01
Where this one gets misread
RAG grounds an answer in retrieved material. It does not verify the answer. A confident, well-written response built from irrelevant passages is the ordinary failure mode, not the exotic one.
Glossary
What is retrieval-augmented generation (RAG)?
Retrieval-augmented generation fetches source material and hands it to a model to answer from. It grounds the answer without proving the answer correct.
By Marella AI · Updated 21 Aug 2026
02
Questions to ask
Ask for a worked example on your own material, and the evidence needed to reproduce it.
- What was retrieved for this specific answer?
- Does each claim map to a passage?
- What happens when retrieval returns nothing useful?
- How is that case tested rather than assumed?
03
How Marella uses the term
We use “Retrieval-augmented generation (RAG)” only where a product mechanism or an evaluation method backs it up, and we say when the behaviour depends on how a deployment is configured.
- Backed by a product mechanism or an evaluation method
- Deployment differences flagged
Caveats
- B1A definition is not a claim about how the product performs.
- B2Vendor implementations vary.
- B3Test the term against a representative workflow.
Related pages
Test it on your own documents
Pick a real piece of work, agree what a good answer looks like, then go through the results together.