Netskope has launched AgentSkope, a library of AI agents designed to automate tasks across security operations and network operations, as organisations grapple with increasing alert volumes and operational workload.
In an email to media, the company cited research claiming 40% of security alerts remain uninvestigated due to lack of capacity. Netskope said AgentSkope is intended to help reduce manual workflows in areas including policy creation, triage, investigation and troubleshooting.
According to Netskope, AgentSkope is launching with six agents, with additional agents planned. The initial set targets data loss prevention (DLP) analysis, insider threat triage, private access configuration auditing, and digital experience management (DEM) troubleshooting and insights, as well as natural language querying of cloud and SaaS application risk attributes and compliance certifications.
Netskope said a global professional services organisation participating in beta tests used one of the agents to triage and investigate 14 million daily security alerts in minutes, with the vendor claiming the automation reduced hours of manual work and allowed staff to focus on remediation.
“Security and network operations teams today are overwhelmed by an endless loop of manual triage, and bogged down by repetitive tasks across disparate tools, leading to severe analyst burnout, an inability to innovate at speed, and unchecked risk,” said Sanjay Beri, co-founder and CEO of Netskope.
Stuart Walters, partner and chief information officer at BDO UK, said the firm expects automation that supports security and networking workflows to become more important as AI adoption increases and data movement grows.
IDC research manager Pete Finalle said security and network teams face mounting workloads as AI expands the threat landscape, and argued organisations should consider “agentic security automation” to augment human resources.
Netskope said AgentSkope and five of the six agents are generally available, with the Insider Threat AISecOps Agent in private preview.
