Medical Triage Voice AI Agent

Designed and deployed a voice-based medical triage AI agent that answers patient phone calls, collects intake information, verifies insurance, and schedules appointments through natural, real-time conversations.

The AI agent handles inbound patient calls end-to-end, reducing front desk load while routing structured, verified information to clinical and administrative staff when human follow up is required.

Automated phone-based patient intake and appointment scheduling using a voice AI agent, reducing reliance on staff-led call handling.

Enabled seamless patient interactions via SIP telephony integration (Twilio).

Implemented a HIPAA-compliant system architecture to securely process protected health information.

Delivered real-time insurance verification, improving scheduling accuracy and reducing manual processing.

Automated structured email handoffs to staff, streamlining back-office workflows.

Integrating multiple AI services (OpenAI LLM, Deepgram STT, Cartesia TTS) into a real-time agentic voice framework using LiveKit Agents.

Designing low latency conversational flows that balance data accuracy with natural patient interaction.

Building reliable SIP telephony infrastructure with error handling and call-quality safeguards.

Implementing real-time insurance and scheduling integrations without blocking voice interactions.

Designing and validating a HIPAA-compliant voice pipeline for live PHI handling.

Healthcare phone-based intake and scheduling remains manual, fragmented, and resource intensive, creating long hold times, staff burnout, and inconsistent patient experiences.

AI & Voice: OpenAI LLM, Deepgram STT, Cartesia TTS, agentic AI, conversational NLP

Telephony & Real-Time: LiveKit Agents, Twilio, SIP, real-time audio pipelines

Backend & Systems: Python, API development, system integration

Compliance & Security: HIPAA-compliant system design, secure PHI handling

Automation: Patient intake, insurance verification, appointment scheduling, workflow handoff