Microsoft Rebuilds Copilot App Around Office, Natural-Language Coding, and Agents

Microsoft is rolling out a restructured Copilot app with a tabbed interface, direct integration of Office apps, natural-language app building, and a new agentic mode the company calls Autopilot, according to a blog post published today by Jared Spataro, Microsoft's Chief Marketing Officer of AI at Work.

The redesign is the latest in a series of facelifts for Microsoft's dedicated AI app, following a minor refresh earlier this year. It arrives at an odd moment in the company's AI strategy. Microsoft may be pulling back from embedding AI and Copilot features directly into Windows, but it continues to invest in the standalone Copilot app as a destination for both consumers and businesses.

The stakes are clearest on the enterprise side. While Microsoft has not made a convincing case for serious consumer adoption of Copilot, businesses that constantly process data, generate reports, and create presentations represent a far more concrete market for the technology. The new app is built squarely around that use case.

A Tabbed Interface With Three Modes

The redesigned app splits functionality across three tabs: Home, Code, and Autopilot.

The Home tab serves as the default workspace, where standard AI chats and other work occur. Microsoft says the tab will eventually do more than host conversations. The company plans for Home to intelligently route requests between Copilot's different modes, matching the complexity of a query to the right capability. A simple question might stay within Chat mode. A more complex query could trigger the full Copilot mode, where users can delegate work to the assistant rather than simply ask it questions.

Microsoft has not specified when the intelligent routing capability will ship.

Natural-Language Coding Without Programmers

The Code tab targets users with no programming skills. It lets people create small tasks and tools through natural-language requests, with Copilot handling the actual construction.

"Describe an app, tracker, dashboard, automation, or workflow in natural language, and Copilot chooses an approach and builds it," Spataro wrote in the blog post.

The output ranges from simple desktop widgets to full dashboards for managing data. The positioning puts Copilot in direct competition with a growing category of no-code and AI-assisted development tools, but with the advantage of living inside an app that Microsoft is already distributing to its existing customer base.

Autopilot: Agents That Work Without Supervision

The most significant addition is the Autopilot tab, which unlocks agentic capabilities. Users can direct Copilot to automatically respond to threads or handle recurring tasks without any further input.

Spataro offered a concrete example of what a full agent-driven process looks like in practice.

"It can do things like set up and manage a full supplier review process: building the schedule and workback plan, then handling prep, meetings, and follow-ups on its own, right down to reaching out to stakeholders for updates," Spataro wrote.

That description goes beyond the chat-and-generate pattern that defines most consumer AI assistants. A supplier review process that includes scheduling, meeting preparation, and stakeholder outreach is a multi-step workflow with human dependencies — the kind of task that has historically required a project manager or coordinator.

Why It Matters

The redesign signals where Microsoft sees the near-term revenue and retention for Copilot: not in casual consumer chat, where adoption has been uneven, but in enterprise workflows where the cost of manual coordination is measurable. Direct Office integration makes the app a control surface for the productivity suite that already dominates corporate environments.

The Autopilot tab also aligns Microsoft with the broader industry shift toward agentic AI, in which assistants act autonomously over extended workflows rather than responding to one prompt at a time. If autonomous multi-step execution like the supplier review example works reliably, it would mark a meaningful change in what office AI tools can credibly promise. If it does not, the gap between the demo and daily use will define how quickly enterprises adopt it.

Microsoft has not announced pricing changes or general availability dates tied to the new tabs.