GEO Is the New Snake Oil (Or Is It?)

Last week a client forwarded me a pitch from a "GEO agency" offering to "optimize their AI search presence" for $3,000 a month. The pitch included terms like "generative engine optimization," "answer engine optimization," and "LLM visibility enhancement."

Three thousand dollars a month. For something that, as far as I can tell, nobody can reliably prove works yet.

Here's the thing: the underlying concern is real. AI is changing how people find information. But the industry that's sprung up around it smells familiar. Very familiar.

Another Three-Letter Acronym to Sell You Something

Generative engine optimization (GEO) is the practice of optimizing your content to get cited in AI-generated answers. ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude. When someone asks an AI a question and it pulls from your website, that's what GEO is supposedly about.

The concept was formalized in a 2024 paper from Princeton and IIT Delhi. Solid academic work. But then the marketing machine got ahold of it.

Now we have GEO, AEO (Answer Engine Optimization), LLMO, GSO, AIO. Five different acronyms for the same thing. Sound familiar? Every time a new channel emerges, an entire industry of consultants materializes overnight. It happened with voice search optimization (remember that?). It happened with social media SEO. And now it's happening with generative engine optimization.

The pattern is always the same: take a real shift in how search works, package it as a new discipline with its own acronym, and sell it as something separate from what good SEO already covers.

The Numbers That Make GEO Sound Urgent

Let me be fair. The numbers behind the AI search shift are not made up. They're genuinely significant.

ChatGPT has 800 million weekly users in 2026. Google Gemini has 750 million monthly users. AI Overviews now trigger on 48% of all Google searches, up 58% year over year. AI-referred sessions grew 796% between 2024 and 2025.

The scary part: organic CTR dropped 61% on queries where AI Overviews appear. Gartner projects organic search traffic to commercial sites will decline 25% by 2026.

These numbers are real. The panic is understandable.

But panic is also profitable. For whoever's selling the solution.

Fewer than 12% of marketing teams have a documented AI search strategy. That gap between "this is clearly important" and "nobody knows what to do" is where snake oil thrives. When people are scared and confused, they buy certainty. Even if that certainty doesn't exist yet.

What GEO Agencies Actually Promise (and Charge For)

The typical GEO service package includes structured data optimization, llms.txt file creation, "AI-friendly" content reformatting, citation building, and brand mention seeding.

The Princeton study claimed that content with citations, statistics, and direct quotations can improve AI visibility by 30-40%. That's the number every GEO agency puts in their pitch deck.

Some agencies claim "double-digit citation rates within 60 days." Tools like Otterly, AthenaHQ, and HubSpot's AEO Grader have popped up. There's even a GEO Conference in 2026. It's already a full industry.

Here's my problem with most of this: the tactics they're selling are just good content practices with a new label. Include real data in your writing? That's always been good advice. Structure your content clearly? That's been SEO 101 since 2010. Use authoritative sources? Obviously.

Repackaging fundamentals as a separate discipline to charge separately for it is where this crosses from legitimate concern into sales pitch. If your SEO consultant isn't already thinking about how AI search affects your content, you don't need a GEO agency. You need a better SEO consultant.

The llms.txt File That Does Nothing

This is the part that should make everyone pause.

SE Ranking analyzed nearly 300,000 domains to test whether having an llms.txt file (a file that tells AI crawlers what your site is about) actually helps you get cited by AI systems.

The result: no significant correlation. None.

Only 10.13% of the analyzed domains had an llms.txt file. Adoption was nearly identical across traffic levels: 9.88% for low-traffic sites, 10.54% for mid-traffic, 8.27% for high-traffic. So much for the idea that "big sites know something you don't."

Here's the kicker: when SE Ranking removed the llms.txt feature from their prediction model, the model's accuracy actually improved. The file wasn't just neutral. It was noise.

I wrote about what llms.txt is and how it works last year. The concept makes sense in theory. In practice, at least right now, the data says it doesn't move the needle.

This is the most rigorous study we have on a specific GEO tactic. And it says the tactic doesn't work. That should make you question every other GEO claim that comes without data.

What Actually Gets You Cited (It's Boring)

So if the fancy GEO tactics don't work, what does?

First, you need to understand that AI platforms cite very differently. Perplexity cites sources 97% of the time. Google AI Overviews cites sources 34% of the time. ChatGPT cites sources only 16% of the time. So "optimizing for AI citations" is not one strategy. It's a different challenge on every platform.

The content types that get cited most: listicles (21.9%), articles (16.7%), and product pages (13.7%). The most cited domains across all AI platforms? Wikipedia and Reddit. Not exactly sites known for their GEO budgets.

Here's the most useful data point I found: 44.2% of all LLM citations come from the first 30% of your text. Your intro matters more than your conclusion. If your key information is buried at the bottom of a 3,000-word post, AI systems probably won't find it.

Content depth, readability, and freshness rank higher than traditional SEO metrics like backlinks for AI citation purposes. And when your brand IS cited in an AI Overview, organic CTR is actually 35% higher than when it isn't.

The boring truth: write clear, well-structured, authoritative content. Include real data. Be a genuine source. Put your best information up front. This has always been the answer. The shift toward AI search doesn't change the fundamentals, it just makes them more important.

What I'm Actually Telling My Clients

When clients ask me about generative engine optimization, here's what I say.

Don't pay for standalone GEO services. If your SEO consultant isn't already factoring AI search into their strategy, that's a red flag about the consultant, not a reason to hire a second agency.

Focus on being a genuine authority. Not gaming citation algorithms. Write content with real data, real expertise, real specificity. AI systems are getting better at distinguishing genuine authority from optimized content. Bet on the trend, not against it.

Track your AI visibility, but don't obsess. Tools like Otterly exist for monitoring where you show up in AI answers. Worth checking. Not worth losing sleep over.

Put your best information in the first 30% of every page. That's where 44% of citations come from. Lead with your insight, not your preamble.

Don't bother with llms.txt yet. 300,000 domains. No correlation. The data is clear.

Keep doing good SEO. The overlap between "what gets you ranked on Google" and "what gets you cited by AI" is enormous. Good content, clear structure, genuine expertise. The channel is changing. The fundamentals aren't.

Is GEO Snake Oil? My Honest Answer

Not entirely. The underlying concern is legitimate. AI is genuinely changing how people discover information. 800 million weekly ChatGPT users and AI Overviews on 48% of searches are not numbers you can ignore.

But the packaging? The three-letter acronyms, the $3,000/month retainers, the conference circuit, the tools promising "AI visibility optimization"? That's the part that smells like snake oil. Taking good content practices that have always worked, rebranding them as a new discipline, and charging a premium for the rebrand.

I keep thinking about that Reddit post from an SEO practitioner who wrote: "Six months into GEO work and I still can't figure out what's actually moving the needle." That person is probably doing fine. Not because GEO doesn't matter, but because nobody can isolate what moves the needle yet. The field is too new, the data is too thin, and the platforms change too fast.

I don't have all the answers here either. AI search is a moving target. Anyone who tells you they've cracked the code for generative engine optimization is selling something. The honest answer is that we're all figuring this out in real time.

What I do know: the businesses that will come out ahead are the ones that focused on being genuinely useful, genuinely authoritative, and genuinely worth citing. Not the ones that paid someone to game a system that hasn't finished being built yet.

Want to make sure your SEO strategy accounts for AI search without overpaying for acronyms? Let's talk about what actually makes sense for your business right now.