A stateful multi-agent system monitoring inventory deficits, querying vector memory, and executing automated purchase order re-routing.
Global supply chain networks face constant disruption from maritime port congestion, customs clearance holds, and sudden carrier capacity cancellations. The client operated 14 international fulfillment hubs across APAC and Europe, where manual logistics dispatchers took 24 - 36 hours to detect inventory bottlenecks and negotiate alternative shipping routes.
This latency led to **$2.4M in annual inventory holding penalties**, missed customer SLAs, and excessive reliance on costly expediting fees.
Rohit designed an autonomous **CrewAI + LangGraph State Graph** architecture with persistent vector memory in **Qdrant**:
Polls ERP streams every 15 minutes, detecting low inventory thresholds and calculating buffer safety days.
Queries 30+ carrier APIs to compare real-time transit times, container availability, and freight rates.
Audits import tariffs, customs risk scores, and budget constraints before issuing booking orders.
Triggers REST API calls directly to SAP S/4HANA to issue purchase order modifications autonomously.
"Rohit transformed our supply chain dispatch operations. What used to take our human dispatch team 30+ hours to analyze now happens in under 4 minutes autonomously. His agentic workflow cut our shipment delay penalties by $1.4 Million in the first year alone. Rohit is a world-class AI architect."
Schedule a 1-on-1 technical session directly with Rohit to review your data architecture.
Book Consultation