Demand Forecasting and Cancellation Risk for OEM Supply Chains
The Challenge
Unpredictable OEM order flow made it impossible to dimension real production capacity. Last-minute cancellations caused bottlenecks and immobilised raw material inventory. Randomness across the supply chain created procurement instability and SLA pressure.
What We Built
ML models trained to discriminate statistical noise produce a refined net demand signal, separating firm orders from those with high cancellation probability. Early identification of at-risk orders enables dynamic MRP parameter adjustment before disruption reaches the production floor.
Operational Impact
Production aligned to real net demand with no unplanned stops or overtime. Drastic reduction in raw material safety stock. Improved SLA towards OEMs without structural cost overruns.
Technology Used
AWS



