TriageAI

TriageAI uses AI to combine patient scans, surveys, and clinical notes into real-time risk scores, so clinicians can quickly identify who needs urgent attention and make smarter scheduling decisions.

Comment

About the Project

Inspiration

We built TriageAI to reduce the time between a patient reporting symptoms and receiving meaningful clinical guidance.
The idea came from a real gap in care workflows: triage is often delayed, inconsistent, and difficult to scale when clinicians are overloaded. We wanted to design something that helps teams prioritize risk faster while keeping clinicians fully in control of final decisions.

How We Built It

We developed TriageAI as a full-stack platform with:

  • A patient intake and clinician dashboard web experience
  • Backend API routes for triage and AI-assisted survey generation
  • Supabase for authentication, profile automation, and clinical data storage
  • Structured workflows to turn symptom inputs into actionable risk signals
  • Guardrails so outputs support decisions instead of replacing clinical judgment

Our focus was to combine speed and usability with a safe, reviewable workflow clinicians can trust.

What We Learned

This project taught us that healthcare AI is less about flashy model features and more about reliability and accountability.
Some of our biggest lessons were:

  • Good data design is foundational to everything else
  • Prompt behavior needs testing and clear iteration loops
  • Transparency matters for clinician trust and adoption
  • UX quality directly impacts whether tools are used in real workflows
  • Human-in-the-loop design is essential in sensitive domains

Challenges We Faced

The hardest challenges included:

  • Converting messy symptom descriptions into structured, consistent inputs
  • Balancing rapid prototyping with responsible clinical boundaries
  • Handling ambiguous or incomplete patient responses
  • Designing a schema that is both flexible and auditable
  • Keeping AI output helpful without becoming overconfident or unsafe

Final Reflection

TriageAI is our attempt to make triage more proactive, consistent, and scalable.
The biggest takeaway was that meaningful health tech is built at the intersection of engineering quality, thoughtful product design, and clinical humility.

Built With

Share this project:

Updates