Our client is a world leader in the beverage industry with several tens of Billions in revenue. As an ERP, they use SAP. In addition, they have developed a custom supply planning interface with a proprietary algorithm. Tests have shown the performance of the GenLots algorithm and a potential reduction of inventory by 15% and a reduction in deliveries by 30%.
The challenge
The unique IT landscape of our client led to the following objective:
- Replace the current supply plans in the existing custom interface with the ones from the GenLots Machine Learning algorithm. The planners should not even realize that the recommendation engine has been changed
Our approach
- Together with the clients IT organization, we defined a REST API (Application Programming interface) to exchange data
- The client developed their API on their custom system
- GenLots uses the data provided to run calculations and write back the supply plan recommendations
- We put in place a close monitoring over the first 2 months to test the acceptance of the new recommendations
- During this time, we used the GenLots reporting functions to monitor KPIs and compare the existing supply plans with the GenLots recommendation engine
Results
- GenLots has been confirmed to generate the targeted savings on inventory and deliveries
- We detected and corrected faulty master data (example: wrong minimum order quantities) in the process
- We optimized the clients supply planning without any apparent change for the user
- Therefore, users need no training and a global rollout as well as support can be very fast with little effort
- We are in talks to collaborate further to solve thorny planning problems such as the transfer of supplies between production sites
