RAG knowledge assistant for internal documents

A retrieval-augmented generation (RAG) assistant that lets staff ask questions in plain English and get answers drawn from the organisation's own policies, manuals and contracts, with links to the source.

Our role: Lead developer & team lead

RAG pipelines
Vector search
Source citations
Access control

What is our notice period for contractors?

Contractor policy · §4.2 Staff handbook · p.18

The problem

Staff wasted time hunting through shared drives and long documents, and answers given from memory were often out of date.

What we built

1 Ingestion of PDFs, Word documents and intranet pages, split and indexed for search
2 Vector search combined with keyword search to find the most relevant passages
3 Answers that cite the source document and section, so they can be checked
4 Permission-aware retrieval, so people only see content they are allowed to access
5 Feedback and evaluation tooling to measure answer quality over time

Services behind this project

AI & Automation Data & Analytics

Described by sector and capability, without client names, screenshots or confidential details. The illustration is our own.

Previous project AI customer support chatbot Next project B2B marketplace for industrial supplies

Planning something like this?

Tell us what you want to build. We will share how we would approach it, what to build first, and what we learned building AI solutions before.

Start a conversation

We use essential cookies only. They keep the site secure and the contact form working. No advertising or tracking cookies. Read our cookie policy.

Cookie policy