Concept
Most AI systems optimise for agreement.
We propagate disagreement.
That is where signal lives.
Problem
Clinical decisions are rarely simple. Guidelines standardise care, but complex cases expose uncertainty.
MDTs are essential, but they can be slow, inconsistent, and vulnerable to groupthink.
System
Input → Independent Agents → Structured Debate → Moderation → Revision → Synthesis
Clinical focus
Built first for cancer MDT decision-making. Initial domain: Thyroid cancer.
Architecture
Multiple independent agents reason from the same case. They challenge assumptions.
A moderator evaluates competing arguments. The system produces a structured synthesis for clinician review.
Decision position
Clinical decision support. Not autonomous diagnosis. Clinician remains responsible.
Validation
Designed for real MDT cases. Initial testing with thyroid cancer clinicians.
Focus on uncertainty detection, reasoning transparency, and clinical usefulness.
Positioning
Not a chatbot.
Not diagnostic AI. Not a guideline engine.
A clinical reasoning architecture.
A new reasoning layer for multidisciplinary medicine.
Peregrine.Reason
Concept
System
Clinical