Prerequisite: None
Retrieval-Augmented Generation (RAG) is the preferred architecture for deploying trusted AI systems in enterprises, but governance often lags when teams treat compliance as an afterthought. However, governance is crucial for RAG systems to ensure reliable and grounded AI outputs, especially in high-stakes situations.
Addressing hallucinations, data quality differences, and evaluation framework limitations are key challenges in governing RAG systems. So this session focuses on practical issues such as designing for verifiable outputs and auditing knowledge bases.
We'll also consider the trade-offs organizations face when balancing rapid experimentation with the risk controls demanded by regulated environments.
This session provides a practical framework for teams who need their systems to work reliably, not just impressively.