Case study · Healthcare
Voice AI for a healthcare operation
Agents that handle hundreds of thousands of patient calls a month, on GCP.
Delivery work from our track record. We've left the client unnamed; the numbers are real.
60%
faster support resolution
300%
average ROI
35%
more appointments kept
99.5%
uptime
The challenge
A healthcare provider was drowning in inbound calls - scheduling, follow-ups, records - and couldn't hire their way out of it. They needed software to take the routine load without dropping the ball on HIPAA.
What we did
- Designed a multi-agent system with 30+ tools for scheduling, record retrieval and treatment workflows.
- Built a voice assistant that transcribes in real time and routes the hard calls to people.
- Ran 15+ AI services on Kubernetes with autoscaling, holding 99.5% uptime.
- Wired call data into BigQuery and dashboards so the practice could see what was happening.
Stack
GCPVertex AILangGraphGoogle ADKGKEBigQuery