- Anthropic collapsed Claude chat and Cowork, its autonomous work agent, into a single Claude interface on September 16, so users no longer pick a mode before starting.
- The update ships Claude Docs and Claude Slides in beta and pulls Claude Design into every conversation, with editable documents and decks downloadable as PowerPoint or PDF.
- Rollout starts on Pro and Max across web, desktop and mobile, with Team and Free plans next and enterprise admins given 30 days notice before features switch on.
Anthropic collapses chat and agent into one Claude
Anthropic stopped treating conversation and autonomous work as separate products. As of September 16, the Cowork agent that ran multi step tasks on a user's files and the standard Claude chat window are the same surface. A person can ask a question, then ask Claude to build the document that answers it, without leaving the thread or choosing a tool first. The company described the change on its own blog as removing the need to decide upfront whether a request is a quick chat or a longer job.
The release also adds two products aimed squarely at the office. Claude Docs and Claude Slides, both in beta, let Claude produce editable documents and presentations inside a conversation, deliver a deck live, or export to PowerPoint and PDF. Claude Design, previously a standalone tool, now appears in any chat. Claude still asks before acting by default, and users can raise that autonomy so it works through a task and surfaces only the items that need review.
Distribution is staged. Pro and Max subscribers get the merged experience first over the coming weeks, Team and Free follow, and enterprise administrators control activation with 30 days advance notice. Existing users need to do nothing, since the capabilities appear when a conversation calls for them.
| Claude Docs | Editable documents built inside a conversation, beta |
| Claude Slides | Presentations you can deliver live or export, beta |
| Claude Design | Now available in every chat, previously standalone |
| Exports | PowerPoint and PDF |
| First on | Pro and Max, then Team and Free |
| Platforms | Web, desktop and mobile |
| Enterprise | Admin controlled, 30 days advance notice |
Why the superapp race now runs through the chat box
The merge is a strategy statement, not an interface cleanup. Anthropic is betting that the AI agent belongs inside the same box people already type into, rather than in a separate workspace they have to remember to open. That collapses the distance between asking and doing, the same shift Santage tracked when ChatGPT moved from asking to doing with its own agentic push.
I could have Claude pull up my legal research database, and it would pull all the cases, read them, figure out which other cases I might need.Andrew Keller, describing the merged Claude, via Anthropic
The competitive logic is direct. OpenAI is assembling ChatGPT, its Codex coding agent and potentially its Atlas browser into one destination, and Google is threading Gemini through Workspace. A single Claude that writes, codes, designs and executes is Anthropic's answer to the everything app, and it protects the consumer relationship at the moment agents make chat the control layer for real work.
The cost math is the catch. Agentic sessions burn far more compute than a chat reply, and a Stanford study found coding agents consumed roughly 1,000 times the tokens of a chat reasoning task. Folding autonomous work into the default interface means more users will trigger that heavier path by accident, which is why Anthropic keeps the ask before acting setting on and why it just locked in $517 billion in compute commitments. The company that makes the agent effortless to reach also has to pay for every time someone reaches for it.
Santage is committed to independent, transparent journalism. This article is produced in accordance with Santage's Editorial Standards and aims to provide accurate and timely information. Readers are encouraged to verify information independently.