Satya Nadella Says Microsoft’s AI Agents Must Be Managed Like Employees

Satya Nadella explained that Microsoft assigns identities, access limits, and control policies to its AI agents. The company’s Agent 365 tool is part of this approach...

Scraper SystemJune 05, 2026, 11:06 PM

Satya Nadella Says Microsoft’s AI Agents Must Be Managed Like Employees

Microsoft has started treating its growing number of AI agents like employees to manage them effectively. The company’s top executive, Satya Nadella, stated that it is essential to define what these agents can access, what they cannot, and how they are monitored.

Nadella explained that managing AI agents requires assigning them identities, defining secure workspaces, and establishing policies to govern them. According to him, these steps form the foundation for controlling agents used within the company.

Need for Identity, Secure Zones, and Policies

Nadella emphasized that agents must be structured similarly to employees. In this context, each agent should have a unique identity, operate only within authorized areas, and have its activities subject to later review.

It was also shared that the company has developed a toolset called Agent 365 for this purpose. This package includes Entra, used for digital identity and network access, and Purview, which tags data generated by AI agents.

The Most Challenging Aspect of Management

Nadella noted that managing AI agents alongside human employees is not an easy challenge for companies. He said that while many organizations invest heavily in AI, how these agents will collaborate with human workers remains unclear.

He mentioned that this issue has become a daily management concern for him personally, stating he can run about 100 AI coding agents simultaneously. Trying to direct so many agents individually through a chat interface creates a significant mental load.

Emphasis on Security and Traceability

Nadella said trust in AI agents can only be achieved through security, restrictions, manageability, and traceability. He explained that these factors will determine how agents are used within corporate systems.

During the discussion, Reid Hoffman also announced that after 10 years he will leave Microsoft’s board to return to what he calls a "founder mode." However, the main focus of the conversation was how companies will regulate AI agents.

As AI investments grow, the role of these systems within organizations is becoming a more debated topic. Microsoft’s approach positions agents not just as technical tools but as a digital workforce requiring corporate policies and oversight.

Satya Nadella Says Microsoft’s AI Agents Must Be Managed Like Employees

Satya Nadella explained that Microsoft assigns identities, access limits, and control policies to its AI agents. The company’s Agent 365 tool is part of this approach...