Anthropic merges Claude chat and Cowork, and adds Docs and Slides for document output

Anthropic has combined Claude's chat and Claude Cowork, the assistant's working environment, according to ITmedia News. The stated consequence is that a single conversation is now said to cover everything from simple back-and-forth exchanges through to multi-step, complex work. That is the whole of the architectural claim: previously the conversational assistant and the place where work got done were presented as separate surfaces, and the merge is described as collapsing that seam into one thread.

The report also says document and slide creation arrive with the change: Claude Docs, and a slide-making capability. The headline names both Docs and Slides; the summary text names Docs and refers to a slide creation function. Either way, this is an addition on the output side — producing the artefacts people actually circulate — rather than a change to the underlying reasoning.

Rollout is described as gradual, beginning with the Pro and Max tiers. Nothing in the evidence says when the rollout completes, which regions are covered first, what the tier limits are, or whether anything changes for API users as opposed to the chat product.

Several things a practitioner would want to know are simply absent. How much context carries across a long multi-step task, whether Cowork's own features survive the merge rather than being replaced, what document formats or sizes are supported, and whether any of this alters cost or rate limits — none of it appears in the source. The report is a product announcement summary, not a hands-on test, and it should be read that way until someone measures it.

Who should care: people who currently split work between a chat window and a separate tool, and who feel that handoff as the slow part of the job. The interesting claim to test is not that documents can be generated — plenty of tools do that — but whether one conversation really can carry a multi-step task from discussion to finished document without the user reconstructing context along the way.

A bounded way to check this for yourself: pick one recurring task you already do in several steps, run it end to end in a single conversation once the feature reaches your account, and note every point where you had to restate something the assistant should already have known. Those restatements are the real measure of whether the merge changed anything for you, and they are the notes worth keeping while the rollout is still gradual.