Claim Management
System Performance Dashboard
Patient Management
Dashboard
Agent Status Dashboard

Madza: Agentic Healthcare Claims Processing System

Problem Statement

While I was in the United States until a few months back, over a period of time I did see the challenges faced as a consequence of too much back and forth between Healthcare providers and Insurance with respect to claim processing leading to serious delays in payment realizations and unnecessary hours wasted behind followups. I wanted to come up with a AI/ML based solution to alleviate the problem and streamline the process.

Solution Overview

It tries to bring an Multi-Agent based Agentic AI Solution to the entire process right from Patient Registration, Claim Processing, Claim Management, Processing of Explanation of Benefits and ultimately the entire Claim Process. It also has a Chatbot built into it to assist end users with any query they might have.

Technical Architecture

I created a React Application with a Python Backend - AWS Bedrock Agent Core Client has been used for all AI Analysis as part of different flows like Patient Registration, Claim Processing, EOB Analysis, Suggestions etc. Strands Agents Framework from AWS for chatbot and EOB generation simulation. I have incorporated both research based and meta-tool approach so as to enable end users get a wide spectrum of benefits and having a true multi-agent system at their disposal.

Deployment Challenges

Deploying to AWS sometime ran into issues. Since I was always having cost in mind could not use some of the other AWS advancements for deployment etc.

Achievements

I am proud of Madza and that I have been able to create a truly agentic system automating a highly complex flow with few steps within few days but several hours. Hope to launch it as a product with AWS support.

Learning Objectives

How to use AWS Bedrock Agent Core
Strand Agents etc hosted on Lambda
Kiro development environment

Future Vision

Launch as a product with blessings from AWS and help the community as a whole...Bring in several more features to make the lives of people easier like:

Document Processing: Scanning capabilities for medical documents
Payment Tracking: Payment disbursal tracking systems
Analytics: Historical claim analysis and insights
Enhanced Automation: Further streamlining of complex healthcare workflows

Building the future of healthcare claims processing through intelligent automation