Once the technology is taken care of, what decides an AI transformation is no longer technical. It is the willingness to win, organizational agility and managerial courage. A week in China and Singapore made that plain.

AI use : 25%.

I was in China and Singapore last week. I try to go at least twice a year, and I have done so for a decade. Every trip gives me the same shot of energy and speed. This one also caught me off guard with a simple principle: once the AI technology is taken care of, organizational agility, courage and the willingness to win are the defining factors.

Six months

I visited XPeng, often called the Chinese Tesla: flying cars, humanoid robots and very good cars. Over the last six months, they implemented internally one of the products a venture we are building at OSS Ventures is creating right now. The technology that makes it possible has existed for about six months. So the gap between “this is technically possible” and “this runs in production” was close to zero.

The speed of diffusion was immense. When I asked the manager how they handled scaling, the hardest part of every deployment I see in Europe, the answer was that scaling was “never the bottleneck, as everyone wants to win.”

Context matters. A listing is coming in the months ahead, so nearly everyone in the company has a stake in it going well. And I am well aware that the Chinese model should not be copied and pasted: the labour market, the incentives and the social contract are different, and not all of it would translate to Europe. But the contrast with even the best-organized companies in the West was striking.

The technology is no longer the constraint

For most of the past three years, the honest answer to “why hasn’t AI changed this process yet?” was the technology. The models hallucinated, the integrations were brittle, the costs were high. That answer is expiring. The same models, the same tools and the same cloud are available in Shenzhen, Singapore, Paris and Detroit, often within weeks of each other. When everyone can buy the same technology, the technology stops being the difference.

What remains is everything that happens after the demo works. In my experience, three things decide it.

  • Willingness to win. Not a slogan on a slide, but a shared, felt stake in the outcome. When people want the company to win more than they want to protect their current role, a new tool is a weapon rather than a threat. Adoption stops being a change-management programme and becomes a pull.
  • Organizational agility. The number of meetings, sign-offs and committees between “this works on one line” and “this works on every line”. Most large Western organizations are not short of ideas or pilots; they are short of the ability to turn a successful pilot into the default within one quarter.
  • Managerial courage. Someone has to decide to change how work is done, accept that the first version will be imperfect, and own the result. Waiting for the next model, the next benchmark or the next steering committee feels prudent. It is, more and more, the riskiest choice on the table.

Why the West stalls after the pilot

I have watched dozens of AI pilots in European and American companies. The pattern is familiar: a strong proof of concept, enthusiastic users, a business case that holds, and then a long plateau. The plateau is rarely technical. It is a procurement cycle, a works council that was consulted too late, a middle manager whose objectives were not changed, an IT roadmap already booked for eighteen months, a leader who prefers one more pilot to one decision.

Each of these brakes is reasonable on its own. Together, they turn a six-month technology into a three-year transformation. In a world where the technology itself moves every six months, a three-year transformation means deploying something that is two generations old by the time it lands.

What to take, and what to leave

The lesson is not to import the Chinese model wholesale. It is to stop pretending that the bottleneck is still technical, and to work on the three things that actually are:

  • Give people a real stake in winning. Incentives, ownership and recognition tied to the outcome of the transformation, not to the preservation of the existing process.
  • Shorten the path from pilot to default. Fewer gates, clear owners, and a rule that a pilot that works is scaled or killed within a quarter, never left to drift.
  • Make decisions with imperfect information. The technology will be better in six months. It always will be. Courage is deploying what works now and improving it in production.

The companies that win the AI era will not be the ones with access to better models. Everyone will have those. They will be the ones that want it more, move faster and decide sooner.

Anyways, time to build.