Your security team doesn’t know about half its users (www.techradar.com)

A new report highlights a critical challenge for security teams as enterprises increasingly integrate task-specific AI agents into their workflows, expected to rise from under 5% to 40% by 2026. Traditional security models are ill-equipped to address the rapid execution of tasks by these agents, which operate at machine speed without the human judgment required to identify suspicious activity. This has created a significant visibility gap in security frameworks that typically rely on recognizing user behavior and intentionality, leading to heightened risks of data breaches and unauthorized actions. To mitigate these risks, security leaders must shift their focus to embedding controls within browser sessions, where over 85% of enterprise workflows now take place. Unlike traditional security measures that assume static checks and manual oversight, real-time governance at the session level is necessary to monitor and manage both human and agent actions effectively. With agents operating autonomously, security teams need to adopt distinct identity management practices and session-level visibility to close the existing gaps. This proactive approach ensures that security can keep pace with the rapid evolution of AI-driven workflows, safeguarding enterprises while enhancing productivity.