Just now, Claude has become the only access point, and the super "all-in-one office suite" is going head-to-head with Microsoft.

Starting from today, Claude has officially completed an epic integration!

Anthropic has just announced that Claude Cowork and Chat have been fully merged, leaving only one unified "Claude" entry for users going forward.

From now on, no one will ever have to struggle with deciding whether to send a task to Chat or assign it to Cowork.

Whether you want to ask a casual small question, or need to write a high-stakes report that takes hours to polish, you can just hand it all over to Claude.

You only need to open the new version of Claude to access all the capabilities of Cowork, Design, Docs and Slides.

Claude's full office suite has now been fully interconnected for good.

Most importantly, all your accumulated historical conversations, Skills and connectors are preserved completely without any loss.

After receiving a task, Claude will independently judge the complexity of the task and invoke the corresponding capabilities.

Even if you close your laptop and leave, Claude can continue to work on the task.

However, users will need to wait a little longer: the unified version of Claude will be rolled out to Pro and Max subscribers in the coming weeks.

The Ultimate Integration of Claude

The Real All-in-One Office Suite Is Here

The most groundbreaking part of this evolution is that it completely breaks the traditional human-computer interaction paradigm.

In the past, people were accustomed to the pattern of "you ask one question, the AI gives one answer, and everything resets as soon as you close your computer".

However, this has led to a new gap: there are more and more separate entrances —

You go to Chat for Q&A, open Cowork for complex tasks, switch to Design for creation, and jump to another tool to make presentations.

In reality, real work is never divided along the lines of product features.

A business analysis report may require multiple capabilities at the same time: collecting information, analyzing data, drafting documents, making presentations, designing visual elements, and iterating repeatedly based on feedback. It naturally spans multiple AI capabilities.

Many users have reported that it is very troublesome to constantly judge which workspace a task should belong to!

Therefore, Anthropic decided to eliminate this unnecessary choice.

Now, no matter if you just throw out a vague casual request, or directly upload an entire dozens-of-page complex business report to it, Claude can take full autonomous control and keep making progress in the background even after you close your laptop and shut the screen.

Word, PPT, Design Software

All Integrated Into Claude

Strictly speaking, Cowork has not disappeared.

All the capabilities it previously undertook have been integrated into the main Claude entry.

The old pattern was: Claude Chat was responsible for answering questions; Claude Cowork was responsible for executing complex tasks; Claude Design was responsible for visual creation.

Now, Claude uses one unified entry, allowing the AI to independently judge the task requirements, whether it needs to call Chat, Cowork, Design, or document, spreadsheet and PPT capabilities.

The less judgment work left to humans, the higher the quality of the delivered results.

What's more impressive is that Claude's conversation interface has now directly evolved into a one-stop native "creative factory".

You don't need to jump between different software at all, you can complete typesetting, slide production and visual design directly in the chat box:

Claude Docs: Supports quickly drafting a one-page proposal or a full-length document;

Claude Slides: Can convert the document you just wrote into well-structured presentation slides with one click;

Claude Design: Can seamlessly generate matching visual drafts and interface prototypes in the same workspace.

On the surface, this is Anthropic adding more office features, but the deeper change is: Claude is starting to transform from an "answer window" to a "result workspace".

The previous traditional AI interaction pattern was: user asks a question -> AI outputs text -> user copies the text to other software.

Now, the interaction pattern of the brand-new Claude has become: user puts forward a goal -> AI creates a document -> AI makes a presentation -> AI generates designs.

All content stays in the same context.

The interaction granularity is even as precise as the pixel level —

You can either edit a specific line of text in the document yourself, leave a comment on a certain page of the PPT for Claude to rewrite, or directly select a certain design element with the mouse to drag and adjust it freely.

When everything is ready, you can directly present the slides from Claude, export the documents in your commonly used format, or share all the content through a single link.

This is essentially a complete Harness reconstruction.

For example, before you go out now, you can ask Claude about the progress in the business pipeline from last week, and then add: "Write it in our usual format, mark all delayed items, and organize the key points into five slides for the leadership meeting."

If there is any unclear point, Claude will take the initiative to ask you for clarification.

On your way to the office, you can check the progress on your mobile phone. The moment you sit down, the report document is ready, with comments from your colleagues included, and the PPT can be opened directly.

Moreover, since both are generated from the same conversation, the content of the PPT is completely consistent with the report.

In addition, you can set it to run automatically every Monday, and Claude will write the report on its own without you reminding it!

OpenAI, Anthropic, Google

Competing for the Super Entrance

This adjustment of Claude also reflects that AI giants are moving towards a similar direction.

OpenAI hopes to make ChatGPT the unified entrance, integrate Codex, so that chat, coding, research and Agent capabilities are all centered around ChatGPT.

Anthropic's path is even more clear: to make Claude an AI assistant that can execute work over a long period of time.

Google is trying to deeply embed Gemini into its existing ecosystem including Gmail, Docs, Workspace and other products.

The three companies have different routes, but similar goals: to compete for the new operating system entrance in the AI era.

The large-scale integration of Doubao's resources in China may also follow a similar line of thinking.

Fewer Buttons, More Intelligence

In the past, software competition was about who had more features, who provided more tools, and who had more entrances.

But the AI era may be completely the opposite: the best product may present users with the fewest possible options.

Why does this change happen? Because all the complexity is absorbed by the AI.

Users only need to say: "Achieve this goal."

The remaining questions: what tool to call? How to break down the task? When to ask the user for feedback? When to continue execution?

All of these can be handled by the AI on its own.

Most users only use the Chat feature, wasting the powerful Agent capabilities of separate modules such as Cowork and Design.

The integration of Claude Cowork into Chat is not a simple product merge.

The signal it sends is: Agent is evolving from a special standalone feature to the underlying architecture of AI products.

The competition between AI assistants in the future will not only be a competition of model capabilities, but also a competition of who can build a more powerful Harness, so that the same model can truly be integrated into the workflow of human work.

References:

https://claude.com/blog/cowork-is-now-claude

This article is from the WeChat official account "AI Era" (ID: AI_era), edited by Ma Ke, Taozi and Aeneas, and republished with authorization from 36Kr.