What happened
On Oct 9, OpenAI published that Sophos now uses OpenAI Daybreak agents to cut cyber-threat investigation time by 96% and automatically resolve 52% of MDR cases while keeping human oversight for the rest — the first disclosed large-scale production deployment of agentic AI inside a major MDR operation.
Why it matters
Concrete, named-vendor evidence that agentic SOC tooling is moving from pilot to production ROI at managed-detection scale: if ~half of MDR cases resolve without human investigation, it resets analyst-load economics for the whole SOC-as-a-service market and validates the 'autonomous defense' procurement thesis.
Applicability
MDR providers and large SOC teams evaluating agentic investigation platforms (OpenAI Daybreak vs. vendor-native agents) should monitor Sophos-deployment metrics and the human-oversight model.