Microsoft CEO Satya Nadella Says AI Agents Should Be Managed Like Employees

Microsoft CEO Satya Nadella believes businesses will need to start treating AI agents more like employees as artificial intelligence becomes increasingly integrated into corporate operations.

According to comments reported by Business Insider, Nadella said organizations will need to provide AI agents with identities, sandboxes, and policies that govern how they operate. He argued that AI systems are evolving beyond simple assistants and are becoming active participants in workplace processes.

The remarks come as Microsoft continues to expand its focus on artificial intelligence, making security, quality, and AI transformation central priorities across the company.

AI Agents Becoming a New Workplace Reality

Nadella said the technology industry is undergoing a significant platform shift, moving away from traditional application-centric computing toward systems built around AI agents.

“We’re moving from building operating systems and devices for apps to agents,” Nadella said.

The Microsoft chief executive believes AI agents will increasingly perform tasks on behalf of users, requiring organizations to manage them in ways similar to human workers. This includes assigning identities, defining permissions, and establishing clear operational boundaries.

His comments echo concerns raised by Microsoft AI CEO Mustafa Suleyman, who recently warned that advances in artificial intelligence could affect many white-collar jobs within the next 18 months.

Project Solara Highlights Microsoft’s Vision

At the company’s Microsoft Build conference, Nadella unveiled Project Solara, a new operating system designed to function largely behind the scenes.

Unlike traditional operating systems that focus on launching and managing applications, Project Solara is intended to coordinate and support AI agents. A key component of the platform is the Agent Shell, which can load, customize, and manage multiple cloud-based AI agents.

The project reflects Microsoft’s broader vision of a future where AI agents become a major part of everyday business operations.

Security and Oversight Remain Essential

While expressing optimism about AI’s potential, Nadella emphasized that trust will be critical for widespread adoption.

He said organizations must prioritize security, containment, manageability, and observability when deploying AI agents. These safeguards would allow businesses to monitor agent activity, enforce policies, and ensure systems operate within approved limits.

As AI agents gain access to sensitive corporate data and workflows, Microsoft believes strong governance frameworks will be necessary to maintain reliability and accountability.

Managing AI Agents Can Be Challenging

Despite Microsoft’s enthusiasm for AI, Nadella acknowledged that managing large numbers of agents can be difficult.

He revealed that he currently uses around 100 AI coding agents. However, coordinating those agents through chat-based interfaces creates a significant cognitive burden.

The admission highlights one of the emerging challenges of AI adoption: while agents may improve productivity, organizations will also need effective tools to oversee and coordinate increasingly complex AI ecosystems.

In other developments, Microsoft has expanded its MAI family of AI models with new functionality. Meanwhile, AI company Anthropic has suggested that the rapid pace of AI development could eventually require some form of global pause or coordinated oversight to ensure the technology remains safe and beneficial.

Via Windows Central