How Natera designed an agentic AI system for lab ops with Amazon Bedrock AgentCore
What happened
Natera, a company specializing in cell-free DNA testing, developed an AI system called SWARM to manage lab instrument maintenance. The engineering team built SWARM using Amazon Bedrock AgentCore, integrating it with their existing tools to automate support ticket responses. SWARM processes support tickets for lab equipment, delivering diagnostic reports and enabling engineers to access relevant knowledge quickly. The system was implemented to address delays and knowledge loss in lab operations.
Key details
- SWARM achieves 130 times faster first response times compared to previous processes.
- The system automates 96.2% of support ticket responses for lab instrument maintenance.
- Natera handles over 65 support tickets per day across three Jira queues.
- Amazon Bedrock AgentCore provides the managed runtime and security features for SWARM.
- SWARM integrates with Jira, Confluence, Amazon DynamoDB, Amazon SQS, and AWS X-Ray.
Why it matters
Lab engineers at Natera now receive immediate diagnostic support for equipment issues, reducing instrument downtime and delays in sample processing. The automation preserves institutional knowledge, even when experienced technicians leave, and allows the team to handle a high volume of support tickets without hiring additional staff. This results in faster lab operations and more consistent patient reporting.