Microsoft Unveils Agent 365 for AI Agent Management

Microsoft (MSFT) is taking steps to manage the proliferation of AI agents within corporate environments. Following the trend of empowering employees to create AI agents for task automation, the tech giant has unveiled Agent 365, a tool designed for IT specialists to monitor and govern these agents effectively.

The announcement, made at Microsoft’s Ignite conference in San Francisco, introduces Agent 365 as a comprehensive solution providing visibility into all AI agents operating within a company’s systems, regardless of their origin. This includes agents sourced from third-party vendors, offering a centralized management platform for an increasingly decentralized AI landscape.

“Just as you provision identities and access controls for new employees, you’ll now provision the same for your AI agents,” stated Judson Althoff, CEO of Microsoft’s commercial business, highlighting the shift towards treating AI agents as integral components of the corporate workforce. This reflects a growing awareness of the need for structured governance in the deployment of AI within enterprises.

Agent 365 offers administrators a range of capabilities, including the ability to approve new agents, track agent popularity, and quantify time savings achieved through automation. Crucially, the tool also enables the identification and mitigation of security risks, allowing administrators to block potentially malicious or unauthorized agents. End-users benefit from the ability to analyze agent activities, promoting transparency and accountability.

The rise of AI agents, fueled by the success of OpenAI’s ChatGPT, has seen a surge in the development of software capable of automating tasks ranging from software development to ad generation. This has led to increased adoption by traditional corporations seeking to improve efficiency and productivity, but also introduces the need for robust management and security protocols to maintain control over these autonomous systems.

Agent 365 will automatically discover and display agents from vendors like Adobe (ADBE), Cognition, Databricks, Glean, ServiceNow (NOW), and Workday (WDAY). The platform also supports agents built on Microsoft’s Azure AI Foundry and Copilot Studio. According to Ray Smith, vice president of autonomous agents at Microsoft, the company welcomes integration with other AI agent providers, positioning Agent 365 as a central hub for AI agent management.

EY, a leading accounting firm, provides an early example of Agent 365’s effectiveness. Having previously developed its internal catalog of AI agents, the firm is now leveraging Agent 365 to enhance control and governance, as confirmed by Mark Luquire, a managing director at the company. This underscores the value proposition for larger organizations grappling with complex AI deployments.

Cybersecurity companies are similarly attuned to the challenges presented by the growing number of AI agents and the potential for security vulnerabilities. Okta (OKTA), an identity management provider, announced in September that it would also provide tools to discover and manage AI agents, highlighting the broader industry focus on this emerging security concern. The move by Okta signals a growing understanding of the need for comprehensive security fabrics that extend to AI-driven operations.

Microsoft aims to initially rollout Agent 365 to customers participating in its Frontier program, marking the first stage in early access for AI features. Pricing details are yet to be finalized, but the tool positions itself as an essential solution for companies looking to safely harness the cost-saving and efficiency-driving benefits of AI.