Empowering a large energy organisation with secure AI-powered document search
A leading UK energy organisation manages a vast library of technical and operational documentation (engineering standards, safety procedures, project specifications) and finding the right information quickly often relied on specialist staff with deep organisational knowledge. They wanted a secure, AI-powered search experience that could surface accurate, contextually relevant answers while staying fully compliant with stringent security and governance policies, operating entirely within their internal network and supporting multiple formats (PDFs, Word, CAD specs and scanned images).
Agile delivered a secure Retrieval-Augmented Generation (RAG) solution using an open-source LLM hosted in the client’s Azure environment. Documents were ingested and vectorised (with OCR where needed) into a metadata-rich vector database, enabling grounded, citation-backed answers without data ever leaving the secure estate. Deployed as a web app integrated into the internal knowledge portal, the solution reduced average time to locate and validate technical standards from 20+ minutes to under 2 minutes, while reducing dependency on subject matter experts for routine queries.
