Case Study · AI & Data
AI-Driven Supply Chain Optimisation
Global Logistics Company
Voodoo AI designed and delivered a real-time forecasting platform that turned fragmented operational data into a production decision system for inventory and procurement teams.
Challenge
Inaccurate demand forecasting was creating excess inventory, missed sales opportunities, and operational friction across 12 distribution centres.
Approach
We built a streaming data pipeline ingesting more than 2 million events per day, implemented feature-store-backed forecasting models, and integrated predictions directly into warehouse management and procurement workflows.
Impact
The result was a measurable operational improvement, with stronger forecast accuracy, lower inventory waste, and better availability across the network.
Results
60% reduction in forecasting error
25% reduction in excess inventory
40% fewer stockout events
£4.2M annual cost savings
Tech stack
KafkaPyTorchMLflowAirflowAWS SageMakerFastAPI