LEARNER · GLOBAL
What is RAG (Retrieval-Augmented Generation)? A plain-English explainer
RAG is a technique that gives AI models access to external documents (like a bank's loan files or compliance manuals) in real-time before answering questions. Instead of only using patterns learned during training, RAG lets models fetch and cite current, proprietary data—reducing hallucinations and keeping answers factual and audit-ready.
WHY IT MATTERS
RAG powers the most reliable enterprise AI deployments in banking: compliance Q&A bots, KYC-document summarization, and internal-policy finders. Understanding RAG is critical for risk teams evaluating whether a model can be trusted on regulated workflows.
Source: AITechHive · 2026-08-10