Sophos says AI agents cut some response times to 89 seconds, resolving 52% of MDR cases
Sophos says AI agents cut some response times to 89 seconds, resolving 52% of MDR cases
AI Cybersecurity

Sophos says AI agents cut some response times to 89 seconds, resolving 52% of MDR cases

OpenAIOct 9, 2026

Sophos says agents built with OpenAI Daybreak reduced average response time to 89 seconds for cases using them and resolved 52% of MDR cases end-to-end.

Sophos says AI agents developed through OpenAI’s Daybreak program have shortened its handling of some managed detection and response (MDR) cases. In a case study published October 9, 2026, OpenAI reported Sophos figures showing average response time for cases using the agents fell from about 38 minutes to 89 seconds, while AI resolved 52% of MDR cases end-to-end.

Sophos built the agents into Sophos Fusion, its cyber defense system. The system draws on sensor data from more than 500 third-party integrations as well as Sophos products. Those sources produce trillions of events each day, which the company narrows to roughly 1,000 to 2,000 cases for investigation across nine security operations centers.

For each case, an investigation agent gathers customer context, detections, indicators of compromise and relevant threat intelligence. A planning model then lays out and carries out investigative steps, before producing a summary and recommended response actions for analysts. Other agents can perform parts of the response process.

The reported automation does not remove customer controls or human review. Sophos offers three service modes: Notify, where it recommends action for the customer to take; Collaborate, where Sophos and the customer decide together; and Authorise, where Sophos can act on the customer’s behalf. The company says potentially destructive responses remain subject to the appropriate oversight, and cases agents are not suited to handle go to human judgment.

Sophos CTO John Peterson said the earlier process averaged around 38 minutes and that the 89-second figure applies to cases handled using agents. OpenAI’s account says Sophos intends to expand the agents’ capabilities and range of use cases. The results are company-reported and describe Sophos’s deployment; they do not establish equivalent performance for other security operations teams.

Industry Impact

Faster MDR case handling

Cybersecurity and managed detection and response

Overview

Sophos’s reported results show how agent-assisted investigation can reduce response time in a high-volume MDR operation while automating a meaningful share of case resolution. For cybersecurity teams, the deployment offers a model for combining broad telemetry, investigation plans and analyst recommendations without removing customer authorization or human judgment. Near term, providers can benchmark speed and end-to-end resolution against their own workloads. Longer-term expansion depends on safe handling, integration coverage and performance beyond Sophos. These are company-reported results, not proof of industry-wide outcomes.

Key Impact Points

Cases handled using the agents averaged 89 seconds to respond, down from about 38 minutes.

The agents resolved 52% of MDR cases end-to-end, while unsuitable cases remained subject to human judgment.

Three service modes preserve customer control, including oversight for potentially destructive responses.

Opportunities

Agent-assisted MDR workflows

Cybersecurity and managed detection and response

The clearest opportunity is for MDR providers to assess agent-assisted investigation and response workflows against their own case volumes. Sophos’s approach—gathering customer context and threat intelligence, planning investigation steps, and presenting summaries and recommended actions—offers a practical pattern to evaluate. Providers could begin with cases suited to automation, retain analyst review for exceptions and potentially destructive actions, and offer customers defined authorization levels. The article does not establish demand, pricing, or commercial results for other providers, so any revenue opportunity remains unquantified. Expansion should depend on measured response quality, safe escalation and fit with existing security integrations.

  • Evaluate agent-assisted MDR operations

    Problem: High volumes of security events must be narrowed and investigated, while response speed and human oversight remain important.

    Solution: Assess agents that gather case context, plan investigative steps, summarize findings and recommend actions, with human review and customer controls for appropriate cases.

    Target Users: Managed detection and response providers and cybersecurity operations teams.

    Business Value: Potentially shorten case handling and automate suitable cases; Sophos reports 89-second average response times for cases using its agents and end-to-end resolution of 52% of MDR cases.

    Revenue Potential: The article provides no pricing or revenue evidence. Any commercial value for other providers would depend on validated performance and customer demand.

Reference link

https://openai.com/index/sophos

Image Source Link

https://images.ctfassets.net/kftzwdyauwt9/4pg31JOD6Z3cO31LCR6GYq/8c1cedfd2e85c6dd80d2ea8c5ea0c9a1/sophos-social.png?w=1600&h=900&fit=fill

Community Discussion (0)

Loading discussions...