When A Proof Of Concept Takes An Hour, Restraint Becomes The Discipline That Protects The Data Underneath

The proof of concept that once took a month now takes an hour. That compression is the good news and the problem at once, because when anyone in the building can stand up a working AI tool over a lunch break, the harder question stops being whether the organization can move fast and becomes whether it can connect all that motion to reliable processes without leaking data or spinning up tools that answer to no one. The answer is not to slow the experimentation down, but to build the foundations underneath it.

Maurice Schulmann is Chief Transformation, Technology, Data & AI Officer at Groupe Partouche, a French casino and leisure group founded in 1973 that now operates 43 casinos and 12 hotels. The group generated €460.2 million in consolidated revenue in its 2025 financial year, up 6% year over year. Schulmann has spent two decades at the company, starting in operations, then leading marketing and communication, before moving into strategy and general management, a path he credits for letting him understand the business from the inside. His view of AI starts with a blunt point about what the technology runs on.

“It’s like Formula One. You have a beautiful car, but if you put the wrong fuel in, it doesn’t work. It’s the same for AI. It’s very powerful, but you need to put the right project on it to be relevant and see the benefit,” Schulmann said.

• A revolution to organize around: Schulmann draws a hard line between the last technology wave and this one. The arrival of the web was a game changer, he said, but what's happening now is a different order of change. “When I was young, I saw the internet coming. It was incredible, it was a game changer, but not a revolution. What we are living right now with AI is a real revolution, bigger than the internet,” Schulmann said. The consequence he draws is organizational rather than technical: AI can't be treated as one more tool in the kit, but has to sit at the center of how the group is run. This is why he pushes the conversation up to the C-level and the group’s leadership rather than leaving it in IT.

• Where AI is already the work: In parts of the group, Schulmann said, the shift has already happened. “Right now all my developers use AI to develop all our software, to make all the updates and releases. It’s not only a project, it’s really the work today,” Schulmann said. The gains show up in both time and a level of quality the team hadn't reached before. His creative studio tells the same story: motion-design work that once took a week now turns out two or three pieces in a day. Those are the settings where the data is clean and the use case is clear. The friction shows up elsewhere, in decision-support work, where the limiting factor is the state of the underlying data.

For the analytical work, Schulmann keeps running into the same wall. The tools are powerful, but they only pay off when the data feeding them is clean. “We do a lot of proofs of concept, and we feel the limit because of the quality of the data,” Schulmann said. It's a problem he's worked before, in the years his team spent on data mining in the CRM to push the right campaign to the right customer at the right moment, but AI raises the stakes, because the same weak data that produced a mediocre campaign can now undermine a far wider range of decisions.

The economics of experimentation have also inverted. Where a proof of concept once meant a month of work and a paid provider, it now takes an hour, and that, Schulmann said, is the real change. “My issue right now is more to create a limit on the POC,” he said. Because he comes from the group's founding family and has spent twenty years in the business, he can green-light almost any experiment he wants without fighting the hierarchy, which makes his own discipline the binding constraint. The job, as he put it, is to keep calm and not do crazy things with the company’s security and its data.

• When a dealer builds his own tool, the data walks out the door: The grassroots energy Schulmann wants to preserve is also where the exposure lives. With four thousand people across the group, he describes a floor employee, someone with no digital remit, proudly building a tool over a day to help organize the work, using a public AI service, with company data flowing out of the company in the process. That instinct is exactly what he doesn't want to crush, because the barrier to building has collapsed. “The limit right now is creativity. You don’t need to know how to code. You only need a good idea and the curiosity to write a prompt,” Schulmann noted. The task is to channel that initiative rather than ban it, which means standing up the rules and the systems that keep the data inside while the ideas keep coming.

• Governance as the foundation of tomorrow, built under time pressure: Schulmann is candid that the structure doesn't fully exist yet, and that building it is now his central job. He reorganized his team toward more operationally minded, senior people precisely because the technology is no longer a tool to be administered but a shift to be organized around. The work is to gather the scattered experiments into one coherent strategy with real security underneath, drawing on outside cyber partners, and to do it quickly. “It's time to build the foundations of tomorrow, but we don’t have much time, because it's happening so fast,” Schulmann said. He frames it in a saying he returns to: you have to work today if you want to be ready tomorrow.

• Communication is the connective tissue: Pressed on how he scales governance across so many functions and so much volume, Schulmann is straightforward that he doesn't yet have a finished blueprint, and distrusts anyone who claims one this early. “We're at the beginning of the beginning," he said. "We'll take little steps, but you need very powerful communication between everybody to adapt the system to the need." What he leans on is a flat family culture and everyday collaboration tools that keep people linked. Yesterday’s best will not automatically be tomorrow’s best, so the real discipline is keeping everyone’s skills climbing and the conversation open as the system adapts.

For Schulmann, the lesson other enterprises can take from a regulated, brick-and-mortar group moving quickly is that the wave doesn't wait. He compares it to a tsunami rather than the large but survivable swell of the internet era. "Are you going to surf it, or are you going to go under?" Surfing it, in his telling, means neither blocking the grassroots experiments nor letting them run loose, but building the data foundations, the governance, and the cross-company communication that turn one person’s clever afternoon side project into value the whole business can trust.