Priori AI

Local intelligence. Clinical confidence.

About Priori AI

Priori AI simplifies prior authorization by bringing AI to the point of care – but with a crucial difference. Our solution runs locally on physicians' devices, ensuring patient data stays private while dramatically improving the efficiency of insurance approval workflows.

PrioriAI_Demo_8VC_Hackathon.mp4

Our Solution

Priori AI transforms the process of prior authorization from days to mere minutes by:

• Running locally on physicians' devices for maximum privacy protection
• Integrating directly with major EHR systems like Epic and Oracle Cerner
• Analyzing physician documentation for gaps and errors
• Providing real-time suggestions to improve approval probability
• Generating complete prior authorization forms with significantly higher chances of first-pass approval

Key Features

• EHR Integration: Seamlessly connects with Oracle Health (Cerner) and Epic
• Local Processing: Edge-optimized LLaMA 3.1 8B model runs on-device without remote servers
• Dual-Agent Architecture:
  • Agent 1: Analyzes physician documentation and provides suggestions
  • Agent 2: Applies RAG against clinical policies for compliance with payer requirements
• Agentic GAN Approach: Iteratively refines documentation to maximize approval probability
• Approval Probability Estimation: Provides physicians with concrete improvement suggestions and estimated approval probability

System Architecture

Our innovative architecture consists of:

1. EHR Integration Layer (Epic, Oracle Cerner)
2. Physician Interface for drafting prior authorization requests
3. Agent 1 powered by LLaMA 3.2 11B model running locally
4. Agent 2 utilizing a specialized RAG system against insurer clinical policies
5. Iterative Refinement Cycle between Agents 1 and 2 to maximize approval chances

Real-World Example

Priori AI significantly improves the prior authorization process for procedures like Myocardial Perfusion Imaging which can be very expensive for patients. What is typically an opaque, error-prone process with frequent dead ends becomes automated, clear, and completed in seconds.

Benefits

• Higher approval rates
• Faster processing speed
• Lower administrative burden
• Higher patient satisfaction
• Complete patient data privacy

Team

• Varun Madan: Stanford Engineering
• Madhuhaas Gottimukkala: Stanford Engineering
• Krish Sharma: Stanford Engineering
• Jeff Liu: Stanford Engineering

Getting Started

1. Select your EHR system (Cerner or Epic)
2. Search and select patient
3. Start authorization
4. Follow the guided process through our intuitive interface
5. Receive AI-assisted suggestions and improvements
6. Submit your completed prior authorization with confidence

What's Next

Future development includes a dashboard for tracking all progress of prior authorization for each doctor.