EirAI is an AI-powered assistant that streamlines clinical documentation and supports medical coding workflows. Designed for healthcare professionals and medical coders, EirAI transcribes patient-doctor conversations into standardized SOAP notes, extracts ICD-10 codes, and predicts insurance approval rates using historical data.
- 🥈 2nd Place – Themed Track @ HackKU 2025
-
Speech-to-Text Transcription
Converts live or recorded doctor-patient conversations into structured SOAP notes (Subjective, Objective, Assessment, Plan). -
ICD-10 Code Extraction
Automatically identifies and indexes relevant ICD-10 codes from SOAP notes, covering all ~73,000 classification codes. -
Approval Rate Prediction
Cross-references ICD codes with a custom-built dataset of 1.5M+ simulated historical records to estimate likelihood of insurance approval per provider. -
Scalable & Modular Design
Built with extensibility in mind, enabling future integration with EHR systems, payer APIs, and provider workflows.
- Frontend: NextJS, Framer-motion
- Backend: PostgreSQL (supabase)
- AI Models: Open AI Speech Transcription (speech-to-text), Google Gemini for note structuring
- Data: Custom mock dataset simulating 1.5M+ historical claims and all ICD-10 codes
- ~73,000 ICD-10 codes indexed
- 1.5 million+ rows of simulated insurance claim data
- Provider-specific historical approval trends