Ducal Navigator transforms your institutional documents, databases and field sources into an auditable knowledge graph. Every concept traces to its source. Queryable in natural language by your coordinators.
Care pathway coordinators spend hours searching for the right information across scattered sources.
Incomplete directories, scattered institutional documents, unshared tacit knowledge. Each coordinator rebuilds their own information base, with gaps and duplicates.
Standard AI tools (RAG) search text fragments and hallucinate. In care coordination, a wrong referral can have serious consequences. AI must be auditable, not approximate.
No tool provides a view of relationships between organizations, professionals, pathways and care offerings across a territory. Coordination decisions are made without a global view.
Navigator doesn't search text fragments. It builds a structured, validated representation of knowledge, queryable by any professional.
Every concept, relationship and fact traces to its source document. Contradictions between sources are explicitly identified, not masked. An auditor can trace any answer back to the original document.
Our pipeline draws from Grounded Theory: multi-pass coding, constant comparison, saturation detection. AI structures the knowledge. Humans validate. The result: a rigorous graph, not a black box.
Coordinators describe a patient's situation in plain language. Navigator suggests the most relevant referrals with sourced justifications. No query syntax needed.
Before a coordination meeting, Navigator generates a concise briefing: patient context, available care offerings on the territory, referral options with their sources. Coordinators save hours of preparation.
Institutional documents, public databases (ROR, RPPS), field interviews, regional health agency reports, best practice guidelines. Navigator ingests and structures any source type to build a unified view of the territory.
End-to-end traceability, answer explainability, hallucination-free by design. Navigator is built to meet EU AI Act requirements in a YMYL healthcare context.
Coordination Support Platforms — structure your territory's knowledge to refer faster and better.
Territorial Professional Health Communities — share a common view of care offerings among all your members.
Regional Health Agencies — steer your territorial policies with a living, sourced map of the care supply.
Hospital Territory Groups — coordinate patient pathways between facilities with a shared knowledge base.
A RAG chatbot searches for text chunks similar to your question and sends them to an LLM that improvises an answer. Navigator first builds a structured, validated knowledge graph, then reasons over it. Every answer traces to its sources, and contradictions between documents are explicitly flagged.
Any type: PDFs, Word documents, institutional reports, structured databases (ROR, RPPS), interview transcripts, best practice guidelines, meeting minutes. The analysis pipeline adapts to the format and nature of the content.
Navigator is in an advanced proof-of-concept phase, with a first deployment in health coordination in Île-de-France. Contact us to discuss a pilot in your territory.
Yes. Like all Ducal modules, Navigator is hosted on sovereign HDS infrastructure. No data transits through foreign servers. The production AI model is Mistral, hosted in France.
Let's talk. We'll get back to you within 48 hours with an initial proposal tailored to your needs.
Contact Us →