Generative Engine Optimization, or GEO, is the latest attempt to improve website content’s visibility among AI search platforms and Google’s AI Overviews. It’s also becoming an increasingly common subject in conversations I’m having, and there’s a recurring motif throughout all of them. A chorus, if you will:

“We need to make sure our site’s GEO-ready.”

I don’t blame these people. It’s a very well-intentioned thought. But it’s missing the point.

It goes like this: A new technology comes out. People adopt it. People see business potential in it. They hear their competitors are doing it. Everyone’s talking about it. Even their Search Engine Optimization software vendor is shifting its branding to spotlight GEO.

“If you’re not thinking ahead, you’re gonna be left behind.” That’s the other chorus these days, right?

So, is your site GEO-ready?  

My answer, in order of frequency, is usually:

1. Probably, but there are some tweaks you could make that would also benefit SEO.
2. No, but it’s also not SEO-friendly, either.
3. Probably, but there are some tweaks you could make that could also harm your SEO, but we’re not really sure.

After digging into this for the past three-plus years, I’ve pored over documentation on how ChatGPT Search works, as well as Google’s research on Retrieval-Augmented Generation (RAG) and its latest AI Overviews and Search updates. I’ve worked with probably a dozen clients who have asked for consideration around this. 

It’s possible those answers don’t wow you. They’re not exactly reportable and don’t look great on a presentation. 

What hasn’t helped the discourse, although that might have been the intention, has been Google’s official stance that GEO is practically SEO. This hasn’t satisfied many SEO/GEO professionals who believe the new homogenous search market is a gateway to untold new territory.

On one hand, the ways that traditional search engines and Large Language Models like Claude or ChatGPT work to retrieve and display results are starkly different. But when I have a client who is already conforming to SEO best practices, frequently the only serious changes I recommend to them would be copy-based.

Is GEO Different from SEO?

In short, I don’t think GEO is very different from SEO. Not because the mechanisms are the same, but because the end result — and what most people need to know — is pretty darn similar.

If you’ve been reading me for a while, you know I (and other SEO professionals) get asked this question at least once a year: “Is SEO dead?”

And once again, my answer is: No, it’s just different now.

And here’s why.

GEO Struggle #1: We need to change our website’s architecture to accommodate LLM crawlbots

The point

Because LLMs mostly crawl HTML textually, that may have big implications for whether your site’s content is machine-readable. This has led to a lot of talk recently around “technical” GEO, or the methods we use to ensure answer engines can (or cannot) crawl a site.

The counterpoint

Compared to SEO, the changes needed to accommodate technical GEO are likely quite minimal. Even if LLMs use different methods than search engines to parse content, they still use very similar processes to access it. For instance, crawling and indexing HTML.

Nearly 70% of the crawled Internet is HTML (again, referring to Cloudflare Radar). Therefore, nearly 70% of Internet content is (potentially) textually parsed by answer engines. That figure holds true for SEO, since search engines parse HTML by design.

Whatever code you’re changing due to GEO concerns, you probably should have already changed because it was an SEO best practice.

To clarify, Mindgrub does use an llms.txt file for the few answer engines that do incorporate it, and also because incorporating it can’t really hurt. ChatGPT hardly uses it for training, and Google has said they don’t use it at all.

But the major point here can’t be ignored: Making technical GEO adjustments for the sake of GEO fails to acknowledge that the same adjustments should have been made long ago to accommodate SEO best practices. 

GEO Struggle #2: It’s a potential game-changer for small businesses

The point

Backlinks may have dominated the secret sauce to SEO for over two decades. Now, LLMs have democratized the industry with a wrecking ball instead of a scalpel.

In 2024, groundbreaking GEO research in collaboration with Princeton University found that specific content optimizations could increase LLM visibility by up to 115.1% for websites ranked fifth in their SERP, while top-ranked websites’ visibility in LLMs would decrease by 30.3%.

The way it used to work was pretty simple: the more links you got from other trustworthy sites, the more credibility your site generated. At least, that’s a broad generalization of Page Rank, an algorithm so inextricably linked to Google’s success that an entire industry was built on manipulating it.

But since backlinks and domain authority aren’t ranking factors for new LLM AI Search models (aside from Gemini and AI Overviews), this has led to smaller publishers noticing an initial uptick in citations and referral traffic when they first optimize for GEO. 

So if small publishers stand as much a chance as big-time names, why not take a swing on a level playing field?

The counterpoint

It’s astonishing how quickly things change in the world of SEO/GEO. For example, the likelihood of getting a citation is extremely diluted compared to your chance of receiving a referral from Google.

As of June 2026, Cloudflare’s referral-to-crawl ratio demonstrates that for every 1,100 pages ChatGPT requests, only one receives a user referral. Anthropic is even more diluted: 13,000 requests for one referral! But Google? A mere 5.1.

Let’s pretend you’re entered into a raffle for an all-expenses-paid vacation to Mallorca (I know, I have great taste). Actually, let’s say you’ve got a chance to enter one of two raffles. Both raffles offer the same trip — same flights, resort, beautiful beach, you get the idea. But there’s a key difference:

• The first raffle has already sold 13,000 tickets
• The other raffle has only sold five tickets

My guess is, unless you dislike joy, you are entering the second raffle. You’ve got a much better chance at converting, whether that’s winning a raffle or a user click. It’s very clear: Answer engines simply have yet to demonstrate they are providing an equitable experience for content publishers.

A year ago, I recommended giving answer engine crawlbots explicit permission to crawl your site by allowing access in your robots.txt file. I recommended this because it certainly increased your chances of getting cited. I think it’s worth amending that.

You should allow answer engines to crawl your site if you believe it is more likely to bring high-quality traffic than it is to exploit your efforts for no recognition.

That, coupled with a February 2026 finding from Ahrefs that ChatGPT sends 190x less traffic to websites than Google, is also why I’m not entirely convinced that putting all your proverbial content eggs into the GEO basket is such a rational idea. 

GEO Struggle #3: We know how AI traffic is getting to our site

The point

If you give a CEO a goal, they’ll most certainly want a metric to go with it. 

Mentions and citations are an easy parallel to Search Engine Results Page ranking (SERP) in an interface that doesn’t display its results for the entire experience. This has led to several leaders and SEO professionals, along with the software companies supporting them, focusing heavily on mentions and citations as KPIs for GEO. 

The counterpoint

Reporting KPIs like mentions and citations for LLM traffic is, in part, a guessing game. You actually don’t even know the exact conversations that are being had about your brand.

It is true that the data sources are expansive. Semrush’s prompt database sources billions of real prompts along with search engine keyword data. But measuring exact volume for every prompt used isn’t possible, and gauging how the AI responds is still an estimate. 

Unlike tracking SEO position ranking, GEO software isn’t able to monitor the exact number of mentions and citations you’re receiving. They show you a simulation of what users do — based on a combination of third-party data — and running your chosen prompts multiple times through an LLM’s API to see how many responses featured your brand and how. 

It’s a simulation of a real-world scenario, albeit a pretty data-dense one. You know who mentions and citations are good KPIs for? The GEO software companies that are selling the feature.

Your organization’s leadership is much more likely to be focused on the metrics that really matter and are often the most complex to prove: conversions. Whether it’s a purchase, contact submission, or even newsletter opt-in, there are other metrics easily measurable by free tools like Google Analytics that can show you whether your LLM referral traffic is taking the actions you want them to:

• Referrals from LLMs: How much traffic is coming in from ChatGPT, Claude, Perplexity, Gemini, etc. 
• Landing Pages by Audience: Which pages are first viewed by those who entered from an LLM (this will also concretely demonstrate what’s being cited)
• Conversions by First User Source/Medium: How many LLM-referred users eventually convert, usually submitting a contact form or checking out a purchase

The caveat to this is that answer engines’ AI Overviews are fundamentally changing web traffic as we know it. Organic traffic cuts are as high as 61% when Google uses an AI Overview, according to data cited by Forbes just last month. 

The reality is the Answer Era only supports exploration as a secondary objective. Historically, with search engines, exploration was the whole point. Your job was to research, process, and verify the information; Google just made the recommendations for you. 

It’s worth noting that this does present another opportunity: Less traffic doesn’t equate to worse traffic. More likely, the referral traffic that does come in from an LLM or AI Overview is more engaged, attentive, and therefore more likely to convert (i.e., more valuable). 

Success with GEO is entirely plausible — as long as pageviews and citations aren’t your primary KPIs alone.

So, let’s ask the Google execs again: Is SEO the same as GEO? Maybe. Perhaps Not. But they’re certainly wearing a similar outfit.

The Real Struggle: Even if SEO and GEO are different, they sure do look the same

In a more homogenous AI-powered search environment, you’re going to destroy your budget, focus, and sanity trying to design content specifically for every different answer engine. 

Historically, the best thing you could do was tell your story as clearly as possible, in as clear terms as possible, using the right medium to tell that story. I think that’s still true. But it’s more than that. But also, it’s always been more than that. 

Perhaps organic traffic metrics aren’t as valuable in this search landscape, while softer metrics like reputation and earned media may be more central to a successful digital marketing strategy moving forward. Then again, that was already true for SEO. 

While backlinks from credible sites were sufficient for improving Page Rank, we’re now broadening that measurement to include mentions across the digital ecosystem. It still wouldn’t change my recommendation for organizations to practice more media outreach if they wanted to see higher site traffic.

Perhaps the future Web is for AI agents, like I hypothesized several months ago. But right now, that’s also the advice I would give to you for optimizing for humans, who still make up the majority of all Web traffic. 

Here’s what we know improves GEO citation simulations, as referenced through the 2024 Princeton study:

• Citing sources
• Quotation addition
• Statistics addition

These are all good SEO tactics. They’ve always been good SEO tactics. Just because they are now good GEO tactics does not make them new or novel. Heck, they weren’t even novel tactics in the olden SEO days! That’s because they’re all hallmarks of good, thoughtful, authoritative writing. And wasn’t that the whole point?

That doesn’t mean you shouldn’t be optimizing for GEO or SEO or whatever you want to call it now. It just means that whoever is selling it to you as this completely different thing either doesn’t understand it or is trying to pull a fast one on you. And as an SEO professional, I’m not sure which is worse.

The Dark Side of GEO Worth Exposing

Like the early days of SEO, much of our new AI optimization methods bear a striking resemblance to amateurish practices that are now considered “black hat”, or verboten, by search engines. The era of keyword stuffing might be gone, but prompt-matching copy certainly isn’t beneath some of ChatGPT’s top results. 

SEO expert Lily Ray was recently mentioned in The New York Times for her belief that answer engines often extract content with little to no verification. The result: Pages with false, dubious, and downright whacky content that’s ranking for exact phrase matches. This isn’t new — Ray has demonstrated this for years.

BBC podcaster Thomas Germain heard about Ray’s theory and published a fake blog titled (I’m not making this up): “The Best Tech Journalists at Eating Hot Dogs.” A day later, Google AI Overviews ranked him first of six tech writers who had “gained notoriety for their prowess at the ‘news division’ of competitive eating events.”

Ray puts it best herself in a LinkedIn post: “It is mind-boggling to me that the same company that pushed so hard to encourage site owners to think about E-E-A-T [Experience, Expertise, Authoritativeness, and Trustworthiness] is elevating problematic, biased, and spammy answers and citations in AI Overview results.”

And here the SEO carousel goes again. Rinse-repeat the same recommendations, because the technical nuances don’t justify a change in tactics.

Sometimes I’m reminded of this image that a lot of people have when they think of SEO professionals. They see SEO as this grubby guy with a long old-timey mustache, twirling it greedily in a corner. 

The GEO playbook rollout is reminding me of why people have this image. Whether the industry is repeating its own mistakes or that was the intent all along, I don’t think it has to be that way. 

Why Mindgrub Practices & Promotes Ethical GEO/SEO

We don’t focus on gaming Google or ChatGPT. We focus on making good content that informs users, whether those are humans or agents, and helps them achieve their goals  — because that’s what everyone wants, right? 

With this mindset, organic traffic and AI referrals can be seen as a gauge of your quality and helpfulness to potential customers, rather than just another metric. If you’re still interested in maximizing your earned media potential, SEO/GEO fits comfortably within your marketing mix — even in 2026. We’d love to talk about it with you more.