Leading Dutch goat cheese maker gains instant insights into process variations that help ensure consistent quality, while saving time and costs
By applying Infor Coleman artificial intelligence to past manufacturing orders/batches and other production attributes, the machine learning model detects batches that deviate from the target yield, along with an explanation of the main contributors to the yield value. This process runs daily, and insights are delivered through flexible interfaces depending on the end user — CloudSuite Food & Beverage Homepages for Operators and Infor Birst dashboards for management. Amalthea’s AI-driven yield deviation detection and explanation is now fully automated — from data collection to processing and presentation. Implemented in less than 90 days, the solution helps in decision making to stabilize the process for future manufacturing orders by:
“What we also saw and learned is that for some cheese batches, we did not have the correct target value,” Aarts said. “Infor Coleman AI showed us the average milk yield of all the cheese batches since go-live, and we used this information to change the target value to the correct value, so that the right parameters are in place for the best outcome possible.”
With Infor Coleman AI, Amalthea can now steer the cheese production process faster, and with more accuracy:
Another big win for Amalthea is that operators don’t have to spend time in diagnosing issues with yield deviations, as the Infor system automatically generates actionable insights or explains the contributors to yield deviations. Operators are now able to focus their efforts on stabilizing the cheese production process and continuing to grow with a more efficient process.
“The Infor partnership and data science team are great. They really listened very carefully to our requirements, and, in a very short time, they delivered a great tool. It is easy to use and understand, so we can improve our process and yields, which makes my life easier,” said Liesbeth Stoopen, Amalthea operations manager.
Infor CloudSuite Food & Beverage supports Amalthea across the supply chain — from milk reception to the cheese warehouse. The system interfaces with the production automation in the factory, such as weighing units and the processing tanks. It also makes financial closing much faster than it used to be, from weeks to now days. Having the business now fully running in the cloud with Infor is the basis for its future growth, as data will play an increasingly important role in dairy production.
“Infor has a process catalog specifically designed for dairy and can give good insight into milk utilization and yield. Milk utilization is the most important driver for our cost, and Infor already helped us to save on that,” Aarts said. “Now with Coleman AI implemented, we can stretch that even further because we have more real-time and faster information about the milk yield.”
We’re able to use Infor EAM to address our customers’ needs, and to provide them greater depth of wisdom just by the fact that we have a single enterprise solution for maintenance operations.
In 2020, Amalthea modernized its ERP with Infor CloudSuite Food & Beverage and leveraged the power of the integrated cloud technology platform Infor OS to solve a critical milk yield business problem. This was previously hard to tackle due to outdated and disconnected applications, data silos and lack of time and resources. The modernization formed the building blocks to create AI-driven applications in less than 90 days. As innovation is in Amalthea’s nature, that was the next logical step.
For each cheese batch, Amalthea used to set a target yield based on the practical or past knowledge of production. When a cheese batch deviated from the target yield or the right target yield wasn’t set, factory operators had to act and make changes to production to make the process more efficient. Production issues were identified by manually processing data and analyzing it on a weekly basis. With approximately 240,000 liters of milk processed each day, it was not optimal to wait a week or sometimes even a month to get insights and make the necessary system changes to optimize the yield.