I keep getting sent the same kind of pitch. Subject line some version of "win answer engine optimisation in 30 days." Schema audits, FAQ rewrites, structured data fixes. The whole package, sold as a checklist.
It's not wrong, exactly. Those things matter. I do most of them on every project.
But it's the wrong starting point.
What AEO is being sold as
Open any article on answer engine optimisation right now and you'll get the same playlist. Add JSON-LD. Build out your FAQ schema. Make sure your structured data validates. Get cited by Reddit. Get on Wikipedia. Add an llms.txt file. Tighten your headings.
All useful. All necessary. None of it new in spirit, even if the labels are.
What it adds up to is a tidy technical layer on top of whatever brand you already have. Plumbing for the answer engines to read.
The problem is what's underneath the plumbing.
What AEO actually is
If you watch how an AI platform builds an answer about a category of business, the picture is less mystical than people make it. The model is sweeping a defined set of trusted sources, looking for consistent signals about who does what, and assembling a short list of names it can stand behind.
The signals it's looking for are signals brand strategy has always tried to produce. Clarity about who you are. Consistency across surfaces. A specific answer to a specific buyer question. A name attached to a category, repeated by people who aren't you.
That's not new. That's just brand strategy with a new audience reading the output.
Tactical AEO assumes the brand underneath is already clear. If it is, the technical work compounds. If it isn't, you can validate every schema in the world and the model still won't know what to do with you.
The problem most brands have isn't technical
In my experience, when a brand isn't getting cited, the cause is almost never the markup. It's that the brand can't answer a basic question about itself in one sentence.
I've sat with founders who can describe their business for ten minutes without landing on what it actually does. I've read homepages that list eight services with no opinion about which one is the lead. I've audited About pages where the company name appears nine times and the customer's problem appears zero. I've seen four different versions of the same elevator pitch on the same site, on the same week.
A model reading any of that has the same problem a buyer has. Too many possible summaries. No clear answer. So it picks something else, or it makes one up, or it averages across the noise into a description nobody at the company would sign off on.
That's not an SEO failure. That's a positioning failure. The technical fixes don't reach it.
What AI platforms reward, in plain language
A few things, ranked roughly by how much they matter at the input layer.
One named answer to one buyer question. If your buyer asks "who's good at X for businesses like mine," your brand needs to be a defensible answer. Not the answer to twelve adjacent questions. One. The clearer the question you own, the easier you are to cite.
Consistency across every surface. Homepage, About, services pages, social bios, podcast intros, third-party listings. If they describe slightly different businesses, the model has to choose, and it'll usually choose the description that appears most often in the trusted sources, which may not be the one you'd pick.
A clear position, not a list of capabilities. Capabilities are commodity. Position is who the work is for, what problem it solves, and what the studio or company believes about how that problem should be solved. Models pick up on positioning the same way humans do, by noticing what gets repeated and what gets skipped.
Third-party language that matches yours. When somebody else writes about you, do they use the same words you'd use? If your site says "brand studio for founders" and every press mention says "creative agency," guess which one the model is going to lean on.
None of that is markup. All of it is brand strategy.
Why agencies are selling the wrong half of the work
Because the wrong half is easier to sell.
Schema markup is a fixed-scope job with a deliverable you can demo. Brand strategy is a longer engagement with a fuzzier output. So the people who got into AEO from an SEO background reach for the schema work first, because that's the muscle they already have. And the people buying it want a deliverable they can sign off and move on from.
I understand the incentive. I just think the result is a lot of polished plumbing connected to vague brands. The brands that get cited are the ones where the strategy underneath is already in good shape, which makes it look like the technical work caused the citations. Mostly it didn't. The technical work made an already-clear brand legible to the machines.
If the brand wasn't clear to begin with, no amount of structured data is going to rescue it.
What this means in practice
If you're trying to get cited, the order matters. Get the brand clear first. Then make it legible.
Clear means: one sentence about what you do, who it's for, and why someone would pick you over the obvious alternative. Tested with real buyers, not iterated alone in a tab. Repeated on every public surface in language consistent enough that a model could extract it without ambiguity.
Legible means: schema, FAQs, llms.txt, citable About pages, structured data. The technical layer doing its job because the layer underneath is doing its job.
The order is everything. Reverse it and you've spent the budget on plumbing for a brand that still can't be summarised.
A short test
Open your homepage. Read the first paragraph. Now imagine an AI platform asked to recommend a business in your category, given a buyer's specific problem.
Could it use that paragraph as the answer? Word for word, no edits, no clarifications?
If yes, the rest is plumbing.
If no, the plumbing won't fix it.
That's the work. AEO is brand strategy with a new audience. The studios and consultancies that figure that out early are going to spend the next decade getting cited. The ones still selling schema audits as a standalone solution are going to spend the same period explaining to clients why the citations didn't follow.
I know which side I'd rather be on.