Critical information has to be accurate, fast, and easy to retrieve

This case study shows how healthcare and pharmaceutical information can be structured so users can retrieve accurate drug guidance from partial names, packaging, or images during live care journeys. The work focuses on safer access to trusted information rather than uncontrolled summarisation. By creating a governed retrieval layer, the system can connect image or text inputs to validated source material, return understandable information in context, and reduce delays for patients, carers, and professionals who need reliable answers quickly.

Challenge

• Patients and professionals needed quick access to accurate medicine information.
• Leaflet and drug information had to be retrieved safely from trusted sources.
• Delays or inconsistency in access increased operational risk in live care journeys.
• Users were often working from packaging, partial names, or images rather than a clean structured identifier.
• Information had to be understandable in context, not just technically retrievable from a document repository.
• The organisation needed a way to support faster access without creating uncontrolled summaries or unsafe interpretation.

Approach

• Structured leaflet data into a governed pharmaceutical knowledge base.
• Designed a photo-led lookup flow that connected the user to accurate medicine information.
• Delivered the retrieval experience through an API-powered RAG system grounded in governed data.
• Mapped the most common user intents, including dosage questions, warnings, usage instructions, and confirmation that the correct medicine had been identified.
• Normalised leaflet content into retrievable units so the system could return relevant sections instead of forcing users to parse full documents.
• Designed the lookup flow to preserve traceability back to the source content and reduce ambiguity in the returned answer.
• Kept the system grounded in approved pharmaceutical data so retrieval remained governed even when the user journey began with an image.

Outcome

• Faster access to drug information for both patients and professionals.
• Better consistency in information retrieval from trusted sources.
• Reduced risk in how medicine information was surfaced and used.
• Improved usability in live situations where the user did not start with perfect input.
• More dependable responses across repeated high-volume information requests.
• A stronger model for governed retrieval in patient-facing and professional-facing care journeys.

Real-world example

A healthcare organisation needed patients and professionals to access accurate drug information quickly. We structured leaflet data into a governed knowledge base and built a photo-led lookup flow that returned clear medicine information through an API-powered RAG system.