Microsoft on Monday released a new series of AI agent products aimed at enterprise customers, evolving its Copilot tool from an AI assistant to a wider ecosystem of agentic features.
The launch, dubbed by Microsoft “Wave 3” of its 365 Copilot rollout, centers on a new feature called Copilot Cowork, built in collaboration with Anthropic and powered by the independent generative AI vendor’s Claude model. The move represents something of an about-face for Microsoft, which was originally closely tied to Anthropic’s archrival, OpenAI. The tool is designed to handle complex, multi-step tasks from a single user request. By accessing a user’s work apps and context, Microsoft said it can provide more targeted, tailored responses to the task at hand.
Copilot Cowork is being piloted with select customers and is set to enter research preview in March as part of Microsoft’s new Frontier Worker product suite.
Microsoft said Claude will also now be available across the full Copilot Chat experience.
Meanwhile, Microsoft is planning a new generation of agentic features for Word, Excel, PowerPoint and Outlook, with employees gaining an improved chat experience, the ability to create and augment artifacts and tools to build their own agents.
Also, Microsoft said it will make its Microsoft Agent 365 system generally available on May 1. Governance and transparency — increasingly crucial considerations for enterprise AI deployments — featured prominently in this release. “The speed of agent development and proliferation tells us customers see value, but without guardrails the pace of adoption turns into blind spots, diminished ROI and real security risk,” Jared Spataro, Microsoft’s chief marketing officer for AI at Work, wrote in a blog post on the launch. “As AI agents become more capable and autonomous, trust is nonnegotiable.”
Pitched as addressing these concerns, Microsoft said Agent 365 gives IT and security leaders a single platform to “observe, govern, manage and secure” agents across their organization.
The vendor also highlighted interoperability as a key motivation behind the updates, with Microsoft highlighting siloed as causing difficulties for organizations looking to uptake AI technologies at scale.
“Leaders end up managing overlapping tools, inconsistent experiences, and rising costs as teams bring their own AI into the business,” Spataro said. “The result is broken context for users, unnecessary overhead for organizations, and the burden of model selection pushed onto people who just want to get work done.”
