How Agristo and Vintecc are building tomorrow’s smart factory, together

Agristo is one of the largest players in frozen potato processing, active in Belgium, France, the Netherlands and beyond. Its ambition is clear: to become the world leader in potato valorisation, from field to fork. But that growth brings a technological challenge that traditional methods can no longer keep in check.

“Agristo’s rapid growth creates two challenges,” says Kristof Wallays, Chief Global Expansion, Innovation and Sustainability. “Production complexity keeps rising. The number of parameters to monitor, from the quality of the potatoes on arrival to the finished product, is no longer manageable for human operators alone. And at the same time, you lose critical knowledge whenever experienced people want to take on another role.”

Agristo was not looking for a supplier to deliver a one-off fix. It was looking for a technology partner that thinks alongside it about its entire innovation strategy. “We don’t see Vintecc as a party that solves one specific problem,” says Eline Wallays, who leads the innovation track within Agristo. “We see them as a partner that can help shape the whole innovation strategy, including how we scale this, how we can roll it out to other factories around the world.”

That choice matches what Gilke Van Hove, who leads the project on the Vintecc side, describes as the heart of the partnership. “Agristo has the deepest process knowledge in house. They know exactly how a potato has to move through their factory and what the result needs to be. We know how to translate that knowledge into mathematical models, into software, into a system that runs in production and integrates with everything around it.”

In concrete terms, Vintecc’s AI&DataOps team builds AI models that capture process knowledge and translate it into real-time control. The starting point is potato storage, where quality differences early in the process already determine what is achievable further down the line. Less dependence on operators, less variation between shifts, less waste.

Anyone expecting a project like this to start with a fully detailed specification is mistaken. “The project is hugely innovative for the sector, which means the requirements and the development keep changing,” Gilke admits. “Nobody has done this before at this scale. We use simulations to back everything up, but the moment we go live, new insights appear again. That is when we have to switch gears quickly.”

The project started about two and a half years ago. From a first feasibility study to a first simulation, to model adjustments, to testing again. An iterative process that moves ever closer to the operational solution. “Internally we still call it the ‘feasibility study’, even though we have been working on the actual solution for a long time,” Gilke laughs.

That iterative nature demands trust on both sides. And that trust is there. “There is an understanding that something like this does not work perfectly from day one,” Gilke says. “Both parties share the same goal: to make processes better and more optimal. That makes working together a pleasure, even when it gets complex.”

One of the most concrete lessons from this project: scalability is not a feature you bolt on later. The solution being proven in Wielsbeke has to be deployable tomorrow in Tilburg, in northern France and beyond. Agristo is already building factories in the United States and India. That means the architecture of the system, the data structure, the models, everything is designed to be industrialised and scaled.

“The danger in many AI projects is that people think in pilots,” Gilke says. “You prove that something is possible, but you don’t think about how you are going to scale it to ten factories. That is exactly what we are doing together now: proving that it works and building in scalability at the same time. So that the investment in factory two and three is a fraction of factory one.”

Eline confirms that vision from Agristo’s side: “We think from field to fork. This project is one link in a bigger story. But it builds confidence. Confidence that Vintecc can clear the bar as a technology partner, and that we as a company can deploy AI at industrial scale.”

The deepest impact of this project is less tangible than a percentage of waste reduction, but perhaps even more relevant in the long run. It comes down to one question: what happens to crucial process knowledge when the people who carry it leave?

Today that knowledge sits in the heads of operators with years of experience. Tomorrow it will sit in a model that learns continuously, adjusts objectively and performs consistently, regardless of who is at the controls that day. “The algorithm is built on the knowledge of product, process, quality and IT experts, and more,” Gilke says. “We consolidate that knowledge into a technical model. The result is a system that takes the rising complexity off the operators’ shoulders and supports them with the right information at the right moment, so they can adjust more consistently and more confidently where they can make an impact.”

That is exactly the strength of the approach: the operator stays at the wheel, but gains a co-pilot that never gets tired, never has a bad day and watches every shift with the same sharpness. A more consistent end product of higher quality, made with less waste, is then not the result of fewer people, but of better supported people.

The current project is a starting point, not an end point. Agristo’s roadmap stretches across the entire production chain. Planning potato deliveries on arrival, optimising cutting lines, energy management, logistics. Every step is a new opportunity to deploy AI where people still make decisions today based on feel and experience.

“We try to inject ideas, to understand their business questions and to converge on what they ultimately want to achieve,” Gilke says. “Which is: to make the best potato products in the world, as efficiently as possible. We try to help with that in every way we can through technology.”

The partnership between Agristo and Vintecc is therefore more than a project. It is a structural collaboration between two companies that are each an innovator in their own field, and that together reach further than either could alone.

Want to know how Vintecc can help your production company turn process knowledge into measurable results? Get in touch.