Healthcare / Clinical laboratory testing
Natera: Automatically drafting cause hypotheses and troubleshooting steps for clinical-laboratory equipment tickets
- Company
- Natera
- Country
- United States
- Adoption stage
- In operation
- Source published
- Date basis
- The date the source was published. It can differ from the date adoption started.
- How the source was checked
- Read the full source text
The work problem
Natera's maintenance knowledge depended on senior technicians and was lost when they left. The company described more than 65 daily Jira tickets and a 17.4-hour average first response, delaying samples and patient reporting.
Technology and data
Natera operates SWARM on Bedrock AgentCore. It retrieves vendor manuals and resolved tickets to draft cause hypotheses and prioritized checks. Engineers perform verification and repairs and record outcomes for evaluation. Only equipment and workflow information is used; patient identifiers are excluded from the model and vector index.
Results
Natera and AWS reported an approximately eight-minute first response and automatic reports before engineer review for 96.2% of tickets over 90 production days. The evaluation covered 524 tickets; 269 had both a confidence label and engineer-verified outcome. Coverage was not treated as repair completion.
Limits and open questions
Faster first response is not faster actual repair. Low-confidence cases numbered 37, so labels are not interpreted as probabilities of correctness.
Sources
- How Natera designed an agentic AI system for lab ops with Amazon Bedrock AgentCoreAWS / Natera, Accessed
Compiled from public sources. These are not results from ATF Works customers.