Generative AI adoption is exploding, but if your AI agents have blanket access, you’re basically handing them master keys—yikes! In this NDC Copenhagen talk, Ashish Jha explains why traditional access control can’t keep up with the contextual, document-level permissions needed when you’re doing RAG or agentic AI.He walks through how Fine-Grained Authorization (FGA) with tools like OpenFGA and LangChain locks down your systems, ensuring strict multi-tenant isolation, preventing data leaks, and providing full auditability. The best part? You can scale to billions of access decisions without breaking a sweat—absolutely essential whether you’re building an internal copilot or a customer-facing AI.