Overview
This engagement focused on unifying fragmented data sources and deploying an AI forecasting engine to optimize replenishment decisions across 500+ stores.
Approach
- Consolidated POS, ERP, and external signals into a clean feature store
- Built a demand forecasting model per SKU x store with hierarchical reconciliation
- Implemented automated order recommendations
- Enabled control tower visibility with alerting and what-if simulations
Outcomes
- 37% fewer stockouts in Q1 post go-live
- $15M working capital freed from reduced excess inventory
- 3.5% improvement in gross margin