Anthropic merges Claude chat and Cowork, adds Docs and Slides

Anthropic is merging Claude chat and Cowork into a single Claude interface. Instead of choosing a separate mode before starting, a user can describe the task and Claude can answer directly or use Cowork-style capabilities for longer-running work while retaining the conversation’s context, skills and connectors. Claude Design is also being brought into the same conversational surface.

The company is also launching Claude Docs and Claude Slides in beta. Docs supports collaborative drafting and editing, while Slides can be edited and presented inside Claude and exported to PowerPoint or PDF. The unified experience is rolling out first to Pro and Max users on web, desktop and mobile over the coming weeks, with Team and Free plans due to follow.

The product change removes a decision that previously sat with the user: whether a task belonged in ordinary chat, Cowork or a separate creation tool. Anthropic is moving that routing decision into Claude itself, making agentic execution a capability of the default assistant rather than a separate destination.

That also broadens the amount of knowledge work Anthropic wants to keep inside Claude. A conversation can now move from a question to a longer-running task and then to a document, presentation or design without a manual context handoff. The direction increases competition with productivity suites, although beta creation tools are still a long way from proving they can replace mature collaborative document and presentation software.

Anthropic says users retain control over how autonomous Claude is allowed to be. By default, Claude asks before taking actions, while users can choose a setting that lets it continue until something needs closer review. Enterprise administrators will decide when the new beta creation tools become available in their organisations, and Anthropic says it will give them advance notice before the unified experience changes.

That control layer matters because the interface now makes higher-autonomy behaviour easier to invoke from an ordinary conversation. The practical safety question will be whether permissions, previews and auditability remain clear when a task moves from answering to acting.

The rollout is gradual and Docs, Slides and Design remain beta products. Anthropic has not published evidence yet on reliability across complex collaborative documents, presentation fidelity, usage costs or how often users choose the more autonomous setting. The next useful signals will be enterprise availability, administrative and audit controls, pricing or usage-limit changes, and evidence that the new creation tools hold up in production workflows.