CanadaTechnology and telecom2025 – ongoingeSentire
Autonomous threat investigations in the Atlas managed detection and response platform
eSentire detected 41% more confirmed attacks per customer after Claude agents took on its threat investigations
Problem
eSentire is a managed detection and response provider that protects thousands of customers. Its analysts had to trade depth of investigation against speed. Deep expert analysis of every alert was not possible at that volume.
Approach
eSentire built the Atlas platform on Claude Opus 4.6, run on Amazon Bedrock through the Agent SDK. A multi-agent setup runs threat investigations on its own, using tools to pull evidence. The company checks the results against top security analysts in continuous production evaluation.
Results
Takeaway
eSentire measured its agents against its own best analysts in live production rather than on public benchmarks. It also reported outcomes such as attacks stopped, not only investigation counts.
Categories
- Industry
- Technology and telecom
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- AI agentsGenerative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Anthropic's customer story describes eSentire's Atlas platform, which uses Claude agents to investigate security threats. It reports more than 120,000 autonomous investigations in 12 months, a 6-minute average investigation, 99.96% ransomware containment before encryption, and 41% more confirmed attacks detected per customer. It also reports above 90% alignment with top security experts in production evaluation. The page carries no date; the Internet Archive first saved it on September 19, 2025.
Full details are in the original sources.
