Multi-agent supply-chain exception management

Automotive · Supply Chain

AWS Bedrock AgentCorePythonFastAPIAthena / SAPDocker

Challenge — A German automotive OEM's material disposition team manually triaged supply-chain exceptions across SAP data — slow, reactive, and hard to scale.

Solution — A multi-agent system (planning → cognition → action layers) on AWS Bedrock that detects exceptions, runs inventory-forecast and transport analysis, and drafts escalations automatically.

Result — Disponenten get analysis and recommended actions instead of raw alerts — exception handling shifts from reactive to proactive.