LEARNER · GLOBAL
What is Retrieval-Augmented Generation (RAG)?
RAG is a technique that feeds an LLM external documents (e.g. compliance rules, customer data, market reports) *at query time* instead of baking them into training. LLM retrieves relevant chunks, then generates answers citing fresh sources. Reduces hallucination.
WHY IT MATTERS · EDITORIAL ANALYSIS
RAG powers production AI in banking: it keeps compliance chatbots up-to-date with new regulations, lets risk systems cite source documents, and avoids costly model retraining. Essential for regulated BFSI.