RAG (Retrieval-Augmented Generation)
Answering a question by first retrieving the relevant passages from your own documents and then having the model answer only from them — rather than relying on what it memorised in training.
Why it beats fine-tuning for most business cases
Fine-tuning teaches a model style and format; it does not reliably teach it facts, and it has to be redone every time the facts change. Retrieval keeps the facts in a searchable index, so a new contract is answerable the moment it is uploaded.
The part that matters in an audit
Because the answer is generated from retrieved passages, each statement can cite the document and page it came from. An answer you cannot trace is an answer you cannot use in a regulated process.
Related terms
Where this applies
StancorIQ
Need help applying this?
Stancor Technologies engineers answer specification questions directly. Contact us.