (overstated at best)
• AI-specific content formatting and restructuring (Google says: write for humans)
• "Citation architecture" strategy (which is just... having good content with a fancy name)
• AI search audits that recommend all of the above for $2,000-$10,000

Fast Company called GEO "a mirage." Their reporting found that brands only control 5-10% of what AI search actually references about them. The rest comes from third-party sources, reviews, forums, and content that no GEO consultant can control.

I wrote about this when GEO was first gaining traction. My take hasn't changed: the fundamental advice was never wrong (create good content, be findable), but the packaging and pricing were designed to exploit uncertainty.

Google's guide just made that exploitation a lot harder to justify.

But What About ChatGPT and Perplexity?

Here's where I need to add some nuance, because Google's guide only covers Google. And the AI search landscape isn't just Google anymore.

The citation patterns across platforms are genuinely different:

• ChatGPT pulls 47.9% of its citations from Wikipedia and 11.3% from Reddit. It heavily favors high-authority, established domains.
• Perplexity is the opposite: 46.7% of citations come from Reddit. It favors fresh, specific, community-sourced content.
• Google AI Overviews sit in between, with about 21% Reddit citations and a strong preference for pages that already rank well in traditional search.

So Google saying "just do SEO" is accurate for Google. But if you want to show up in ChatGPT's answers, the game is different. Wikipedia presence, third-party reviews, and brand mentions matter more. If you want to understand what ChatGPT actually searches for, it's worth looking at that separately.

The honest answer: there's no single optimization strategy that works across all AI search platforms. Google's guide simplifies things for their ecosystem, which is helpful. But don't assume it applies everywhere.

The Real Takeaway (It's Not "Do Nothing")

I don't want to leave you with the impression that Google's guide says "relax, do nothing." It doesn't. It's actually pretty specific about what works:

Technical SEO fundamentals. Make sure your site is crawlable, meets Core Web Vitals, uses semantic HTML where possible, and provides a good page experience across devices. The basics haven't changed.

Create non-commodity content. This is the phrase Google uses, and I love it. "Commodity content" is stuff like "7 Tips for First-Time Homebuyers," content that could come from anyone and adds nothing unique. Non-commodity content provides a unique perspective, first-hand experience, or expert insight that isn't available elsewhere.

Reduce duplicate content. Having the same content across multiple URLs wastes crawl budget and confuses systems.

Optimize your local and ecommerce presence. Google Business Profiles, Merchant Center feeds, product listings. These feed directly into AI responses where relevant.

Don't over-optimize. Google explicitly warns against creating separate pages for every query variation. Their AI systems understand relevance and synonyms.

Here's what I find refreshing about this guide: it doesn't introduce any new hoops to jump through. It's Google saying, "The thing that worked before? It still works. Stop chasing shortcuts."

And here's what I find honest about it: AI Overviews appear in roughly 25.8% of US searches. They reduce organic click-through rates by 15-46% depending on the query. The zero-click rate hits 83% when an AI Overview shows up. But sites that get cited in those overviews can see CTR increases of up to 35%.

The traffic is shifting, not disappearing. The question isn't whether to optimize for AI search. It's whether you need a new strategy for it, or whether the old one still applies.

Google just answered that question pretty clearly.

If you're doing solid SEO, creating genuinely useful content, and keeping your technical house in order, you're already doing AI search optimization. You just didn't need to call it GEO or pay someone $5,000 to tell you that.

Need help figuring out where your site actually stands? Let's talk about an honest SEO audit, no GEO upsells, I promise.