Retrieval-Augmented Generation (RAG)
Retrieval-Augmented Generation (RAG) is an AI architecture that combines a large language model with an external knowledge source. Instead of relying only on what the model learned during training, a RAG system retrieves relevant documents, database entries, or knowledge-base content at runtime and uses that context to generate a more accurate answer.
A typical RAG workflow has three steps:
This makes RAG useful when answers need to reflect fresh business data, internal documentation, product catalogs, policies, or support content.
RAG helps reduce hallucinations, improves factual grounding, and allows teams to update answers without retraining the base model. It is widely used in AI search, enterprise chatbots, internal assistants, customer support tools, and knowledge management systems.