LEARNER · GLOBAL
What is RAG? Retrieval-Augmented Generation lets AI look up fresh information instead of relying only on training data
RAG (Retrieval-Augmented Generation) is a technique where an AI model pauses before answering, searches a database or web for relevant documents, reads them, then generates a response informed by what it found. Think: a chatbot that fact-checks itself mid-answer using a library. Financial example: a risk assistant that retrieves the latest SEC filings before summarizing a company's leverage. Without RAG, the model would only know facts from its training data, which gets stale. With RAG, it stays current and can cite sources.
WHY IT MATTERS
For compliance, audit, and risk teams, RAG-enabled systems can provide auditable, up-to-date answers because they show which documents were retrieved. No 'black box' hallucination.
Source: Illustrative · 2026-06-02