Glossary
RAG
Retrieving relevant passages from your own documents and giving them to the model, so answers are grounded and citable.
- Also called
- retrieval-augmented generation
Retrieval-augmented generation splits the problem in two: find the right passages, then write an answer from them. Because the answer is built from retrieved text, it can carry a citation back to the source.
The accuracy of a RAG system usually lives in the retrieval half. When these systems give wrong answers, the model was typically fine and the right passage was simply never fetched.
Permissions are the requirement most often discovered late. If the assistant retrieves across a corpus where different users have different entitlements, filtering has to happen at retrieval time, bolting it on afterwards means the model has already seen material the user cannot access.
Commonly misunderstood: RAG is often proposed as an alternative to fine-tuning, but they solve different problems. RAG handles knowledge that changes and must be cited; fine-tuning handles style, format and task behaviour.
Related terms, in context
The concepts you almost always meet alongside rag.
- Vector database
- A store for embeddings that answers 'what is semantically closest to this?' quickly.
- Embedding
- A numeric vector representing text, images or audio, positioned so similar things sit close together.
- Chunking
- Splitting documents into passages small enough to retrieve precisely but large enough to stay meaningful.
- Hallucination
- A model producing fluent, confident output that is not true, the failure mode that makes evaluation non-optional.
Where this shows up in our work
RAG is not an abstraction for us. It is a decision we make on live projects. It shows up most directly in rag & knowledge retrieval, ai search implementation, where getting it wrong has a cost someone can measure.
If you are evaluating a vendor on this, the useful question is not whether they can define the term. It is what they measure, what they would refuse to do, and what happens in their system when the assumption behind rag stops holding.
Questions
What is RAG?
Retrieving relevant passages from your own documents and giving them to the model, so answers are grounded and citable.
What do people get wrong about rag?
RAG is often proposed as an alternative to fine-tuning, but they solve different problems. RAG handles knowledge that changes and must be cited; fine-tuning handles style, format and task behaviour.
Does Orqent Labs build this?
Yes, RAG & Knowledge Retrieval Systems and AI Search Implementation. We work across India, covering all 19,238 PIN codes remotely.
Building something that involves rag?
We will tell you honestly whether it is the right approach for your problem.
Or email bd@dtrasglobal.com · call +91 74118 77878