Extended Conditional Access protections for Agent's user accounts

Conditional Access now provides broader controls to secure AI agents that have a user account. Administrators can:

• Target agent's user accounts with greater precision by including or excluding individual agents or dynamically grouping agents using Custom Security Attributes.
• Protect against risky agent activity by applying Conditional Access policies based on Agent Risk.
• Require compliant devices for agents running on managed endpoints, including Windows 365 for Agents, ensuring agents can only operate from devices that meet your organization's compliance requirements.
• Apply device platforms, filter for devices and compliant network conditions to agents running on endpoints, enabling policies based on device state and trusted network locations.

These capabilities extend Zero Trust protections to agent's user accounts while leveraging the familiar Conditional Access policy experience. For more information see: Target agent identities in Conditional Access policies.