NHS — South Yorkshire ICB · Healthcare
Clinical Knowledge Assistant
Delivery context
Personally delivered by Raga AI founder Usman Bukhari through Usman Bukhari Ltd before establishing RAGA AI LTD.
- Sector
- Healthcare
Patients and staff needed reliable answers drawn from clinical guidelines, policy documents and regional service information.
Business challenge
A greenfield build in a clinical setting has no tolerance for hallucination or stale information, while service discovery was scattered across dozens of regional sites.
What was delivered
- A clinical knowledge assistant grounded in approved NHS guidance, with answers traceable to their source material
- Real-time web grounding blended with proprietary knowledge so responses reflect current regional information
- Policy documents made searchable through OCR and combined keyword and meaning-based search, shortening the time to find an answer
- The assistant exposed inside the NHS South Yorkshire GP app for patients and staff
- An automated pipeline keeping regional service information up to date across many local sources
Security, governance and operational considerations
- Answers grounded in approved sources, with no hallucinations observed across the agreed evaluation dataset
- Retrieval kept within approved knowledge sources, with sanitisation applied before content reached the assistant
Technical implementation
- Azure AI Agent SDK with LangGraph state orchestration, LangChain retrieval chains and LlamaIndex sanitisation
- Document Intelligence OCR over policy PDFs into Azure AI Search with hybrid lexical and semantic retrieval
- MCP server integration into the NHS South Yorkshire GP app
- Event-driven Azure Functions pipeline with Crawl4AI collection across regional sites
Technology stack
- Azure AI Search
- LangGraph
- Azure Functions
- Crawl4AI
Related service: AI implementation services
Facing something similar?
Request a discovery conversation and we will talk through your workflow, the systems involved and whether this approach would work for you.