When AI Starts Picking Brands for Users, Is Your Company Still Chasing Keyword Rankings?
This article argues that Generative Engine Optimization (GEO) is displacing traditional SEO as users increasingly discover brands through AI assistants like ChatGPT and Google AI Overviews. It covers GEO market growth, budget shifts, buyer-seller dynamics, and risks such as platform dependence and regulation.
A while back, a friend who runs a consumer-goods business met me for coffee.
He looked miserable.
Said his team had spent three years clawing their way onto Google's first page — only to watch traffic drop nearly twenty percent this year. Where did it go? He couldn't even say.
I asked him, how do you search when you're buying something these days?
He paused, then said, "Honestly… I just ask ChatGPT."
And there's your problem.
One Number That Stopped Me Cold
$848 million.
That's the size of a brand-new business in 2025. It's called GEO — Generative Engine Optimization.
By 2034, that number is projected to hit $19.8 billion.
A compound annual growth rate of 50.5%.
What does that mean? The hottest SaaS categories of recent years consider 30% growth excellent. 50.5% is nearly doubling every year.
Why the explosive growth?
Because the entire foundation of how people use the internet is being swapped out.
What Do I Mean by "Swapped Out"?
Used to be, you'd search for something and Google would give you ten blue links. You'd click the first one, the second, page through a couple results. Traffic was distributed by ranking, and brands fought for position with SEO.
And now?
You ask ChatGPT "what's a good protein powder," and it doesn't throw a list of links at you. It just hands you an answer that mentions three brands.
Three.
Research shows the average AI-generated answer cites no more than three sources.
Think about that. A single category can have hundreds of brands. AI mentions three. The rest, in that moment, might as well not exist.
That's "winner-takes-all."
The brands AI mentions own the conversation for the entire market. The ones it doesn't mention aren't on page two — they've vanished entirely.
Where Did All the Users Go?
Let me walk you through the math.
By the end of 2025, ChatGPT's weekly active users topped 300 million. Google's AI Overviews showed up on more than 45% of search results pages in the US. Perplexity's growth rate exceeded 300% year over year. Microsoft Copilot was embedded in Office 365 for over 400 million commercial users.
Combined, these platforms now process more than 15 billion AI queries a month.
A year ago, that number was hovering around 7 billion.
In other words, users aren't "maybe someday" going to search with AI. They already are. And they're doing it more every month.
McKinsey's late-2025 survey was even blunter: 54% of Millennials and Gen Z trust an AI assistant's recommendation over a search engine's links when making a purchase.
Fifty-four percent.
More than half.
This means the main gateway for brands to reach people is moving from Google's search box into ChatGPT's chat box.
SEO Is Breaking
For two decades, digital marketing rested on two pillars: one called organic ranking SEO, the other paid search.
Now both pillars are wobbling.
Take SEO first. AI Overviews push the traditional blue links further down the page. Between 2024 and 2025, the organic click-through rate on Google's first-page results dropped about 18%.
Now look at paid. Ad slots are shrinking while advertisers keep pouring in. Cost per click on competitive keywords has been climbing more than 22% a year.
On one side, organic traffic is falling. On the other, paid costs are rising. Who fills the gap in the middle?
GEO.
Someone tracked a group of brands that secured top citations inside AI answers. Their customer acquisition cost was 41% lower than those relying on traditional search alone.
Any CFO who sees that number can't sit still.
Where the Money Is Moving
In 2024, only 18% of Fortune 500 CMOs listed GEO among their top three annual priorities.
In 2026, that number hit 67%.
In two years, nearly a fourfold jump.
Money follows strategy. In 2023, the average enterprise GEO contract ran $72,000 a year. In 2025, $185,000.
Nearly tripled.
Where's the money going? Mainly four buckets.
The first bucket: "AI visibility analysis." That means I first need to know whether I'm being mentioned inside each AI engine, whether the mention is praise or criticism, and which competitors are getting cited more. This slice accounts for 34.2% of the entire market — the most fundamental, must-have layer.
The second bucket: "content optimization for LLMs." Making sure your website, knowledge base, and product materials are structured and worded in a way that matches how large language models pick citations. 27.8%.
The third bucket: "brand citation monitoring." Keeping real-time tabs on how the major AI platforms talk about you.
The fourth bucket: "AI answer placement." The most expensive tier, averaging $280,000 a year per contract. The goal: get your brand to firmly occupy the top citation slot in AI answers for high-intent queries. Early customer feedback shows an average 67% lift in AI citation rate within six months.
Six months, 67%. Those numbers speak for themselves.
Who's Buying, Who's Selling
On the buyer side, large enterprises are the most aggressive. They account for 44% of the entire market. The logic is simple: big companies have the most brand equity to lose when AI overlooks them, and the biggest marketing budgets. It's no longer unusual for Fortune 500 firms to have dedicated GEO teams.
The second-largest buyer group is digital marketing agencies, at 27.3%. They may not be the end users themselves, but they package GEO and resell it to small and mid-sized brands, making them the largest channel in this business. Small and mid-sized brands make up 18.2% — a group unlocked by self-serve tools that let them get started for a few hundred dollars a month.
On the seller side, the market is still quite fragmented. In 2025, no single vendor held more than 18% market share. Pure-play GEO newcomers, traditional SEO platforms pivoting over, and AI-native companies doing this on the side — three camps all brawling for ground.
That fragmentation tells you one thing: the market is still early.
Early means the window of opportunity is still open. Once the top five to seven players swallow more than half the share, the barriers to entry go up.
Don't Rush In Before You See These Three Traps
I'm bullish on this business. But there are risks, and I need to lay them out.
First, platform dependence. GEO's lifeblood is how each AI engine selects citations. But that logic is a black box — vendors don't publish it. Unlike Google, which at least releases search quality guidelines. The day GPT upgrades or swaps in a new batch of training data, the strategy you painstakingly optimized could stop working overnight. CFOs hate that kind of thing.
Second, platforms might sell placement directly. If AI engines ever start selling "sponsored citations" — the way Google now runs paid search ads — the space for winning position through optimization gets squeezed. On the flip side, that would spawn an entirely new market for AI paid advertising.
Third, regulation. The EU AI Act's transparency requirements could force AI engines to disclose their citation mechanisms. Once that's public, the "secret sauce" GEO vendors have built up gets diluted.
These three traps aren't meant to scare you off. They're a reminder: when you bet on GEO, don't go all in on one platform or one playbook. Spread your bets, keep iterating.
Looking One Step Further
What actually excites me is the next variable: AI agents.
Right now, users ask AI, AI gives an answer, the person glances at it, and decides whether to buy.
The next step is AI placing the order for you.
In enterprise procurement, AI agents are already comparing prices, picking suppliers, and running approvals — some companies are doing it today. Once that behavior goes mainstream, what AI cites isn't just brand image anymore. It's real money — actual orders.
Whoever's product sits at the front of the agent's decision logic wins the deal.
This isn't a marketing question anymore. It's the lifeblood of revenue.
Research projects AI agents reaching meaningful penetration in enterprise scenarios between 2027 and 2028. Which means the window for brands to build an AI-citation advantage is about two or three years.
Back to That Friend
After we finished our coffee, my friend went quiet for a moment, then asked: so what should I do?
I said, start with one thing.
Open ChatGPT. Open Perplexity. Open Google's AI Overviews. Ask about your single most important product keyword — ask five times, ten times.
See whether the answers include you.
If they do, great. Hold that ground.
If they don't, that's the biggest gap you need to close next year.
Because users have already started letting AI pick brands for them. Whatever AI doesn't pick isn't even an option.
It's not page two.
It's never even making it onto the field.