An AI-powered insurance claim validation and decision platform that automates claim assessment using OCR, rule validation, multi-agent reasoning, and premium reporting.
Frontend: Vercel
Backend: Render
Storage/Auth: Supabase
Claim Reflex AI helps insurance teams instantly validate policy and claim documents, detect mismatches, and generate AI-backed approval/rejection decisions with confidence scoring.
Instead of manual verification, the system uses:
- OCR text extraction from uploaded PDFs
- Intelligent policy vs claim validation
- Multi-Agent AI Debate Engine
- Final approval/rejection reasoning
- Downloadable premium PDF reports
- Clean enterprise-grade UI
- Upload Policy PDF
- Upload Claim PDF
- Preview uploaded files before processing
Uses 3-layer reasoning:
Finds evidence supporting approval.
Finds issues / risks / mismatches.
Compares both sides and gives final verdict.
- Policy Number Match
- Claim Number Presence
- Holder Name Fuzzy Match
- Policy Expiry Check
- Claim Amount Detection
- OCR Noise Handling
- Missing Fields Detection
- Animated processing states
- Smooth transitions
- Approval / rejection result cards
- Confidence progress bar
- Expandable AI Debate Chamber
- Recent analysis history
- Responsive mobile design
Generated after every analysis:
- Final decision
- Confidence %
- Validation summary
- Approval Agent opinion
- Rejection Agent opinion
- Final Judge verdict
- Audit Trail
- HTML
- CSS
- JavaScript
- Python
- Flask / FastAPI
- OpenAI API
- Supabase
- Vercel
- Render
- Upload Policy PDF
- Upload Claim PDF
- Click Analyze Claim
- AI validates documents
- Debate engine reasons internally
- Final result shown instantly
- Download premium report
frontend/
├── index.html
├── app.html
backend/
├── main.py
├── validator.py
├── ai_engine.py