What is retrieval-augmented generation (RAG)?

Retrieval-augmented generation, or RAG, is a way to make AI answers trustworthy about your specific information. Instead of asking a language model to answer from memory alone — where it can confidently invent things — a RAG system first retrieves relevant passages, records or documents from your own data, then feeds them to the model so its answer is grounded in real source material. It's the standard approach for getting an AI to answer questions about content it was never trained on, like your internal documents.

iDBQuery uses retrieval to ground its work — for example, finding the right tables, columns and documents relevant to your question — but it's important to be clear about how it differs from a pure RAG chatbot over documents:

• It runs real queries, not just text retrieval. For structured questions, iDBQuery writes and executes actual SQL against your live data, so a number like "total revenue" is computed from your rows, not paraphrased from a nearby paragraph.
• It cites the exact source. Every figure traces back to the specific rows or documents it came from, which is stronger provenance than a typical RAG answer.
• It spans structured and unstructured data — databases, spreadsheets and warehouses alongside PDFs and folders of files.

So RAG is part of how modern grounded AI avoids making things up, and iDBQuery applies that grounding principle — then reinforces it by executing verifiable queries and citing every result.