Microsoft's new offering, Agent 365, presents AI agents as “digital coworkers” that organisations will need to register, manage and monitor alongside human staff. Microsoft frames this as an evolution of workplace infrastructure, one where AI agents must fit into identity, access and governance systems originally designed for people.

According to Microsoft, the tool will let companies register their AI agents in the Entra identity registry, define what they can access, connect them with Microsoft 365 applications and monitor their behaviours with telemetry dashboards and alerts.

The Age Of Agent Proliferation

Research cited by Microsoft suggests there could be as many as 1.3 billion AI agents operating globally by 2028. As the number of these agents grows, the question of how to organise, secure and integrate them becomes critical.

Agent 365 is Microsoft’s answer to that problem, a central control plane for AI agents, just like platforms exist today for human workforce identity, device management and access.

The Infrastructure And Ecosystem Angle

The product is built to support not only Microsoft’s own agent investments but also those from Adobe, NVIDIA, ServiceNow, Workday and others.

For businesses, this means the platform could serve as a unified surface for managing diverse agent workflows, from inventory agents in supply chains to out-of-stock agents to customer-service bots, under one governance umbrella.

Governance, Trust And Operational Risk

With agents comes risk: unexpected behaviour, access to sensitive data, linkages between agents and humans that are hard to trace. Agent 365 aims to address these by providing real-time dashboards that map connections between agents, people and data, making oversight visible rather than hidden.

From a regulatory or enterprise-governance standpoint, this matters because it opens the possibility of audit trails for agent actions, clearer responsibility chains and better risk controls in environments where AI agents act autonomously or semi-autonomously.

Enterprise Evolution

Enterprise IT has long evolved from simple device management to identity platforms, to zero-trust access, to cloud-native services. The shift to Agent 365 marks a next phase: treating AI agents not as assistants but as first-class entities within the workforce ecosystem.

If a decade ago companies had mobile-device management systems and five years ago brought in machine-learning models, the next layer is managing fleets of agents, code-enabled, decision-touching, workflow-embedded. Agent 365 sits at the frontier of that evolution.

What To Follow

In the coming months it will be important to see how quickly businesses adopt Agent 365 and how they define their agent-governance frameworks. Will organisations treat agents as employees, contractors or entirely new class of entity?

We should also watch how Microsoft and its partners refine the agent-integration model: how agents plug into human workflows, how organisations measure return on agent deployment and how safety and audit functions evolve.

Another key question: Which types of agents gain early traction in real-world business use, procurement bots, finance agents, customer-service bots or others, and how Agent 365 supports, monitors and scales these use cases.