Skip to content

OpenAI

Sophos lets Daybreak agents close 52% of MDR cases

Written by an AI model from the linked sources and checked against them by a second model. How we work

All about: OpenAI models and products →
Image: openai.com

Sophos now runs agents built on OpenAI’s Daybreak that resolve 52% of its MDR cases end to end, while human analysts keep control over consequential actions, OpenAI said in an October 9 customer story. If your team is weighing Daybreak for security operations, treat these as figures the vendor published: OpenAI’s own number is a 96% cut in investigation time, and the before-and-after minutes appear only in later reports on the case.

Those reports put Sophos’s average security event response time at about 38 minutes before Daybreak and 89 seconds after, which matches the 96% figure. The labels differ, though. OpenAI calls the metric investigation time, and the reports call it response time.

MDR, or managed detection and response, is the service in which a vendor watches and investigates security alerts for its customers. That work is one of the uses OpenAI lists on its Daybreak page, under defensive operations: investigating threats, improving detections and supporting response. The same page has people review consequential changes and independently verify deployed fixes, which is the line Sophos draws between agents and analysts.

Since August, Daybreak has come in two tiers. TechCrunch reported that the Blue tier covers incident response, malware analysis and patch validation, and that OpenAI calls Blue its “recommended starting point for most defenders”. Red adds models trained specifically for cybersecurity.

Outside the commercial deals, the same program now covers civilian cyber defense in Ukraine, and OpenAI has committed $1 billion in subsidized access for state and local governments, critical-infrastructure operators, community banks, nonprofits and open-source maintainers.

No comments yet. Be the first.

Join the conversation

Sign in to leave a comment. It appears after moderation.