Natera's SWARM AI System Cuts Response Times and Boosts Efficiency
Discover how Natera's groundbreaking SWARM AI system, built with Amazon Bedrock AgentCore, is transforming lab operations by automating support and preserving vital institutional knowledge for improved patient outcomes.
Key Facts
- SWARM reduced first response time from 17.4 hours to 8 minutes, enhancing operational efficiency.
- 96.2% automated coverage of support tickets reveals significant labor cost savings potential.
- Loss of senior technician knowledge highlighted vulnerability; SWARM mitigates this risk effectively.
- Improved diagnostic accuracy from 71% to 75% shows data quality's critical role in financial outcomes.
- Integration with AWS tools indicates strategic alignment with cloud services for scalable AI solutions.
Summary
Summary
Natera, a leader in cell-free DNA testing, faced significant challenges with lab instrument downtime due to inefficient knowledge management among their engineering team. To address this, they developed SWARM, an agentic AI system using Amazon Bedrock AgentCore, which improved first response times to support tickets from 17.4 hours to approximately 8 minutes and achieved 96.2% automated coverage for ticket responses.
Background
Natera operates in the healthcare industry, specializing in cell-free DNA testing. Prior to the deployment of SWARM, the company struggled with maintaining lab instruments due to a lack of standardized knowledge management. The engineering team relied heavily on individual expertise, which led to the loss of critical knowledge when senior technicians left the company. With over 65 support tickets generated daily, the existing system was unable to keep pace with the demand for timely diagnostics.
Challenge
The primary challenge Natera aimed to solve was the continuity of knowledge within their lab operations. The absence of a structured system to capture and deploy institutional knowledge resulted in prolonged instrument downtime and delays in patient reporting. The engineering team needed a solution that could provide rapid, accurate diagnostics across various equipment types without relying solely on human intervention.
Solution
Natera collaborated with AWS to build SWARM, an AI system that utilizes Amazon Bedrock AgentCore. SWARM consists of two main components: the swarm-agent, which autonomously generates diagnostic reports before an engineer opens a ticket, and swarm-hive, which allows engineers to interact in real time. The system integrates deeply with existing tools like Jira and Confluence, enabling it to process high volumes of support tickets autonomously. SWARM employs a structured knowledge base that includes vendor manuals and historical data to enhance diagnostic accuracy.
Results
The implementation of SWARM resulted in a dramatic reduction in first response times, dropping from 17.4 hours to roughly 8 minutes—an improvement of about 130 times. Additionally, SWARM achieved 96.2% automated coverage for support tickets, meaning that the majority of tickets received a diagnostic report generated by the system before any engineer intervention.
Key Insights
Key lessons from Natera's deployment of SWARM include the importance of treating knowledge management as a structured process, the necessity of building traceability into AI systems, and the value of separating batch processing from interactive querying to optimize performance. These insights can guide other organizations looking to implement similar AI solutions in operational workflows.
Customer Testimonial
“We needed a managed runtime that would let us focus engineering effort on diagnostic accuracy rather than infrastructure plumbing—AgentCore gave us that, along with built-in identity management and the flexibility to swap models without rewriting agent logic.” — Ariel Jirau, Lab Engineering Team Lead at Natera.
Entities Mentioned
Companies
Products
Technologies
People
Key Concepts
Definitions
- agentic AI
- A type of artificial intelligence system that autonomously performs tasks and makes decisions based on institutional knowledge.
- SWARM
- Systematic Workflow Analytics and Resolutions Manager, an AI system developed by Natera for lab operations.
- RAG
- Retrieval Augmented Generation, a method for converting documents into embeddings for improved search and retrieval.
- HIPAA
- Health Insurance Portability and Accountability Act, a U.S. law designed to provide privacy standards to protect patients' medical records.
- observability
- The ability to measure and monitor the internal states of a system based on the data it generates.
Use Cases
- →automated ticket response
- →diagnostic report generation
- →real-time engineer collaboration
- →knowledge base construction
- →clinical compliance management
- →data governance in healthcare
Frequently Asked Questions
What is the primary function of SWARM?
SWARM is designed to automate the diagnostic process for lab operations, providing engineers with immediate access to institutional knowledge and generating structured reports for support tickets.
How does SWARM ensure compliance with HIPAA?
SWARM operates on HIPAA-eligible AWS services under a Business Associate Agreement, ensuring that no patient-identifiable information is processed, and all agent actions are audit-logged.
What are the benefits of using Amazon Bedrock AgentCore?
Amazon Bedrock AgentCore provides a managed runtime for building and deploying AI agents, allowing teams to focus on diagnostic accuracy without the burden of maintaining custom infrastructure.
How does SWARM improve diagnostic accuracy?
SWARM continuously evaluates its performance by comparing its diagnostic outputs against engineer-verified resolutions, allowing for iterative improvements based on data coverage and accuracy.
What architectural patterns can other organizations adopt from SWARM?
Organizations can adopt patterns such as separating batch processing from interactive querying, building traceability into skill libraries, and using continuous evaluation methodologies to enhance their own AI systems.