For the past two years, most enterprise AI has been a smarter search box. You ask a question, you get an answer, you go do the thing yourself. That dynamic just shifted.

On March 9, 2026, Microsoft launched Copilot Cowork — an AI agent built into Microsoft 365 that doesn’t just answer questions. It takes on work. Real work: triaging your inbox, drafting documents, coordinating multi-step projects, and taking action across Outlook, Teams, Excel, and the rest of the Microsoft 365 suite.

This is what the industry has been calling “agentic AI,” and it’s now available inside the tools that most knowledge workers already use every day.

What Copilot Cowork Actually Does

The pitch is simple: describe the outcome you want, and Cowork handles the steps to get there.

You tell it something like “prepare a summary of all client feedback from last month’s meetings and draft follow-up emails for anything still unresolved.” Cowork builds a plan, pulls signals from your emails and meeting transcripts via Microsoft’s “Work IQ” layer, drafts the emails, and checks in with you at each stage before taking action.

It’s not running autonomously without oversight. Microsoft was deliberate about that. Every action Cowork recommends is visible. You can pause, redirect, or approve at any checkpoint. The goal is what they’re calling “controlled delegation” — you hand off the work without losing visibility or control.

Built on Anthropic’s Claude model, Cowork runs entirely in the cloud within a customer’s Microsoft 365 tenant, meaning it’s covered by existing enterprise data protection policies. That’s a meaningful distinction from Anthropic’s own Claude Cowork product, which runs locally on a user’s device. For organizations with strict data governance requirements, the cloud-native approach matters.

The Bigger Picture: Microsoft 365 Wave 3

Copilot Cowork is the headline feature of what Microsoft is calling “Wave 3” of Microsoft 365 Copilot — a platform overhaul that includes two other significant products:

Agent 365 is the control plane for managing every AI agent operating in an organization. IT and security teams get a single dashboard to observe, govern, and audit agent behavior. It’s generally available on May 1 at $15 per user per month.

Microsoft 365 E7 is a new enterprise bundle ($99/user/month) that packages productivity, AI, identity, and security under one license — a direct response to enterprises asking for simplicity as they juggle multiple AI tools.

Copilot Cowork itself requires a $30 per user per month license and is currently in Research Preview, with broader rollout expected through the Frontier program by end of March.

Why This Happened When It Did

The timing is not accidental. Anthropic launched its own Claude Cowork product in January 2026 — and the market reaction was swift. Software stocks dropped sharply as investors realized that AI companies were starting to compete with the productivity software incumbents that had hosted them.

Microsoft lost roughly $220 billion in market cap in one week following that launch. Copilot Cowork is the response: rather than watch Anthropic eat into its market, Microsoft has turned its massive distribution advantage and the $30 billion compute deal it struck with Anthropic into a product that lives where enterprise workers already are.

It’s a smart move. Enterprise buyers don’t want to add another tool. They want their existing tools to get smarter.

What This Means for Business

If you’re a business owner or executive still thinking of AI as a research assistant, this launch signals that the transition period is ending.

The shift is from AI as a tool you use to AI as a workforce participant that handles work on your behalf. That’s a materially different operating model, and it affects how you think about headcount, workflows, and where human judgment should actually sit.

A few practical implications:

Rethink what “delegation” means. Tasks that currently require a coordinator — pulling together information from multiple sources, tracking follow-ups, preparing briefing documents — are now candidates for AI delegation. The question is no longer “can AI do this?” but “do we have the workflow clarity for AI to do this well?”

Governance becomes essential. Agent 365 launching alongside Cowork is not a coincidence. As AI agents begin taking real actions in enterprise systems, knowing what they’re doing and why becomes a compliance and risk question, not just a productivity question.

The real competitive advantage shifts. If every business eventually has access to the same AI delegation tools, the edge goes to companies that are better at defining outcomes, structuring context, and reviewing AI output critically. Data literacy and workflow design become core competencies.

Skills gaps will widen faster. Teams that understand how to work with AI agents — not just use them — will outpace teams that treat this as just another software feature. The businesses that close that gap quickest will move faster than their competitors.

Enterprise DNA has been preparing businesses and professionals for exactly this transition — from understanding data, to deploying AI agents, to building the organizational capacity to scale with tools like Copilot Cowork. The infrastructure is arriving. The question is whether your team is ready to use it.

---

If you’re building with Claude or Codex right now, grab the free Working With Claude field guide. Thirty-two pages on the full ecosystem, Claude Code in depth, and how to roll agents out properly. Get the free guide.