AWS Machine Learning describes Natera’s SWARM agentic AI system for lab operations

AWS Machine Learning describes how Natera and AWS built SWARM, a lab-operations agentic AI system based on Amazon Bedrock AgentCore. Designed to capture institutional knowledge and work with Jira and Confluence, it automatically generates diagnostic responses for support tickets and runs with built-in security and observability. In a 90-day production evaluation, Natera says SWARM reduced first response time from 17.4 hours to about 8 minutes and delivered automated first responses for 96.2% of tickets.