Inspiration

Medicare is designed to make healthcare accessible, especially for seniors, yet medical billing errors remain widespread. While many disputed bills are successfully corrected or negotiated, only a small percentage of patients challenge them due to the time, complexity, and uncertainty involved. That's why we built Gem, an AI-powered medical billing advocate that identifies potential errors, simplifies disputes, and handles the tedious work of negotiating bills. Gem empowers patients to challenge unfair charges while keeping them in control, allowing them to focus on what matters most: their health.

What it does

Gem acts as an AI-powered medical billing advocate, allowing patients to upload their Explanation of Benefits (EOB), Medicare Summary Notice (MSN), or provider bill for a comprehensive audit. The app extracts individual charges, compares them against official Medicare reimbursement rates, flags potential errors like duplicate charges, overcharges, and incorrect billing codes, and prioritizes discrepancies by financial impact.

How we built it

Step 1: Gemini API Integration We integrated the Google Gemini API to power Gem's multi-agent medical billing audit workflow. Step 2: Document Extraction Built an AI agent to extract and structure billing data from MSNs, EOBs, and provider bills. Step 3: Discrepancy Detection Developed an engine that compares extracted charges against official CMS reimbursement data to flag potential billing errors. Step 4: Prioritization Created a Strategist agent to rank discrepancies by financial impact and confidence. Step 5: Dispute Generation Built a Composer and Negotiator agent to draft evidence-backed dispute letters using Gemini. Step 6: Patient Approval Added a review step so patients can verify findings and approve communications before sending. The result: Gem combines Gemini's AI capabilities with public CMS data to make medical billing audits simpler and more actionable.

Challenges we ran into

One challenge was making the negotiation caller sound natural and respond at a comfortable pace, without abrupt replies or long pauses. It also had to make useful arguments while keeping the conversation flowing

Accomplishments that we're proud of

We're proud of how we leveraged the Google Gemini API to turn a complicated, time-consuming medical billing dispute process into a simple, AI-powered experience. By combining intelligent automation with reliable Medicare data, we built Gem to empower patients, reduce financial stress, and make healthcare billing more transparent and accessible.

What we learned

We started off our project thinking the hardest part was building an AI that could do the task. We finished realizing the harder problem was building one that patients could trust to act on their behalf. That influenced every design decision, from grounding recommendations in public CMS data to requiring user approval before any communication is sent.

Built With

Share this project:

Updates

Submission history