ClaimsAgent - Healthcare Claims Intelligence Platform

An AI-powered healthcare claims processing system that predicts claim outcomes, provides billing guidance, and applies automated corrections using LangChain and Microsoft Dataverse.

Features

• Claim Prediction: AI-powered prediction of claim approval/denial with confidence scores
• Billing Guidance: Interactive copilot for billing SOP queries
• Auto-Correction: Rule-based claim correction engine
• Analytics Dashboard: Visual insights into claim patterns and trends
• Microsoft Authentication: Secure Dataverse integration via MSAL

Tech Stack

Backend

• Python 3.x with FastAPI
• LangChain for LLM orchestration
• OpenAI GPT-4.1-nano for predictions
• Microsoft Dataverse for claim data storage
• Azure Identity for authentication

Frontend

• React with Vite
• TailwindCSS for styling
• MSAL React for Microsoft authentication
• Axios for API communication

Setup

Prerequisites

• Python 3.8+
• Node.js 16+
• OpenAI API key
• Microsoft Dataverse instance
• Azure AD app registration (for MSAL)

Backend Setup

1. Clone the repository

   git clone <repository-url>
   cd claimopsagent

2. Create virtual environment

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

   pip install -r requirements.txt

4. Configure environment variables

   cp .env.example .env

   Edit .env and add your credentials:

   OPENAI_API_KEY=your-openai-api-key
   DATAVERSE_URL=https://your-org.crm.dynamics.com
   LANGCHAIN_TRACING_V2=false
   MOCK_MODE=true  # Set to false for production
   

   💡 Mock Mode: When MOCK_MODE=true, the application returns mock data without making real API calls to Dataverse or OpenAI. Perfect for local testing!

5. Run the backend

   uvicorn app.api:app --reload --port 8000

Frontend Setup

1. Navigate to UI directory

   cd ui

2. Install dependencies

   npm install

3. Configure environment variables

   cp .env.example .env

   Edit ui/.env and add your MSAL configuration:

   VITE_MSAL_CLIENT_ID=your-azure-app-client-id
   VITE_MSAL_AUTHORITY=https://login.microsoftonline.com/your-tenant-id
   VITE_DATAVERSE_SCOPE=https://your-org.crm.dynamics.com/.default
   

4. Run the frontend

   npm run dev

5. Access the application Open http://localhost:5173 in your browser

Project Structure

claimopsagent/
├── app/
│   ├── api.py              # FastAPI endpoints
│   ├── chains/             # LangChain processing chains
│   │   ├── prediction.py   # Claim outcome prediction
│   │   ├── correction.py   # Auto-correction logic
│   │   ├── guidance.py     # Billing guidance copilot
│   │   └── analytics.py    # Analytics generation
│   ├── dataverse/          # Dataverse integration
│   │   ├── client.py       # Dataverse API client
│   │   └── tools.py        # LangChain tools
│   ├── engine/             # Business logic
│   │   └── rules_engine.py # Correction rules
│   └── core/               # Core utilities
│       ├── config.py       # Configuration
│       └── telemetry.py    # Observability
├── ui/                     # React frontend
│   ├── src/
│   │   ├── pages/          # Page components
│   │   ├── components/     # Reusable components
│   │   └── authConfig.js   # MSAL configuration
│   └── package.json
├── requirements.txt        # Python dependencies
└── .env.example           # Environment template

Usage

Claim Analysis

1. Sign in with your Microsoft account
2. Navigate to "Claim View"
3. Enter a claim ID (UUID or string format)
4. Click "Analyze Claim" to get AI predictions
5. Apply auto-corrections if needed

Billing Guidance

1. Navigate to "Copilot"
2. Ask questions about billing procedures
3. Get instant SOP guidance

Analytics

1. Navigate to "Dashboard"
2. View claim trends and insights
3. Export reports as needed

Development

Mock Mode

For local development and testing without real API connections:

• Set MOCK_MODE=true in .env
• Mock data will be returned for Dataverse queries
• OpenAI calls will be skipped with mock predictions
• Perfect for testing the UI and workflows

Note: You can still use a dummy Dataverse URL if needed, but MOCK_MODE is the recommended way to enable mock responses.

Running Tests

pytest tests/

Security Notes

• Never commit .env files - they contain sensitive credentials
• Use .env.example as a template
• All secrets should be stored in environment variables
• MSAL tokens are handled securely by the browser

License

[Your License Here]

Contributing

[Your Contributing Guidelines Here]