AI lessons from the boardroom
How Albert Heijn is bringing AI to the shop floor: Wies Mensink on starting practically and learning quickly
By Rutger Huizenga
AI is now on almost every agenda. Not as a gadget, but as a strategic question: what should you use it for, how do you scale it, and how do you prevent experiments from getting stuck in the pilot phase? In the interview series "AI lessons from the boardroom," we speak with executives and board members who have to make those choices in practice. In this edition: Wies Mensink, VP Data Science & Analytics at Albert Heijn, Etos and Gall & Gall. “My teams have a large toolkit with possible solutions for complex problems. GenAI is one part of that.”
Wies Mensink is not the kind of leader who gets swept up in doomsday scenarios about AI taking over jobs. In practice, she sees something different: work that is fundamentally changing. She uses AI tools herself to sift through mountains of information more quickly, sharpen texts, or rapidly learn new concepts. But she sees the biggest shift in her own teams of around 150 specialists: machine learning engineers, data scientists, and advanced analytics experts.
The days when data scientists typed out every line of code manually are behind us. “You still need deep knowledge of code, but AI now acts as a flywheel. Things move faster, but the human eye on the output is more important than ever. After all, you have to understand what the machine is building for you in order to guarantee quality.”
No romanticism
That practical stance also runs through the way Albert Heijn looks at GenAI more broadly. “We don’t start with AI,” Mensink says. “We start with the problem we are trying to solve.”
If a simple algorithm can solve a logistics issue, then that is the route they take. GenAI only enters the picture when it genuinely adds value. “You don’t want to find out too late that something won’t work,” she says. “Usually, you can tell during the process whether something needs to be complex, or whether a simpler solution will do.”
A tangible example is “Steijn,” Albert Heijn’s digital assistant. Whereas the search function in the app used to be static, Steijn now helps customers with a daily dilemma: “What are we having for dinner tonight?” But the value goes beyond recipe advice alone, Mensink believes. “For Albert Heijn, it is a way to better understand what customers actually need. It changes the relationship from one-way sales to a supportive dialogue.”
On the shop floor too, AI is mainly about something very practical: helping employees more quickly. To that end, Albert Heijn developed, among other things, an AI assistant. “Employees can simply use it on their mobile phones. Where is this product located? When will it arrive? Which wine goes with this dish?” That may sound convenient, but Mensink sees it as essential. Many employees are young and work part-time. “Then you don’t always know where everything is,” she says. “So it’s very useful if you can simply look it up on your own phone.”
The scale of that use shows how quickly AI can become part of day-to-day work. Around 80,000 colleagues in Albert Heijn stores have already asked the AI Assistant questions, and every week more than 40,000 conversations are launched.” In that way, we make work more enjoyable and help our colleagues work more efficiently.”
Another practical example can be found in the bread department, where Albert Heijn combines smart baking and dynamic pricing to better match supply to demand and reduce food waste by more than 2 million kilograms a year. Internally, bakery teams use Bak, an AI app developed by Albert Heijn that helps predict and adjust bread demand throughout the day. In the second half of 2025 alone, Bak helped reduce bread waste by 815,000 kilograms.
Redesign processes
On paper, these are practical applications. A customer who quickly figures out what to put on the table tonight. A store employee who immediately sees where a product is located. A bakery team that has to throw away less bread.
But Albert Heijn is also looking beyond these day-to-day improvements. The company is exploring how AI can help redesign processes themselves. Not just making the current way of working faster or more efficient, but asking whether that work should still be organized in the same way at all. That, Mensink says, is where the real challenge begins. “Then it is no longer just about working smarter. It becomes a question of genuine transformation, just as we saw twenty years ago with digital transformation.”
And that is where AI becomes more than a tech project. When the work changes, people have to move with it. Not just the early adopters who are already eager to get started, but the wider organization too.
At Albert Heijn, that challenge is being approached as part of a broader HR and AI strategy: helping employees across the business build AI skills and confidence step by step; through training, a large internal community, newsletters, experts who support that community, and tools that employees can use themselves. Peer-to-peer learning also plays a role. “The community trains each other as well,” Mensink says.
Learning from KickstartAI
The learning doesn’t stop within Albert Heijn itself. Through KickstartAI, the company also exchanges knowledge with the other founding partners: ING, NS, and KLM. For Mensink, this makes a lot of sense as all of these organizations are in the middle of the same transformation. “We can exchange a lot of knowledge,” she says, “in terms of architecture, how you build certain things, how you set up projects, and what the lessons learned are. We benefit a lot from one another.”
For Albert Heijn, the value stretches beyond knowledge-sharing between organizations. The data scientists themselves also get something in return. Through KickstartAI, they occasionally step outside their own Albert Heijn context and work on other challenges, for example for the police or the education sector. “I notice that they really enjoy that,” Mensink says. “And that they learn a great deal from it too.”
One example she is especially passionate about is the food waste project between Albert Heijn and KickstartAI. The same team later applied those insights to a similar challenge at KLM, showing how knowledge carries across organizations rather than staying siloed. And according to Mensink, the whole of the Netherlands can benefit from that.
At the same time she also sees a challenge. Developments are moving so fast that the government and society are struggling to keep up. A gap is emerging, in knowledge and in the speed at which roles are changing. “As large AI organizations, we can bring that knowledge to the government. It has to be a joint effort to ensure that talent stays in the Netherlands.”
Her message to other executives is therefore twofold. Internally: “Just get started. Experiment across different subareas, adopt a build-test-learn mentality, and constantly communicate about what does and doesn’t work.” Externally, she is just as clear: “If we, as major Dutch corporates, join forces, we will not only help our own businesses, but move the whole of BV Nederland forward. The door is open; who’s next?”
Wies Mensink’s 4 AI lessons
1. Don’t start with GenAI, start with the problem
GenAI is not an end in itself. It is one tool within a broader toolkit of data science, advanced analytics, and technology.
2. Make AI practical for day-to-day operations
AI only becomes valuable when customers and employees benefit from it directly: from meal inspiration to product information on the shop floor.
3. Sometimes a simple algorithm is better
Not every problem calls for the latest technology. If an existing solution works, there is no need to build something more complex.
4. Experiment broadly and learn quickly
Build, test, learn. Start in different areas, see what works, and pivot quickly if something does not create impact.
