Natera’s intelligent appointment scheduling with Amazon Bedrock AgentCore
Natera deployed a voice agent handling real appointment bookings with 100% tool accuracy and sub-7-second latency. Here's how they solved the dual-WebSocket and latency-masking problems.
Why it matters
Production agent case study with engineering depth: real deployment metrics, architectural patterns (event-driven latency masking, progressive-trust auth), and tool-calling reliability at scale. Practitioners building voice agents can extract specific patterns; enthusiasts see proof that agent infrastructure is hardening.
The key facts
14 to knowNatera deployed voice agent for mobile phlebotomy appointment scheduling
100% tool-calling accuracy achieved
Sub-7-second latency (voice-agent responsiveness benchmark)
Built on Amazon Bedrock AgentCore
Dual-WebSocket bridge architecture
Event-driven latency masking pattern
Progressive-trust authentication implemented
Production deployment with patient-facing real use case
Natera built voice agent on Amazon Bedrock AgentCore
Sub-7-second latency for appointment booking
Dual-WebSocket bridge architecture for latency masking
Event-driven progressive-trust authentication
Mobile phlebotomy appointment scheduling use case
Production deployment handling real patient conversations
Go to the source
AWS Machine Learning Blogaws.amazon.com
Publisher excerpt: Learn how Natera built an automated voice agent on Amazon Bedrock AgentCore that lets patients book mobile phlebotomy appointments through natural conversation. The post covers the dual-WebSocket bridge, event-driven latency masking, and progressive-trust authentication behind 100% tool-calling…