GeNotes
GeNotes: commissioned by NHS England National Genomics Education programme.
The GeNotes system incorporates an advanced AI-driven chatbot that employs Retrieval-Augmented Generation (RAG) — a method that combines large language models with authoritative reference material. This approach ensures that all responses are grounded in the verified GeNotes dataset, providing consistent, transparent, and evidence-based information.
Built using the robust library LangChain, the system efficiently identifies and delivers relevant content in response to user queries.
Planned developments include several key innovations designed to enhance accuracy and usability:
- Graph-based retrieval to better represent relationships between genetic conditions, tests, and clinical guidelines.
- Automated guideline summarisation using large language models to generate concise, user-friendly insights.
- Query rewriting to improve the interpretation and precision of complex clinical questions.
The underlying codebase has been intentionally designed to be adaptable and scalable, supporting potential extensions beyond genomics. This flexible architecture could be applied to other use cases, such as clinical guideline retrieval, HR policy access, or education and training resources across the NHS.
Opportunities for collaboration with other teams and programmes are in discussion, with the shared goal of scaling and productising these solutions to deliver broader value across the health and care system.