Retrieval-augmented generation (RAG)
Retrieval-augmented generation (RAG): A technique in which relevant documents are retrieved from a company's own data and supplied to a language model so its answer is grounded in, and can cite, those sources.
Why it matters
Most RAG quality problems are retrieval problems: the right passage was never found, or was split badly when documents were indexed. Evaluate retrieval separately from generation, show sources to users so they can check answers, and keep the document index fresh, because answers can only be as current as what was retrieved.
Related terms
- Vector database: A data store that indexes embeddings (numerical representations of text or images) so semantically similar content can be retrieved quickly, commonly used in RAG.
- Grounding: Constraining an AI model's output to verifiable source material, such as retrieved documents or database records, to reduce unsupported claims.
Where it comes up
In practice
In a Sudo Sapient sprint, this is a building block of AI Product Development: agents, copilots and AI-native features shipped in 6–10 weeks. Sudo Sapient's services describe the scope.