Share via


Secure AI agents at scale using Microsoft Agent 365

As organizations adopt AI agents to automate workflows and boost productivity, securing these agents has become a critical concern. Unlike traditional applications, AI agents operate autonomously, interact with sensitive data, and execute tasks across multiple systems - making them high-value targets for intentional attacks and also vulnerable to unintentional compromise caused by misconfigurations or excessive permissions.

Microsoft Agent 365 provides a unified control plane that lets you oversee the security of all AI agents in your organization. It integrates with Microsoft’s security suite - now extended to secure AI agents - to secure agents built in Microsoft Copilot Studio, Microsoft Foundry (formerly called Azure AI Foundry), and third-party solutions.

This article outlines the core security capabilities that Microsoft Agent 365 provides based on Microsoft’s extended security infrastructure for AI agents.

For more information about Microsoft Agent 365, see Microsoft Agent 365 documentation.

A unified control plane for managing agent security

Agent 365 integrates with Microsoft 365 Admin Center, giving IT teams a familiar interface to configure policies, apply Conditional Access, and monitor compliance across the agent fleet.

This control plane provides centralized visibility and lets you drill down into Microsoft's suite of security tools - the same tools you use to manage, secure, and govern users - to manage posture, configure policies, investigate issues, and remediate risks for agents.

A screenshot showing the Agent 365 overview page in Microsoft 365 Admin Center.

Microsoft security infrastructure extended to AI agents

Agent 365 extends your existing security and governance practices to AI agents, so teams can use familiar tools and processes without disruption.

By integrating with Microsoft’s security suite - enhanced with agent-specific controls and capabilities - Agent 365 lets you manage AI agent security across key areas. Drill down into each security product from the Agent 365 control plane to act on agent-specific insights and recommendations.

The following sections outline the core security capabilities that Microsoft Agent 365 provides by integrating with Microsoft’s security suite.

Identity management

Microsoft 365 admin center | Microsoft Entra Registry

Get the complete view of all of the agents used in your organization, including agents with agent ID, agents you register yourself, and shadow agents.

A screenshot showing the agent identities tab in Microsoft 365 Admin Center.

Learn more about identity management in Microsoft Agent 365:

Access control

Microsoft Entra lifecycle workflows | Microsoft Entra ID governance | Microsoft Entra Conditional Access and Identity Protection

Limit agents' access to resources they need only and prevent agent compromise with risk-based conditional access policies.

  • Define guardrails - Set policies that define who can create, onboard, and manage agents, assign agent sponsors to keep them under supervision, and ensure that agents start secure with security policy templates.

  • Governance access - Enforce least privilege access by only giving agents access rights to the apps and resources they need to complete tasks.

  • Set conditional access - Enforce real-time intelligent access decisions based on agent context and risk, conditions you define and resource they're trying to access.

A screenshot showing the Conditional Access page in Microsoft 365 Admin Center.

Learn more about access control in Microsoft Agent 365:

Security posture

Microsoft Defender | Microsoft Security Exposure Management

Understand your agent and data security posture and attack paths that attackers can create from agents to other critical assets. Remediate misconfigurations, exposures, and vulnerabilities in agents.

A screenshot showing the Overview tab on the Agent page in Microsoft Defender.

Learn more about managing agent security posture using Microsoft Agent 365:

Detection and response

Microsoft Defender

Detect known and emerging threats targeting agents and enable a rapid response with a complete view of the cyberattack chain and prioritized investigation and response at the incident level.

A screenshot showing the Incident page in Microsoft Defender.

Learn more about detecting and responding to threats using Microsoft Agent 365:

Runtime defense

Microsoft Defender | Microsoft Entra Secure Access Service Edge | Microsoft Purview Insider Risk Management

Use AI-powered intelligence to block prompt injection attacks, malicious traffic, and prevent data exfiltration due to risky agent behavior in real time.

A screenshot showing the Security and Compliance tab of the Agent page in Microsoft 365 Admin Center.

Learn more about runtime defense for agents in Microsoft Agent 365:

Data security

Microsoft Purview Data Loss Prevention | Microsoft Purview Information Protection | Microsoft Purview Data Security Posture Management

Gain visibility into AI-related data exposure risks and dynamically block agent interactions with sensitive data based on data security labels and policies.

A screenshot showing the Agent page in Microsoft Purview.

Learn more about protecting data from agents in Microsoft Agent 365:

Next step

Learn more about Microsoft Agent 365: