Microsoft has moved Agent 365, its AI agent management platform, from preview to general availability, reflecting the urgent need for governance of autonomous AI in enterprises. Agent 365 acts as a unified control portal allowing IT and security teams to monitor and regulate AI agents across Microsoft, third-party clouds like AWS Bedrock and Google Cloud, employee devices, and partner SaaS agents. A critical focus is on managing ‘shadow AI’—AI agents installed by employees without IT approval, presenting fresh security challenges. Microsoft identifies key risks including accidental exposure of sensitive data, malicious prompt injections, and data loss due to non-agent-aware security systems. The platform provides comprehensive governance controls, asset relationship mapping, and runtime blocking, integrated with Microsoft Defender and Intune, to protect enterprise data and operations. Additionally, Agent 365 supports cross-cloud agent lifecycle management and includes sandboxed Cloud PC environments for secure agent use. A broad partner ecosystem further enhances its reach, positioning Agent 365 as a crucial tool for managing the evolving AI agent landscape in enterprise settings.
Back
