Skip to content

starlite20/compiled2

Repository files navigation

MedFlow UAE — Hackathon Submission

1. Intro

  • Team Name: Team MedFlow
  • Project Name: MedFlow UAE
  • One-liner: An agentic healthcare orchestrator that transforms fragmented hospital visits into a seamless, AI-guided city journey.

2. Problem

In the UAE, world-class medical care is often hidden behind fragmented operational silos. Patients are forced to manually navigate a "bureaucratic triathlon"—independently managing hospital discovery, document verification, physical queues, and pharmacy fulfillment. This reactive process creates friction, anxiety, and inefficient utilization of city-wide healthcare capacity.


3. Demo

Focus: The Patient Journey Our demo showcases a high-fidelity, hybrid interface where an AI Agent and a visual Journey Dashboard work in perfect sync:

  • Pre-Visit: The AI agent validates Emirates ID and Insurance documents via chat, unlocking the "Ready" status before the patient even leaves home.
  • During-Visit: A real-time virtual queue tracker replaces waiting-room uncertainty with a live, visual timeline of the consultation stages.
  • Post-Visit: Instead of waiting at a crowded hospital pharmacy, the agent routes prescriptions to the patient's preferred city-wide provider (e.g., Aster or Life), ensuring medicines are ready for pickup immediately.
  • The Outcome: A calm, predictable experience where the city’s infrastructure adapts to the patient’s intent.

4. How It Works

The Tech Stack

  • Frontend: Next.js with a "Hybrid UI" (Left pane: Agent Chat | Right pane: Visual State Dashboard).
  • Agent Brain: n8n serves as the orchestrator, using webhook-driven workflows to manage state logic and decision-making.
  • Intelligence: LLMs (via n8n) to handle intent classification and structured JSON outputs.
  • Localization: Lingo.dev for native, culturally-aware English and Arabic adaptation.

Interesting Technical Choices: We moved away from static forms in favor of Agentic Flow Control. The AI doesn't just provide info; it actively controls the SPA state—locking or unlocking journey stages based on document verification and real-time hospital "load" data.


5. Wrap-up

What’s Next? We plan to move beyond mock data by integrating with the Malaffi health information exchange and implementing real-time GPS-based "Arrival Triggers" to automate the check-in process the moment a patient enters the hospital geofence.

What we’d build with more time: We would develop a City-Wide Healthcare Load Balancer. This predictive engine would suggest hospitals not just based on distance, but by analyzing real-time department wait times across the city to optimize patient distribution.


Submitted for the Adaptive City — Healthcare Track

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors