Navigator-AI

AI-powered healthcare referral automation tackling the $150B annual industry crisis. Automates document extraction to appointment confirmation using Raindrop orchestration and Vultr cloud deployment.

Comment

Inspiration

Healthcare systems in the U.S. are hemorrhaging over $150 billion annually due to referral mismanagement. Nearly half of all referrals never complete, costing individual hospitals $200-500M per year while patients waste $1.9B in unnecessary expenses. We set out to solve this crisis with AI-powered automation.

What it does

Navigator-AI automates the entire healthcare referral workflow from document upload to appointment confirmation:

  • AI-Powered Extraction: Automatically extracts patient and provider information from referral documents using multi-model AI (95%+ accuracy)
  • Smart Orchestration: Uses Raindrop's workflow engine to intelligently route referrals, schedule appointments, and track status
  • Real-Time Dashboard: Provides coordinators with live visibility into every referral's progress
  • Patient Portal: Enables patients to confirm appointments and receive notifications
  • Cloud-Native: Deployed on Vultr infrastructure for enterprise scalability

How we built it

Technology Stack:

  • Backend: Node.js API server with Raindrop integration (SmartSQL, SmartMemory, AI, Object Storage)
  • Frontend: Next.js with React for responsive UI
  • AI/ML: Raindrop's multi-model AI for document extraction
  • Cloud Infrastructure: Vultr for production deployment (VM, storage, networking)
  • Database: PostgreSQL via Raindrop SmartSQL
  • Storage: Raindrop Object Buckets for document management

Architecture: Mono-repo structure with separate frontend, backend, and extraction service components. Backend handles RESTful API endpoints documented with OpenAPI, while frontend provides intuitive upload forms and real-time dashboards.

Challenges we ran into

  • Document Variability: Medical referral forms come in dozens of formats - we had to train our extraction pipeline to handle diverse layouts
  • State Management: Coordinating complex workflows across multiple stakeholders required careful orchestration design
  • HIPAA Compliance: Ensuring secure data handling throughout the pipeline
  • Cloud Deployment: Setting up production-ready infrastructure on Vultr with proper networking and security

Accomplishments that we're proud of

✅ End-to-end workflow completion in under 2 minutes ✅ 95%+ extraction accuracy on structured medical forms ✅ Successfully deployed production system on Vultr cloud ✅ Addressing a $67.92B market opportunity (projected by 2034) ✅ API-first design ready for EHR integration ✅ Complete demo with real referral documents (anonymized)

What we learned

  • How to leverage Raindrop's powerful orchestration capabilities for complex healthcare workflows
  • Best practices for deploying scalable cloud infrastructure on Vultr
  • The critical importance of healthcare referral management and its massive economic impact
  • Techniques for handling diverse document formats with AI extraction

What's next for Navigator-AI

  • EHR Integration: Connect with Epic, Cerner, and other major EHR systems
  • Advanced Analytics: Predictive insights for referral leakage prevention
  • Multi-language Support: Expand to serve diverse patient populations
  • Mobile App: Native iOS/Android apps for coordinators and patients
  • Enterprise Features: Multi-facility management, custom workflows, advanced reporting
  • AI Improvements: Fine-tune models for specialty-specific referral types

Market Validation: With the patient referral management software market growing from $16.14B in 2025 to $67.92B by 2034 (17.31% CAGR), Navigator-AI is positioned to capture significant value while solving a critical healthcare problem.

Built With

Share this project:

Updates