I didn’t set out to build an agency.

I was spending hours — more than I’d like to admit — jumping between ChatGPT, Perplexity, and Gemini, asking the same questions in different ways, then cross-referencing where each platform was pulling its information from. Not for a client. Just because I was genuinely curious about the mechanics.

And the more I dug, the more I noticed something that should have been obvious but wasn’t: the brands showing up in AI answers weren’t necessarily the best brands in their category. They weren’t always the highest-ranked on Google. They weren’t the ones with the biggest ad budgets.

They were the ones that happened to have the right signals in the right places.

That observation turned into an obsession. That obsession turned into Cited.

The Construction Company That Changed How I Think About This

Early on, I ran a citation audit for a company in the civil construction space. Decent-sized firm. Established reputation in their market. The kind of business that’s been around long enough that they assume people know who they are.

I ran a set of queries across the main AI platforms — the questions their potential clients would genuinely ask. Things like: “Which civil construction firms specialize in infrastructure projects in the Southwest?”

They appeared almost nowhere.

What struck me wasn’t the result itself. It was their reaction when I showed them. They had no idea this was even a thing to think about. They hadn’t considered that a potential client might open ChatGPT before picking up the phone. The concept of AI recommending — or not recommending — their business simply hadn’t entered their world yet.

That’s the gap I keep seeing. It’s not a technology gap. It’s an awareness gap. Most businesses have no idea that AI tools are making recommendations about their industry every single day, and that they have almost no presence in those recommendations.

What GEO Actually Is

When I describe what we do at dinner parties — not the pitch version, just the honest version — I say this:

We make sure that AI recommends your business when potential clients search for your niche using AI.

That’s it.

The formal name is Generative Engine Optimization. GEO. Researchers at Princeton University and Georgia Tech coined the term in 2023 after running 10,000 test queries through AI systems. They found that specific optimization techniques can boost your visibility in AI-generated answers by up to 40%. The signals that drive this — original data, structured content, entity authority, cited sources — are different from what drives Google rankings.

Only 6.82% of ChatGPT results overlap with Google’s top 10. These are genuinely separate channels with different rules.

The simple version: SEO gets you ranked on Google. GEO gets you cited in AI answers. Same goal — be found by buyers — completely different mechanics.

The Argument I Have Constantly

Here’s where I’ll lose some people, and I’m fine with that.

A lot of marketers hear “GEO” and immediately slot it into their existing mental model of SEO. They assume the process is similar, the timeline is similar, and therefore the pricing should be similar.

It isn’t.

SEO at this point is a mature discipline with decades of tooling, established benchmarks, and a well-understood playbook. You can buy SEO tools for $99 a month and execute a reasonable strategy yourself. The ceiling for a basic SEO engagement is well-known.

GEO is a different animal. The methodology is newer, the platforms are less predictable, the entity work is more technical, and the measurement is more complex. Tracking your Share of Model Score across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews — running structured query sets, scoring sentiment, tracking competitive citation share — this is not the same as checking keyword rankings in Ahrefs.

When agencies price GEO like SEO, they’re underdelivering. They’re applying SEO frameworks to a problem that needs a different framework entirely. And clients who expect SEO timelines and SEO results from GEO work end up frustrated — not because GEO doesn’t work, but because it was set up incorrectly from the start.

And then there’s the Google problem.

In 2026, Google’s official position is that GEO and SEO are the same thing — at least for Google AI Overviews. Their stance: the same ranking systems that determine organic search results also determine what appears in AI Overviews. Do good SEO, get cited in AI. Simple.

I understand why Google says this. It protects their existing ecosystem. It keeps brands investing in Google-centric strategies. And for Google AI Overviews specifically, it’s not entirely wrong.

But it’s only true for Google.

Perplexity doesn’t use Google’s ranking systems. ChatGPT’s web search runs on Bing. Claude uses Brave Search for real-time retrieval. Gemini outside of Search Grounding has its own retrieval layer. None of these platforms have confirmed that Google organic rankings drive their citation decisions — because they don’t.

The research bears this out. Only 11% of domains are cited by both ChatGPT and Perplexity. If GEO were just SEO, you’d expect significant overlap. You don’t get it. These are different systems producing different results for different reasons.

Industry leaders outside Google see it the same way. Perplexity measures source authority differently — recency, citation density, structured content, direct answers. Platforms like Otterly, Superlines, and Siftly have built entirely separate measurement frameworks for AI citation tracking because Share of Voice in Google simply doesn’t translate to Share of Model in AI. Princeton’s original GEO research identified nine distinct optimization strategies — none of which were “rank higher on Google.”

Google’s position is commercially convenient for Google. It’s not a complete picture of what’s actually happening across the AI search landscape.

That’s the argument I keep having. And I’m comfortable having it.

The Mistake That Kills Most GEO Programs Before They Start

This is the one I see most often, and it frustrates me more than anything else in this space.

Brands treat GEO like a quick fix.

They’ve spent years in an SEO culture that rewards rapid iteration — publish more content, build more links, see results in 60 days. Some of that speed is real. A lot of it is manufactured urgency from agencies that need to show monthly wins.

GEO doesn’t work like that. I tell every client the same thing: early citation signals at 30 to 60 days. Measurable Share of Model growth at 90 days. Full compounding effects at 6 to 12 months.

Sometimes longer.

This isn’t a failure of the discipline. It’s how AI knowledge graphs actually work. When you build a Wikidata entity, it takes weeks to propagate across AI systems. When you publish original research, it takes time for other sites to reference it and for AI models to register that co-citation. When you earn a mention in an authoritative publication, ChatGPT’s training cycle doesn’t update overnight.

The brands that win at GEO are the ones who treat it as a long-term compounding investment, not a campaign with a 90-day deadline.

The brands that fail at it are the ones who give up at month three because they haven’t seen the results they expected from a discipline they underpriced and under-resourced from day one.

What We Actually Check

When we audit a brand’s AI visibility, five things tell us almost everything we need to know.

Can AI crawlers actually reach your site? This sounds basic, but it’s wrong more often than you’d think. Cloudflare’s default setting blocks AI crawlers. JavaScript-rendered sites are invisible to most AI bots. ChatGPT’s web search is powered by Bing — if you’ve never submitted to Bing Webmaster Tools, you’re invisible to the most-used AI platform on earth.

Does the AI know who you are? I ask ChatGPT and Perplexity: “What is [your brand]?” and “Who founded [your brand]?” If the answer is wrong, vague, or non-existent, your entity signals are broken. This is fixable, but it takes time.

Is your content structured for extraction? AI tools pull chunks of text to cite. If your content doesn’t work as standalone chunks — if each section doesn’t make sense without the surrounding context — it won’t get cited. Clean headings, direct opening answers, tables, lists. Not because it looks nice, but because that’s how AI models extract and attribute information.

What’s your Share of Model Score? This is the number. How often does your brand appear when we run 50 relevant queries across the five main platforms? Most brands start under 6%. After a structured GEO program, our target benchmark is 50%+ Share of Model Score within 150 days.

Where are your competitors getting cited that you aren’t? The gap in the competitive analysis is usually where the priority list writes itself.

Why Now

Two things changed in early 2026 that I think about a lot.

ChatGPT confirmed that paid advertising is coming. OpenAI’s CFO said it directly. Once paid AI placement exists, the organic citation space compresses. The brands that build AI authority now — while organic citations are still purely merit-based — will be harder to displace when paid placement changes the dynamics.

At the same time, AI-referred web sessions grew 527% year over year in 2025. ChatGPT referral traffic converts at 15.9% versus 1.76% for Google organic. This isn’t a niche channel or an experiment. It’s a real, growing, high-converting source of business that most brands are leaving entirely unaddressed.

I’m not trying to manufacture urgency. The numbers do that on their own.

What We’re Building

Cited is a managed GEO agency. We run AI visibility programs for B2B brands — tracking Share of Model, identifying citation gaps, implementing the technical and content work that moves the numbers, and reporting on it every month.

But the larger project is building the category itself.

The resource library at getaicited.co — the pillar guides, the benchmarks, the methodology documentation — exists because the brand that defines a category becomes the authority that AI tools cite when someone asks what that category actually is.

We’re applying GEO to ourselves.

If you’re a B2B brand and you’ve never asked “how often does ChatGPT recommend us?” — that’s the starting question. You can get the answer, across 50 queries and five platforms, in 48 hours.

No call required.

Wahid Ryland is the founder and CEO of Cited (getaicited.co), a GEO agency helping B2B brands get cited in ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews.

Free AI visibility audit → getaicited.co
Connect on LinkedIn: linkedin.com/in/l-wahid-ryland-a0584781