Quick answer
On September 16, 2026 Anthropic began merging Claude Cowork into the main Claude chat. In the merged experience you describe what you need and Claude decides whether to answer a quick question or take on longer work such as research, reports, spreadsheets, and presentations, which continues after you close your laptop. By default Claude asks before acting; you can opt in to continuous work with check-ins only for bigger decisions. Claude Docs and Claude Slides launched in beta on paid plans, and Claude Design now works inside conversations. Rollout: Pro and Max first across web, desktop, and mobile, then Team and Free, with Enterprise administrators getting at least 30 days' notice.
Cowork was Anthropic's separate mode for longer, agent-style tasks. Keeping it separate meant users had to guess which door to walk through. The merge removes the choice: one Claude, and it works out how big the job is.
What is new
- One experience: no Cowork toggle; Claude assesses the task and uses the capabilities formerly behind Cowork from a normal conversation
- Claude Docs (beta): documents created and edited collaboratively inside the conversation
- Claude Slides (beta): presentations you can present directly or export to PowerPoint and PDF
- Claude Design in conversations, also in beta on paid plans
- Work persists: chats, projects, artifacts, connectors, and skills remain available, so a task can be resumed later
The permission model
The default is to ask before acting. Users who want Claude to run through a task can opt in to continuous work, in which it checks in only for closer decisions. That is the opposite default from "agent first" products and matches the approach Anthropic has taken since Computer Use: the model proposes, the person approves, unless the person has said otherwise.
How it compares
Thirteen days later OpenAI launched ChatGPT Space, with Pages and Team Tasks, and the always-on Dots. The two companies are converging on the same idea, chat that becomes work, from different directions: Anthropic from a single assistant that produces documents and decks, OpenAI from a team workspace where agents hold standing jobs. For an individual on a Pro plan, Claude's version is available now and simpler. For a team on Business or Enterprise, Space does more coordination.
Related reading
Bottom line
The merge is a usability fix with a strategic point: the assistant and the agent are the same product now. If you are on Pro or Max you already have it; if you are on Team or Free, it is coming; if you administer Enterprise, you have at least a month to prepare.
