Microsoft expands Project Perception to supervise agentic insiders
Event summary
Microsoft published official posts stating that Project Perception is intended to help organizations defend with agents and can supervise agentic activity in real time while providing traceability for practitioners to examine what happened and why risk emerged. The posts present a product-focused approach for treating agentic systems as operational/insider risks and for giving practitioners evidence to update security controls.
CHRONOS Wire · October 20 · Alert 12
Publication details
- Published
- Updated
- Revision
- r1
- Source
- Microsoft Security Blog
Cliff Notes
- Microsoft published posts expanding Project Perception to supervise agents and provide traceability; excerpts do not support specific feature names or detailed mechanics.
U.S. Impact
U.S. Impact assessment pending.
ELI5: Plain-English Explanation
Microsoft published posts describing upgrades to Project Perception so organizations can better detect and supervise software agents running inside their environments and have traceability to examine why agent risks emerged.
Why Urgent Level 3
Enterprises running or evaluating AI agents should review vendor guidance and consider short-term assessments of agent discovery and supervision capabilities.
What Changed
Microsoft published official posts describing an expansion of Project Perception focused on defending with and supervising agentic systems, including real-time supervision and traceability for practitioners to examine what happened and why risk emerged.
What Is Genuinely New
Official Microsoft posts describe expanding Project Perception to help organizations defend with agents and to supervise agentic activity in real time with traceability for practitioners.
CHRONOS Bottom Line
Microsoft has published product-focused guidance and capabilities for treating AI agents as operational/insider risks and to provide supervisory controls and traceability for agent activity.
Uncertainties / Known Unknowns
Microsoft posts describe capabilities and architectural approach but do not provide detailed rollout timelines, pricing, or integration specifics in the quoted excerpts.
Detailed Analysis
Microsoft's official posts present Project Perception as a product/architectural approach to defend with agents and to supervise agentic activity in real time with traceability for practitioners. The available excerpts support that Microsoft is positioning Project Perception to help enterprises detect and investigate agent-originated risks, but the excerpts do not specify feature names, continuous scoring mechanics, or precise product release/availability details.
Sources / Evidence
- 01Agents are your newest insiders. Security must move at their speed.Microsoft Security Blog2026-10-19
- 02Securing the agentic insiderThe Official Microsoft Blog2026-10-19
Follow CHRONOS ALERTS
Published alerts also appear on Instagram and Threads as @chronosalerts.