GEO Is Not An SEO Problem, And Pitching It Like One Is Why It Stays Broke
Everyone in my world has the same line right now. If you are not showing up in ChatGPT and Google's AI answers, you are invisible, and you had better start optimizing for it yesterday. It is repeated at every conference, in every newsletter, on every panel where somebody wants to sound early.
And you know what? The underlying point is basically right. People are asking machines for recommendations, and the machines are answering with a shortlist that does not include you. That matters.
But here is what I keep watching happen. The people who understand this problem best are also the people worst positioned to fix it. Because they keep walking into the room and pitching it as an SEO project. A tactic. A new checkbox on the same old list. And a tactic gets a tactic sized budget, which is to say almost nothing.
So let me pressure test the whole approach. Not whether GEO matters. Whether we are selling it in a way that could ever actually get funded.
The survey that should have made more noise
TransUnion recently put a question in front of roughly a hundred marketing leaders, and the answers described a blind spot most of us already feel in our gut. These leaders know AI search is shaping what customers see and believe. What they cannot do is see inside it. They cannot reliably tell you when an AI tool recommended them, how often, to whom, or why. The influence is real and the visibility is close to zero.
If that shape sounds familiar, it should. We have lived through this exact movie before.
We called it the walled garden problem. For years the biggest platforms happily took your money and your attention and handed back a version of reality they controlled. You could see the numbers they wanted you to see. Conversions inside their own house looked incredible. Anything that happened outside the wall, or across a few walls at once, turned into fog. You knew the platform was doing something. You could not prove what, and you certainly could not compare it fairly to anything else.
AI search is not a new problem. It is the walled garden with better manners and a friendlier voice.
That is the real lesson buried in the TransUnion findings. The AI answer box is the newest walled garden, except this one does not even show you the ad auction or the impression count. It just quietly decides whether you exist in the conversation, and moves on.
Why the SEO frame quietly kills the budget
Now here is the part I want the SEOs in the room to sit with, because I say this as one of you.
When you walk into a leadership meeting and say GEO is the next phase of SEO, you have just told everyone in that room exactly how much it is worth. You have filed it under a line item they already understand and already underpay. You have said, in effect, this belongs to the same corner of the org that fusses over meta descriptions and page speed. Important work. Not board level work.
And the moment it lands in that corner, three things happen.
The budget stays small, because it is measured against last year's SEO budget instead of against the revenue at risk. The ownership stays narrow, because it is one person's problem instead of the company's. And the measurement stays impossible, because you have no authority to reach across web, PR, product, sales, and brand to actually see the full picture.
That is the trap. You cannot solve a company wide visibility problem with a department sized mandate.
A hammer is a fine tool. But if you show up to a meeting about the whole house and you introduce yourself as the person who owns the hammer, do not be surprised when they only ask you about nails.
What the AI answer actually eats
Let me make the stakes concrete, because abstraction is how good ideas get ignored.
When a buyer used to research a purchase, they moved through a messy, human trail. A search here. A review site there. A competitor comparison. A blog post from someone they trusted. A sales call. Your brand had a dozen chances to enter the story, and every one of those chances left a footprint you could at least partly track.
Now compress that entire trail into one sentence from an AI assistant. "What are the best options for a company like mine?" And the assistant answers with three names.
You are either one of the three or you are not.
There is no second page to fight your way onto. There is no cheaper keyword to sneak in through. There is a synthesized recommendation, delivered with total confidence, and it draws on your reviews, your press, your product pages, your Reddit threads, your comparison articles, the way your customers describe you, and a hundred other signals that were never owned by the SEO team in the first place.
The AI does not read your website. It reads your reputation, and reputation was never one department's job.
So ask yourself the honest question. If the machine is assembling its recommendation from your whole footprint across the entire internet, why on earth would the fix live inside a single specialist's task list?
Reframe it as a measurement problem, not a ranking problem
Here is the shift I have been pushing with clients, and it changes the conversation every time.
Stop pitching GEO as "help us rank in AI." Ranking is a promise you cannot keep with certainty, because you do not control the model and you cannot audit its reasoning. When you promise rankings and cannot prove them, you sound exactly like every snake oil SEO pitch the CFO has already learned to distrust.
Pitch it instead as a measurement gap that puts revenue at risk. That framing is true, it is defensible, and crucially, it is something leadership already knows how to fund. Companies spend real money to close measurement gaps. They have entire teams devoted to attribution, to brand tracking, to knowing where they stand. A blind spot is a business problem. A blind spot with money flowing through it is an emergency.
The TransUnion angle helps here precisely because it did not come from an SEO. It came from a data and identity company surveying marketing leaders. That is not a niche tactician saying the tactic matters. That is the measurement world raising its hand and saying we cannot see this, and we are worried. When you carry that framing into the room, you are no longer the hammer person. You are the person who noticed the house has no smoke detectors.
Who actually needs to be in the room
If GEO is a company wide problem, then the meeting cannot be you and one skeptical marketing manager. It has to include the people whose signals feed the machine.
The brand and PR people, because the AI leans hard on what credible third parties say about you.
The product and sales teams, because the way you describe your own offer, and the way your customers repeat it back, becomes the raw material for the answer.
The analytics and data folks, because they own the measurement discipline you are going to need to prove any of this over time.
And yes, the SEO and content team, because someone still has to do the structural work that makes you legible to a machine in the first place.
Notice that SEO is on the list. I am not arguing SEO is irrelevant. I am arguing it is one seat at a bigger table, and when it tries to be the whole table, it gets a smaller room.
Measure what you can, admit what you cannot
Let me be straight about the hard part, because the hype crowd never is.
You cannot fully measure AI search visibility today. Anyone selling you a clean dashboard with a confident GEO number is selling you the same fantasy the walled gardens sold for a decade. The honest position is that you can measure some of it, you can triangulate more of it, and you have to be adult enough to say where the fog starts.
You can test the assistants directly and watch whether you show up for the questions that matter to your buyers. You can track branded search and direct traffic for the quiet lift that happens when an AI keeps naming you. You can watch referral patterns as the tools start sending clicks. You can audit the third party sources the models actually cite and go influence those. None of that is a perfect view. All of it beats pretending.
And here is the thing about GA4 and the rest of your stack. They will show you traffic that arrives already convinced, and they will not tell you why. A visitor who lands ready to buy because a machine vouched for you looks, in your analytics, almost identical to any other visitor. The tool cannot see the conversation that happened before the click. Your analytics measure the door. The recommendation happened out in the street.
So the goal is not a false sense of precision. The goal is to name the gap out loud, size the risk honestly, and get resourced to reduce your blindness over time. That is a fundable story. "Trust me, it is working" is not.
The compared to what test
Whenever someone tells me GEO is too new to invest in, or too unmeasurable to justify, I ask the same question I ask about every piece of marketing advice.
Okay, compared to what?
Compared to spending the same money on channels that are getting more expensive and more crowded every quarter? Compared to the cost of quietly disappearing from the exact moment your buyer asks for a recommendation? Compared to the walled garden numbers you are already trusting despite knowing they are inflated inside their own house?
Suddenly the unmeasurable frontier looks a lot less reckless and a lot more like an insurance policy on your relevance.
Where this leaves you
The belief I started with is not wrong. AI search matters, and getting recommended by these tools is going to separate the companies that grow from the ones that slowly wonder where their pipeline went. I believe that.
What I am challenging is the packaging. We keep taking a problem that reaches into brand, product, sales, PR, and data, and we keep shrinking it down to fit a specialist's task list, because that is the room the SEO happens to be standing in. Then we act surprised when it gets funded like a task.
The TransUnion findings are useful because they hand you a different door to walk through. Not "we need to optimize for AI." Instead, "we have a measurement blind spot in the exact place customers now make decisions, it looks a lot like the walled garden gaps we already learned to fear, and it belongs to all of us."
Say it that way and you stop being the hammer.
Stop asking for a bigger SEO budget. Start describing a company sized hole and let the company decide it wants to fill it.
The frontier is real. The blindness is real. The mistake is pretending either one is small enough to fit in a corner. Name the problem at its true size, and for the first time the money in the room is looking at the right thing.
