RAG systems

A model that has not read your documents will improvise, and improvising is expensive. Retrieval is how you make it quote instead of guess.

What it is

We build enterprise retrieval-augmented generation systems: your private corpus indexed, chunked, embedded and searched, with the retrieved passage handed to the model as the only context it is allowed to answer from.

The hard part is not the embedding. It is evaluation: knowing which question your retrieval fails on, and finding out before a customer does.

How we build it

  1. Fix the corpus before the index

    Duplicate policies, superseded versions and scanned PDFs sabotage retrieval more than any chunking strategy saves it. We clean first, because garbage retrieved confidently is worse than nothing retrieved.

  2. Build the question set before the pipeline

    We write the questions the system must answer, with the passage that should be retrieved for each, before tuning anything. Without it you are optimising a number you invented.

  3. Make refusal the default

    When retrieval comes back weak, the system says it does not know. A confident wrong answer from your own knowledge base is the fastest way to lose trust in the whole thing.

Who it is for

A good fit if

  • Organisations with a large internal knowledge base nobody can search.
  • Support and compliance teams that must answer from an authoritative source.
  • Anyone whose data cannot leave their own infrastructure.

Not a fit if

  • Corpora too small to justify retrieval. A long context window is cheaper.
  • Questions with no written answer anywhere in your documents.

See it

Hybrid RAG Knowledge VaultA live query against a fictional company, Harbourline Logistics, held in its own isolated store. Choose who is asking, then ask something above their clearance and watch the system refuse rather than leak it.

Everything here belongs to Harbourline Logistics, a company we made up for this demo. Every figure in it is sample data. Nothing you read is a real claim about Ejentic or a client.

Who is asking
A Customer can see public information only.

Demo only. Here you can switch roles freely, so you can see both the answer and the refusal. In a real deployment you can’t: your access is set by how you sign in to your own system, not a button on a page.

How clearance really works

The engine decides what you can see from a secret key your organisation’s login hands it on your behalf, never from a choice you make in the page. Signed in as a customer, the system only ever retrieves public material; staff and executive documents aren’t hidden from you, they’re never fetched. Ask for anything above your level and the request is refused outright, not quietly downgraded. The only way to read executive answers is to actually be an executive in your login system. Here, with a made-up company and no real users, we let you role-play all three so the boundary is visible.

A reference build

Hybrid retrieval over a private corpus

Keyword search and vector search fail differently, so the vault runs both and merges the results before the model sees anything. The point is not elegance: it is that a question phrased nothing like the document still finds the document.

  • Keyword and vector retrieval merged, not either-or
  • Every answer cites the passage it came from
  • Runs against a private corpus with no external call for the data itself

This is a system we run, not a client case study. We do not publish client numbers we cannot show you the workings for.

Talk to someone who builds this

No qualification call, no discovery deck. Describe the work and we will tell you whether this is the right line for it.

Start a conversationmail@ejentic.xyz

Other lines

AI agentsAutomationsConsultation