Most AI companies I work with arrive at the same conversation. The product is real. The team is strong. The early traction is genuine. And the founders are convinced that what’s holding them back is marketing — that if they could just get the story in front of more people, or hire a stronger growth lead, or run a better campaign, the next stage would unlock.
It almost never is.
What’s actually holding them back, in nearly every case, is that the market doesn’t yet understand which category they belong to. Buyers can’t place them. Investors can’t compare them. Analysts can’t write about them. The product is real, but the category around it is fuzzy — and a fuzzy category is the most expensive marketing problem a company can have, because no amount of marketing spend can resolve it.
What category actually does
A clear category does three things at once. It tells the buyer what kind of company you are. It tells the investor what comparable companies look like and what valuation logic to apply. And it tells the market what to expect from you next — which products, which announcements, which moves are consistent with who you are.
When a category is clear, marketing compounds. Every field note, podcast, conference talk and inbound conversation lands inside a frame the audience already understands. The work of explaining who you are has already been done by the category itself.
When a category is unclear, marketing leaks. Every conversation requires re-establishing the frame. Every pitch starts from zero. Every analyst report places you in a different bucket. The team is working hard, but the work isn’t compounding — because the foundation underneath it keeps shifting.
This is what most AI founders are actually feeling when they say “we have a marketing problem.” They’re not feeling a lack of activity. They’re feeling a lack of leverage. And the lack of leverage is almost always upstream of marketing.
Why AI companies have this problem more than most
AI as a sector is in the middle of a once-in-a-generation category formation. Twenty different sub-categories are being defined simultaneously — clinical AI, agentic AI, AI infrastructure, vertical AI applications, AI-native productivity, AI-augmented services, and a dozen others. Each sub-category has competing definitions. Each has its own analyst coverage. Each has its own emerging set of comparable companies.
That means an AI company founded in 2023 or 2024 is operating in a market where the rules of comparison haven’t been written yet. There’s no default category to default into. If the founders don’t claim a clear position, the market will assign one — and the assigned position is almost always less favourable than the claimed one.
The instinct in most companies is to wait for the market to settle. To let the analysts define the categories. To let competitive positioning emerge organically. This is the wrong instinct for AI in 2026 and beyond. The market is moving too fast. The companies that lead categories in the next five years will be the ones that decided to lead them — with conviction, before the analysts gave permission.
What “claiming a category” actually means
It doesn’t mean inventing a new term. Most successful category claims aren’t new words. They’re sharper definitions of existing terms, or clearer positions within emerging spaces. Harrison.ai didn’t invent clinical AI. It claimed a position inside clinical AI that made it understandable, comparable and investable at scale.
Claiming a category looks like four things, done together. A clear definition of what kind of company you are, written in language a non-technical buyer can repeat. A clear statement of what you do that the next-closest competitor doesn’t. A consistent set of comparable companies you reference in pitches, on the website and in investor conversations. And a narrative architecture that supports the claim across every public surface for the next 18 months minimum.
Most companies do one or two of these. Almost none do all four. The ones that do all four are the ones that look — from the outside — like they came out of nowhere to dominate.
Why this is upstream of marketing
None of this work is marketing. It’s strategic positioning. Marketing is what happens after the category is claimed — the campaigns, the content, the events, the inbound funnels. Without the category being claimed, marketing is just expensive noise. With it, the same marketing spend compounds.
This is why marketing budget rarely fixes the underlying problem. The team can run more campaigns, hire more content writers, attend more conferences, and the needle barely moves — because the foundation is unstable. Once the category is sharp, the same team produces dramatically more output, and the output compounds rather than leaks.
For founders and CEOs reading this and recognising themselves in it: the work of fixing the category problem is rarely something a marketing team can do alone. It requires the founder’s conviction, the executive team’s alignment, and a deliberate decision about what kind of company to be in the market. That’s strategic work, not executional work — and it’s why I spend most of my advisory time at exactly this layer.
If your company is preparing to raise, reposition or scale and you suspect what you have is actually a category problem rather than a marketing problem, that’s the conversation worth having before the next round of marketing investment.