How Tesco Uses GenAI to Understand a 40‑Year‑Old Warehouse Management System - Arshad Hameed

This talk explores Tesco's Warehouse Management System, which has been developed over four decades using mainframe COBOL and is underpinned by numerous interconnected business rules. The speaker discusses how Tesco employed agentic AI and retrieval-augmented generation to decode legacy code, reveal hidden dependencies, and reconstruct undocumented functional knowledge before any modernization efforts could take place. The emphasis is on accelerating understanding rather than automating modernization, utilizing structured RAG pipelines alongside AI-driven agents to provide engineering and product teams with a validated understanding of the system’s operations. Ultimately, this approach created a robust knowledge foundation that will guide future architectural and strategic decisions for this essential platform.