Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
An AI system includes not only the technology, but also the people who use it, the people affected by it, and the environment in which it's deployed. Microsoft's Responsible AI FAQs are intended to help you understand how AI technology works, the choices system owners and users can make that influence system performance and behavior, and the importance of thinking about the whole system, including the technology, the people, and the environment. You can use Responsible AI FAQs to better understand specific AI systems and features that Microsoft develops.
Responsible AI FAQs are part of a broader effort to put Microsoft's AI principles into practice. To find out more, see Microsoft AI principles.
What is Microsoft Agent 365?
Agent 365 is Microsoft’s strategic initiative to enable, manage, and scale AI agents across the enterprise.
What can Agent 365 do?
By extending the infrastructure that you use for managing people to agents, Agent 365 equips your agents with the same apps and protections, tailored to agent needs, saving IT time and effort on integrating agents into business processes. Agent 365 includes leading Microsoft’s security and productivity solutions: Defender, Entra, and Purview to protect and govern agents; Microsoft 365 apps and Work IQ to accelerate their productivity; and Microsoft 365 admin center to manage agents.
Registry: Get the complete view of all agents in your organization, including agents with agent ID, agents you register yourself, and shadow agents.
Access control: Bring agents under management and limit their access to only resources they need. Prevent agents from compromise with risk-based conditional access policies.
Visualization: Explore connections between agents, people, and data and monitor agent behavior and performance in real time to assess their impact on your organization.
Interop: Equip any agent with apps and data to simplify human-agent workflows. Connect them to Work IQ to provide context of work to onboard into business processes.
Security: Protect agents from threats and vulnerabilities, detect, investigate, and remediate attacks that target agents. Protect data that agents create and use from oversharing, leaks, and risky agent behavior.
What are Agent 365’s intended uses?
Enterprise employees can utilize AI agents as autonomous entities capable of acting, collaborating, and delivering outcomes across enterprise workflows.
CTOS and IT admins who are accountable for enabling their organizations to become frontier and need to do it securely, keeping agents under control.
How was Agent 365 evaluated? What metrics are used to measure performance?
We evaluated XPIA and ungrounded risks using a variety of assessment activities, including AI red teaming and evals, where defect rates were identified (if any) and rationalized.
What are the limitations of Agent 365? How can users minimize the impact of Agent 365’s limitations when using the system?
Human in the loop (HITL), sensitivity labels, and sharing permissions are known limitations that will be addressed in the future. Awareness to this is noted on various touchpoints: When you add this agent, any files or chats you share can be summarized and included in the agent’s responses to any user, even those without permissions to the files or chats and regardless of sensitivity label.
What operational factors and settings allow for effective and responsible use of Agent 365?
- To manage external collaborations settings for agents in Microsoft Teams, admins must configure tenant policies in Teams admin center.
How do I provide feedback on Agent 365?
Any feedback for Agent 365 should be provided on the surface in which it is experienced, including but not limited to Microsoft 365 admin center, Teams admin center, Teams Store, Agent Store, Copilot, Purview, and Defender.