Anthropic is killing off Claude Cowork as a separate product and folding its agentic behavior into the main Claude chat, while launching Claude Docs and Claude Slides, two web-native document tools that live at a shareable link. The move puts Anthropic in direct competition with Google’s Workspace suite and its Gemini integration.
The Verge first reported the update, which Anthropic calls “one Claude.” The change removes the choice users had to make between a regular chat and a Cowork session. Until now, Cowork was Anthropic’s agentic workspace inside Claude Desktop, built on the same architecture as Claude Code but without the terminal. Users picked Cowork for longer, multi-step tasks and regular chat for everything else, and Anthropic says people told them the frustrating part was deciding where a task belonged.
“We built Cowork as a separate place for bigger work, and Design for visual work,” Anthropic said in its announcement. “People used both, and told us the frustrating part was deciding where a task belonged.”
Under the new model, that distinction disappears. When a conversation needs to become a longer-running task, Claude handles it without forcing the user to switch modes or restart. The Cowork toggle will vanish as the rollout reaches each account. Anthropic says the underlying behavior is not going away: longer-running work, connector and tool use, task decomposition, and the ability to keep working while the user is away all become part of how Claude operates generally.
Docs and Slides are web-native, not file exports
The bigger product bet is the two new tools. Claude Docs and Claude Slides, both in beta, are working documents rather than files generated at the end of a chat. They live at a single shareable link and remain editable, including from a phone, which is closer to how Google Docs works than to how chatbots have traditionally handled output. Claude Design, previously its own surface, now works from inside any chat session as well.
Exports are supported but secondary. Claude Slides can be downloaded as PowerPoint or PDF, and Claude Docs can be exported to Microsoft Word or Google Docs, with more destinations planned. Enterprises that have already worked in Cowork keep their chats, projects, artifacts, connectors and skills, so nothing is lost in the transition.
One naming wrinkle: Anthropic has used the phrase “Claude Docs” before, but for something else entirely. In September 2025 the company renamed its developer documentation site from Anthropic Docs to Claude Docs as part of a brand consolidation. The document-authoring product announced this week is new and separate, which means searches for the term will now return two unrelated things.
The competitive math
The target here is not hard to identify. Google has been wiring Gemini into Docs, Sheets and Slides for two years, and Microsoft has done the same with Copilot across Office. Both incumbents treat the AI as a feature inside the document. Anthropic is inverting that: the chat is the primary surface and the documents are artifacts the chat produces and maintains.
That inversion has a real advantage in agentic work. A document that Claude can keep reading, editing and updating as part of an ongoing task behaves differently from a file that was generated once. If a user asks Claude to draft a report, pull data, revise it after new numbers arrive, and keep a version history, the chat-native document handles all of that in one thread. The traditional model would require re-uploading or re-prompting at each step.
The risk is equally clear. Google and Microsoft own distribution. Docs and Slides come preinstalled in the productivity stack billions of people already use, with admin controls, compliance certifications and storage bundled in. Anthropic has to convince users that a chat-first document tool is worth a separate tab, and that switching costs are low enough to justify keeping a second window open next to their existing suite.
There is also a pricing question Anthropic has not answered publicly. Google bundles Gemini features into Workspace plans that most organizations already pay for. If Claude Docs requires a Claude subscription on top of Workspace or Office seats, enterprise buyers will want to see a reason for the duplicate spend. The beta does not specify how the tools will be metered across plan tiers.
Context: a busy month for Anthropic
The launch lands amid other Anthropic news. The company has publicly filed for a Nasdaq IPO targeting an October listing at a valuation of $2 trillion or more, a filing that moved from confidential to public documents this week. Federal lobbying disclosures showed it spent a record $1.97 million in the second quarter as it worked through a Commerce Department episode that briefly took its flagship models offline over national security concerns, and CEO Dario Amodei has spent the past week calling for a slower pace of frontier development, a position that drew responses from politicians and rival lab leaders.
On the model side, Anthropic shipped Claude Fable 5.1 and Mythos 5.1 earlier this month, cutting cache-read costs 75% and adding invisible watermarking on outputs produced after August. The release drew some developer skepticism, particularly around the watermarking details. Novo Nordisk also teamed up with Anthropic this week for an AI drug development operation, part of a broader push by pharma companies into foundation-model partnerships.
For users, the practical takeaway is simple: the Cowork toggle goes away, chats gain the ability to run long tasks on their own, and documents produced in chats are now editable, shareable web objects rather than dead downloads. For Anthropic, the takeaway is that the company is done treating productivity as a side bet and is going straight at Google and Microsoft on their home ground, with an IPO balance sheet behind it.