A new offering from Microsoft, called Agent 365, aims to assist businesses in managing their growing number of AI tools. Rather than being a platform to create enterprise AI systems, Agent 365 serves as a management tool that allows businesses to handle their bot workforce as if they were human employees. This functionality is particularly vital for companies increasingly utilizing generative AI agents across their operations.
Agent 365 is part of Microsoft’s early access program and provides a streamlined interface to keep track of various AI agents, monitor their performance, and adjust their settings. Charles Lamanna, president of business and industry for Microsoft’s Copilot, emphasizes the need for management tools for AI agents, similar to those used for managing human employees and devices.
The expectation is that AI agents will proliferate in the workplace. Lamanna foresees a future where businesses might employ up to a million agents for tasks that range from simple email management to overseeing major business processes, noting that Microsoft itself employs millions of agents internally. Without proper oversight, the sheer number of bots could complicate operations and raise security concerns. Agent 365 addresses this by providing a centralized registry for an organization’s active agents, complete with identification numbers and details about their use.
Incorporating security features to monitor the activities of these agents in real-time, Agent 365 strives to ensure that data integrity is maintained as interactions among humans, agents, and applications occur. As companies pilot AI agents, they face critical questions about the technology’s safety, especially regarding vulnerabilities like prompt injection attacks that can manipulate agent behavior.
Despite concerns about the implications of deploying numerous AI agents, Lamanna argues that adapting to this change is essential for future success, likening it to the transition toward personal computers or the internet. Though the technology behind AI agents is still evolving and often exhibits flaws, its integration into workplace settings seems poised to accelerate. Lamanna suggests that 2025 will mark a significant increase in agent adoption, with expectations for even greater prevalence by 2026.
