What Most Enterprises Get WRONG About RAG & AI Agents

Cassie Shum presents a deep-dive on why traditional vector-based RAG fails in enterprise environments and how semantic knowledge graphs provide a solution. The talk covers building GraphRAG pipelines that extract entities from unstructured text and bind them with business logic, while also demystifying AI agents as mere orchestrators. Designed for enterprise architects and ML engineers looking to implement production-grade GenAI systems with better explainability and multi-hop reasoning capabilities.