We talk to teams across a reasonably wide spread of markets building and buying AI tooling. One pattern has become impossible to ignore: American organisations are past the pilot stage. They are running AI in production, against real customers and real workflows, and they are hitting real governance problems as a direct result, data leaving the building through a chat window, an agent taking an action nobody reviewed, an audit trail that turns out not to exist when someone finally asks for it. Almost everywhere else we look, the conversation is still "we're running a proof of concept." Not next quarter's plan. The current state, months into what should have been a rollout.
Two different markets wearing the same label
"AI adoption" gets reported as one global number, and that number hides a split that matters more than the headline. A US enterprise team running AI in production has already made the decisions that create governance exposure: which model, which data gets near it, which employees can act on its output, what happens when it's wrong. Those decisions are live and compounding daily. A team still at proof-of-concept stage hasn't made most of those decisions yet, because a PoC is explicitly the phase where consequences don't count. That is not a small maturity gap. It's a different category of problem, and a vendor selling one governance story to both buyers is selling the wrong thing to at least one of them.
PoC governance and production governance are not the same product
A proof of concept needs a checklist and a demo. Production needs an audit trail that survives a real incident, a policy engine that enforces rather than suggests, and evidence that holds up when a regulator or a customer's security team actually asks for it. Selling PoC-grade governance into a production buyer, or production-grade complexity into a PoC buyer, both fail, just in different directions.
Why the gap exists, and why it's not really about capability
It's tempting to explain this as a technology gap, US teams have better tools, or a talent gap, more AI engineers per capita. Neither holds up well once you look closer. The bigger difference is organisational appetite for making the production decision at all: US enterprises, under real competitive pressure, are choosing to ship AI into live workflows and deal with the governance consequences as they arise. Most other markets are choosing to wait until the governance question is fully answered before shipping anything real, which sounds prudent and mostly just means the PoC never graduates. Neither posture is obviously wrong. But only one of them generates the kind of governance problem that a platform like OBEL actually exists to solve, and it is not evenly distributed around the world yet.
Where ninthLABS sits in this
ninthLABS is a venture studio, not a consultancy that advises on AI from the outside. We build with AI and for AI as our default way of operating, not as a research project bolted onto the side of the business. OBEL exists because we needed exactly the governance problem this gap describes: PII scrubbing, policy enforcement, an audit trail that actually holds up, before we ever considered selling any of it to anyone else. We are not describing the production AI governance problem from a slide deck. We are running our own company on AI, hitting the same failure modes our customers hit, and building the fix before packaging it as a product.
That has a direct implication for how we read this market split. A vendor built for the PoC-stage buyer will keep selling checklists and readiness assessments as the rest of the world eventually catches up to where the US already is. A vendor built the way we build, governance-first because we had no choice, not because a market research report told us to, is already positioned for where every market ends up once its own AI adoption moves from pilot to production. The gap closes eventually. What a company builds while it's still open is what determines whether it's ready when it does.
“We didn't build OBEL because we predicted this gap in a market report. We built it because we hit the governance problem ourselves the moment we moved our own AI usage from experiment to production, and every market eventually makes that same move.”
- Denis Bouton
2
Distinct buyer stages the market lumps into one "AI adoption" number: PoC and production
1
Governance platform, built the same way for both: governance-first, not bolted on later
0
Slide decks it took to find this gap - it came from actually running the company on AI
Denis Bouton is the founding Chief Architect of OBEL™ and Managing Partner at ninthLABS Ventures, where he advises Post-Sales Services and Customer Success organisations on scaling onboarding, adoption, and retention. He leads OBEL's product and platform strategy.
More in Intelligence Brief