With AI, can a day of production planning be reduced to one hour?

Complex Challenge

Bakels Senior has been developing, producing, and distributing custom baking mixes for bakeries around the world since 1904. Nicky Gommers, who is responsible for IT and continuous improvement, emphasizes how complex the challenge Axelio took on is: "We work according to the highest certification standards and plan about 350 products based on customer priorities. Additionally, we deal with allergens and different product colors, which means we can't just produce one product after another. There are 13 values that determine whether the production line needs to be cleaned in between. For example, if a dairy-free product is scheduled, we need to clean the entire production line beforehand. And every cleaning takes time during which we cannot produce. That's why optimal planning is of great importance to us."

Meeting with the Technicians

Despite the complexity of the challenge, Nicky is fully confident that he will succeed with the Axelio team. "They are persistent and love a challenge. I previously worked with them on the implementation of our new CRM system. It's now working perfectly." He appreciates that he can consult directly with the technical responsible parties at Axelio. It speeds up communication and, in his view, reduces noise on the line. "And because of my technical background, I don't need the functional consultant that usually sits between software specialists," Nicky states. He adds: "You are really actively involved in the entire process and get the chance to contribute your thoughts. I can appreciate that."

From Human Expertise to AI

Nicky shares that he had already been tinkering with AI himself. "I just lack the time to really build a good AI agent. When this came up during the CRM project, Axelio offered to dive into it. We hope that production planning with AI will soon be largely automated and at least as accurate as when it is done by our people. Ultimately, the goal is for orders to be automatically placed in the ideal sequence in the production system, and for cleaning orders to be automatically generated by the AI agent as well."

More Complex Than Expected

With Axelio, Nicky has ensured that the AI agent knows the characteristics of all products and when cleaning is necessary. However, they are not there yet. To make the right decisions at every step like the experienced production planners at Bakels, a certain intuition is needed. "We now mainly need to look at the best order in which to schedule the different products so that the AI agent automatically sets the correct priorities. For that, the right knowledge from our people is required."

Breaking Down the Scope

To achieve the desired result with the AI agent, Nicky decided with the Axelio team to break the original scope into smaller pieces. "When we delved into the results after the last tests, it turned out that the order could be even more efficient. The reason is that AI thinks in black and white, while you interpret data more nuanced as a human. As a result, the AI agent sometimes presents an option that we, based on our experience, say is not practical." Nicky admits that this didn't come as a complete surprise to him. "If we let three different people manually create the planning, we get three different solutions," he smiles.

Saving Time and Downtime

For the continuation of the process, Bakels Senior and Axelio decided to only start the next step once the previous phase was satisfactorily completed. "I ultimately don't want to take any risks," Nicky emphasizes. "As soon as the AI agent is working satisfactorily, we will also first run a pilot for a few months." That does not change the fact that Nicky's expectations are still high. "As it stands now, creating the planning will take our people at most an hour. Additionally, we believe that the AI agent can lead to less downtime."

Developing Together

Despite the obstacles he encountered during the development of the AI agent with Axelio, Nicky sees it as a collaboration with a future. "The great strength of Axelio, in my view, is their willingness to develop solutions together. You see that rarely. But precisely because of this, we as clients understand the background of a solution and can manage it ourselves in the future. It also makes you less dependent. I notice this with our CRM system as well. I now make 80% of the changes myself. And if something becomes too complex for me? I can always reach out to Axelio. This way of working also leads to better internal adoption of an IT solution."