Microsoft Unveils Agent 365: Centralize and Manage AI Agents in Your Business!

At the Microsoft Ignite conference, held from November 18 to 21, 2025, in San Francisco, the Redmond-based company unveiled Microsoft Agent 365, a platform aimed at easing the deployment and management of a fleet of AI agents, whether they are built using its own tools, open-source frameworks, or developed by partners.

A “Unified System” to Centralize and Regulate AI Agents in Business

Microsoft’s goal with the introduction of Agent 365 is to create an infrastructure to “govern agents responsibly and on a large scale.” Launched on Tuesday, November 18, 2025, and available to members of the Frontier program (a scheme granting access to the latest AI innovations from the company), Microsoft Agent 365 is intended to provide essential tools for centralizing and controlling the use of AI agents in businesses, as their numbers are expected to skyrocket in the coming years. “The research firm IDC estimates that there will be 1.3 billion agents by 2028,” the American company notes.

Within this “unified system,” which includes Microsoft’s key security (Entra, Defender, Purview) and productivity solutions (Word, Excel, Outlook, etc.), it will be possible to utilize not only agents created in Copilot Studio and Microsoft Foundry but also open-source agents published on GitHub by Anthropic or OpenAI, or those designed by companies like Adobe, NVIDIA, or Genspark. “Our partners are already rallying to the project, bringing their innovation and momentum to the expanding ecosystem of Agent 365,” Microsoft proudly states.

What are the key features of Microsoft Agent 365?

On Agent 365, administrators will have a 360-degree view of the agents’ usage within the organization and will be able to manage them as if they were actual employees, Microsoft explains. Each agent will be associated with an Entra identifier and integrated into a registry. The aim? To enable IT teams to map out the agents in circulation, to know who is using them, and, most importantly, to quarantine the “ghost agents,” i.e., those that have not been previously authorized.

Additionally, it will be feasible to set tailored access for each agent according to their scope, ensuring they only reach the necessary resources. “IT teams can set safeguards on who creates, integrates, and administers the agents,” adds Microsoft.

Lastly, as illustrated in the featured image, a dashboard will provide precise measurements of the agents’ effectiveness, assess their adoption within the organization, and identify the types of tasks they handle. “End-users also gain visibility on task compliance and the business impact of the agents they oversee,” the company further elaborates.