Inventory inaccuracies drove lost sales
Discrepancies between system and shelf inventory reduced availability and customer satisfaction.
We operationalized inventory anomaly detection
Machine learning models were deployed to detect discrepancies and trigger store-level action.
ML-driven inventory monitoring at scale
Binomial models, anomaly detection, and classifiers were orchestrated through Databricks to send restock alerts to store associates.
$80M in recovered sales across 2,300 stores
Improved inventory accuracy reduced shrink and recovered lost revenue.
