Blog Post

Windows IT Pro Blog
4 MIN READ

Unlocking Secure Agentic Productivity with Windows 365 for Agents

Joydeep_Mukherjee's avatar
Mar 17, 2026

The future of enterprise automation isn’t just about building smarter agents, it’s about enabling you to do more with agents, across more of the organization, directly within environments companies already rely on. This means AI agents need access to enterprise applications, not just APIs.

With our latest innovations announced at Ignite 2025 and our vision for AI agents, Windows 365 is becoming a foundational execution layer where people run agents to work through software UI, unlocking productivity while preserving enterprise trust boundaries. Windows 365 for Agents is the next evolution of Cloud PCs—extending the trusted Windows 365 platform from human users to AI agents so organizations can run agentic workloads securely, at scale, and under enterprise governance.

Scaling AI agent productivity securely

While Model Context Protocol (MCP) servers enable automation in modern apps that have an API-based interaction layer, many mission-critical workflows still live in UI-driven legacy apps and bespoke internal sites where APIs are incomplete or unavailable. Equally important, business logic and decision context often live in the UI, not the API: warnings, approvals, conditional flows, and exception handling are communicated visually, and API-only agents miss this operational nuance. As a result, there are limits to how even highly capable agents can operate where real work happens. Windows 365 for Agents closes this gap by giving AI agent workloads a secured Cloud PC to reliably work through the same Windows and browser UIs employees use.

With Agent 365, (generally available May 1) enterprise governance extends natively into the agent execution environment. Agent 365 gives organizations a single, trusted way to observe, govern, and secure agentic workloads—whether built by Microsoft, partners, or organizations—using the same identity, security, and management systems they already use for employees.

That governance also applies to agents working in Windows 365 for Agents Cloud PCs—bringing managed, policy-controlled desktop execution to agentic workloads. Together, Windows 365 for Agents and Agent 365 create a consistent and secure environment to run and govern agents, with each platform serving a distinct role:

  • Agent 365 determines what work an agent should be able to perform and how it fits into the broader business workflows, using shared organizational policies and identity to govern behavior and access.
  • Windows 365 for Agents defines where that work executes, providing Cloud PCs to enable agents to operate inside their own desktops and applications, including UIdriven and nonAPI systems, within a fully managed and auditable environment.

These Cloud PCs can run Windows or Linux (coming soon) and are:

  • Entra ID-joined
  • Microsoft Intune-managed
  • Policy-enforced
  • Auditable by design

This means identity, access controls, compliance policies, and audit logs are inherited from the enterprise environment from day one.

To see what this looks like in practice, the video below shows how an expense report agent managed in Agent 365 can complete a routine expense report task using an assigned Cloud PC.

Managing agents consistently – no matter who builds them

We know that enterprises are increasingly drawing upon internal knowledge to build and deploy AI agents that operate directly within their core Windows applications. At the same time, a growing ecosystem of agent-builder platforms is giving organizations new, external paths to create and deploy agents tailored to their unique business needs. Agent 365 serves as the control plane that brings these internally and externally created agents together under a single governance model.

To support this, Microsoft has partnered with innovators like GenSpark to ensure we’re providing the tools required to bring externally created agents into production securely, compliantly, and at scale—without sacrificing enterprise governance.

“What stood out to us about Agent 365 is that Microsoft focused on trust, not just technology. Security, governance, and compliance are built in from the start, enabling IT leaders to deploy AI with confidence. Through this partnership, we’re able to bring GenSpark’s AI directly into Teams, Microsoft Word, Microsoft 365 apps, Copilot, and computerusing workflows powered by Windows 365 for Agentsall under enterprise IT control. At the same time, employees get a seamless experience right where they work. That’s the future we’re excited to help build.” 
— Eric Jing, CEO, GenSpark

Expanding what agents can do without expanding risk

Combining Agent 365 management with Windows 365 for Agents execution lets enterprises extend automation into the long tail of UI-based work that lives outside the API world—without sacrificing governance or control.

Key outcomes include:

  • Agents operate across applications and browsers,
  • Complete multi-step operational tasks, and
  • Operate under the same policies and controls as human users.

This is what work-aware AI can look like in practice.

Ready to get started?

  • If you’re an agent maker, IT leader, or developer interested in being among the first to try Windows 365 for Agents, sign up here to express your interest in our preview.
  • If you are an enterprise looking for a no-code agent building experience powered by Windows 365 for Agents, get started with Microsoft Copilot Studio today: Automate web and desktop apps with computer use (preview)
  • If you are an agent developer looking to solve agentic work for enterprises – get started with your Agent 365 integration today: Microsoft Agent 365 SDK and CLI | Microsoft Learn

Continue the conversation. Find best practices. Bookmark the Windows Tech Community, then follow us @MSWindowsITPro on X and on LinkedIn. Looking for support? Visit Windows on Microsoft Q&A.

Updated Mar 16, 2026
Version 1.0
No CommentsBe the first to comment