Build · Do Directory
Top RAG & search experts, active now
Retrieval-augmented generation (RAG) lets an AI answer from your own material: manuals, policies, past tenders, product data or a shared drive. The system finds the relevant passages first, then writes an answer from them, ideally with links back to the source.
Quality depends on the unglamorous parts: cleaning and splitting documents well, keeping them up to date and respecting who is allowed to see what. Ask candidates how they measure retrieval accuracy.
0 people
Questions
RAG & search: common questions
Do I need RAG or fine-tuning?
If the AI needs to know facts from your documents, and those facts change, RAG is usually the right first step. It’s cheaper, easier to update and can show its sources. Fine-tuning is better for teaching a consistent style or format.
Can it respect who is allowed to see which documents?
It can and it should. A good specialist will carry your existing permissions through to search, so a junior member of staff can’t surface HR files or board papers. Ask how they handle this before any documents are loaded.
What should I prepare before a project starts?
A sensible set of real documents, a list of 30 to 50 questions people genuinely ask with the answers you’d expect, and someone who knows the material and can judge whether answers are right. That question list becomes the test the system has to pass.
Free to join
Good with AI? List yourself.
Tell us what you build. We review every listing, usually within 2 working days, and you stay visible by confirming you're still active once a month.