Microsoft unveils AI bot monitoring system Agent 365 for enterprise environments

Microsoft announced Agent 365, a new dashboard tool designed to help IT and security teams track, manage, and secure the growing number of AI agents running inside enterprise environments.

The company stated that companies will be able to see and control every AI bot in their network.

As more companies integrate generative AI agents into their workflows, automating everything from email triaging to procurement tasks, IT leaders are increasingly wrestling with “agent sprawl.”

Microsoft estimates that by 2028, organizations could be running 1.3 billion AI agents across workplaces.

Microsoft builds Agent 365 to help companies manage AI bots

Microsoft believes that companies will soon become accustomed to AI agents being part of their teams, as many businesses already utilize them to handle tasks that previously required a person’s full attention. However, the company also notes that customers still struggle to understand how these bots connect, communicate, or integrate into larger workflows, leaving business leaders confused and concerned about potential mistakes. 

The company’s commercial business leader, Judson Althoff, said business leaders spoke directly to Microsoft and requested a tool that monitors all activities of every AI agent in a workplace. These individuals stated that they needed software to help them measure the return on investment or determine whether these agents work well together to achieve their company’s goals.

Microsoft even shared a great example with people, where a company uses an “inventory agent” and an “out-of-stock agent,” yet managers cannot see the whole process from start to finish. They need a system that connects every part and shows where each agent fits in the entire chain. 

Agent 365 helps IT teams monitor, protect, and block AI bots across all platforms

Microsoft introduced Agent 365, which acts as a control plan to monitor all active agents within a company and provide detailed information about the most efficient or problematic bots.

IT teams will also be able to measure the time saved by each agent on a weekly basis, and all this information will be consolidated in a single dashboard to prevent confusion and mistakes.

Teams can also approve new agents before they start working to prevent untested or risky software from running on company systems. Additionally, they have the authority to block agents who fail to behave properly or pose security risks. 

Agent 365 uses the same rules for identity and access that companies have for human employees, so every agent will only see the data it needs to and perform the actions that its workplace allows it to do. With these rules and responsibilities clearly defined, the bots won’t accidentally expose sensitive information or create security risks, and employees can also monitor the actions of the agents they use. 

Those who have started using Agent 365 early have already demonstrated the significant benefits of having a single, central dashboard that controls all these AI agents. Ernst & Young (EY) is already replacing its own internal catalog of AI agents with Microsoft’s tool so that it can track all agents in one place. Other companies, such as Okta, are developing similar tools to monitor and control AI agents, which only underscores the growing demand in the market for AI supervision.

Microsoft decided to launch Agent 365 at a time when other big tech companies, such as Amazon and Google, are also developing their own AI tools, and organizations are investing a significant amount of money in infrastructure. The company has already spent $34.9 billion on capital investments last quarter, with half of that amount allocated to GPUs and CPUs for Azure, supporting the development of AI. 

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.