Production-Grade Healthcare Prior Authorization Platform powered by Agentic AI, Enterprise RAG, LangGraph Orchestration, and UiPath Automation Integration.
HealthAuthAI Enterprise is a modern healthcare prior authorization automation platform designed to streamline payer validation workflows using:
- Agentic AI orchestration
- Retrieval-Augmented Generation (RAG)
- GPT-powered clinical reasoning
- ChromaDB semantic retrieval
- FastAPI enterprise APIs
- PostgreSQL persistence
- UiPath RPA integration architecture
The platform combines enterprise AI workflow automation with explainable healthcare authorization reasoning to modernize legacy authorization operations.
Stateful multi-agent workflow orchestration powered by LangGraph.
Implemented AI agents include:
- Entity Extraction Agent
- Policy Retrieval Agent
- GPT Validation Agent
- Decision Agent
The platform supports:
- Dynamic TXT/PDF policy ingestion
- Multi-document semantic retrieval
- ChromaDB vector indexing
- Metadata-aware filtering
- Specialty-aware policy routing
Supported specialties currently include:
- Orthopedic workflows
- Cardiology workflows
Integrated with GitHub Models GPT-4o for:
- clinical reasoning
- policy validation
- compliance scoring
- missing documentation detection
- authorization recommendations
Supported authorization outcomes:
- APPROVED
- REQUEST_INFO
- DENIED
The system generates:
- compliance scores
- AI reasoning summaries
- audit-ready workflow logs
- request tracking identifiers
Workflow execution data is persisted into PostgreSQL including:
- patient references
- authorization outcomes
- compliance scores
- reasoning summaries
- audit logs
- workflow execution state
Operational dashboard APIs provide:
- approval trend analytics
- workflow throughput metrics
- compliance score distribution
- operational visibility
- audit monitoring
The architecture supports enterprise RPA integration workflows where UiPath bots can:
- monitor intake channels
- upload authorization requests
- trigger FastAPI endpoints
- receive AI decisions
- update legacy systems
- send notifications
Target integration systems:
- SAP Healthcare ERP
- Epic EHR
- Cerner Millennium
- payer authorization portals
| Layer | Technology |
|---|---|
| Backend API | FastAPI |
| AI Orchestration | LangGraph |
| LLM Engine | GitHub Models GPT-4o |
| Vector Database | ChromaDB |
| Database | PostgreSQL |
| Containerization | Docker |
| Orchestration | Kubernetes (Planned) |
| Automation | UiPath |
| Documentation | Swagger / OpenAPI |
| Capability | Status |
|---|---|
| FastAPI Backend | IMPLEMENTED |
| LangGraph Workflow | IMPLEMENTED |
| GPT Validation | VALIDATED |
| ChromaDB RAG | IMPLEMENTED |
| Multi-Policy Retrieval | IMPLEMENTED |
| Metadata Filtering | IMPLEMENTED |
| PostgreSQL Persistence | IMPLEMENTED |
| Dashboard APIs | IMPLEMENTED |
| UiPath Integration Design | ARCHITECTURALLY DESIGNED |
| Kubernetes Deployment | PLANNED ENHANCEMENT |
| Front-End Visualization | PLANNED ENHANCEMENT |
| Method | Endpoint | Purpose |
|---|---|---|
| POST | /agentic/prior-auth |
Executes complete Agentic AI authorization workflow |
| GET | /agentic/dashboard |
Retrieves operational analytics metrics |
| POST | /intake/document |
Uploads healthcare intake documents |
| GET | /health |
Health probe endpoint for Docker/Kubernetes |
| POST | /auth/login |
Authentication gateway |
{
"request_id": "86f0bb7f-4dbf-4058-b820-49d14fc1df15",
"patient_name": "John Doe",
"recommendation": "APPROVED",
"compliance_score": 95,
"missing_items": [],
"reasoning_summary": [
"Insurance validated successfully",
"Diagnosis documentation provided",
"Policy requirements satisfied"
]
}
Detailed installation and execution instructions are available in:
Quick Start:
# Build and start containers
Docker Compose Up --buildOR local execution:
# Activate virtual environment
venv\Scripts\activate
# Start FastAPI server
uvicorn main:appAfter startup:
http://127.0.0.1:8000/docs
{
"claim_text": "Patient Name: John Doe\nInsurance ID: INS9000\nProcedure: Knee Replacement\nDiagnosis: Severe Knee Osteoarthritis"
}Implemented retrieval features include:
- multi-document semantic search
- metadata-aware retrieval
- specialty-based filtering
- dynamic TXT/PDF ingestion
- ChromaDB persistence
- policy indexing pipeline
Future roadmap items include:
- Live graph execution visualization
- OCR extraction pipelines
- Kubernetes production deployment
- Enterprise IAM/OAuth2 integration
- Real-time observability
- SMTP notification services
- SAP/Epic production connectors
- Streamlit/React dashboard UI
The platform demonstrates a modern enterprise AI architecture combining:
- Agentic AI orchestration
- Enterprise RAG engineering
- GPT-powered healthcare reasoning
- Explainable AI workflows
- Healthcare automation
- RPA integration compatibility
- Audit-ready persistence
- Metadata-aware semantic retrieval
The solution architecture aligns strongly with:
- healthcare automation modernization
- payer authorization transformation
- intelligent document processing
- AI-assisted clinical operations
- enterprise workflow orchestration initiatives
This repository currently represents a validated enterprise prototype architecture and technical demonstration platform for healthcare prior authorization automation workflows.
Some components described in this repository are roadmap-oriented architectural enhancements and may not yet represent fully production-deployed implementations.
Internal Enterprise Prototype / Solution Architecture Demonstration