Skip to main content

Anthropic Merges Claude Cowork Into Chat With Docs and Slides

·1445 words·7 mins
Anthropic Claude Claude Cowork Claude Chat Claude Docs Claude Slides AI Agents AI Productivity
Table of Contents

Anthropic Merges Claude Cowork Into Chat With Docs and Slides

Anthropic is changing how users interact with Claude by merging Claude Cowork and Claude Chat into a single experience.

Instead of deciding whether a task belongs in Chat, Cowork, or Design before starting, users can now describe what they want Claude to accomplish and let the system determine which capabilities are required.

The update also brings Claude Docs and Claude Slides directly into conversations, while Claude Design can be invoked without leaving the current chat. Anthropic says the goal is to make Claude capable of handling everything from simple questions to complex, multi-step projects while preserving the same conversation context throughout the workflow.

The rollout begins with Pro and Max subscribers across Claude’s web, desktop, and mobile applications, with additional subscription tiers planned later.

🧩 Cowork and Chat Become One Claude
#

Anthropic says Cowork was originally designed as a standalone workspace for larger, more involved tasks, while Claude Design focused primarily on visual creation.

In practice, users often needed both capabilities during the same project.

That created a workflow problem: users had to decide which tool to use before starting, and work created in one environment did not automatically carry over to another.

The new approach removes that initial decision.

Claude can now determine which capabilities are needed based on the task itself. Cowork and Design functionality can be invoked directly from a normal conversation while inheriting the conversation’s existing:

  • Context
  • Skills
  • Connectors
  • Project information
  • Previous instructions

This turns Claude Chat from a primarily conversational interface into a broader workspace for delegating multi-stage tasks.

πŸ“„ Claude Docs and Slides Launch Inside Conversations
#

The merger also introduces Claude Docs and Claude Slides as integrated creation tools.

Users can ask Claude to create a document, collaborate on its contents, revise individual sections, and download the finished result.

For presentations, Claude can generate slides directly from the same conversation and prepare them for editing, presenting, or downloading.

Supported export formats include:

  • PowerPoint
  • PDF

Claude Design can likewise be accessed directly within conversations instead of requiring users to begin in a separate workspace.

All three capabilitiesβ€”Docs, Slides, and Design integrationβ€”are currently in beta and available to paid users.

Enterprise administrators will have control over when these features are enabled for their organizations.

πŸ”„ Shared Context Connects Documents and Presentations
#

One of the more important changes is that documents and presentations created from the same conversation can share context automatically.

For example, a user could ask Claude to investigate changes in a business pipeline, identify delayed items, write a weekly report, and then summarize the findings in a five-page presentation.

Previously, these tasks might have required moving information between separate tools.

Under the new workflow, Claude can generate both outputs from the same conversation, helping keep the presentation synchronized with the underlying report.

Users can then make manual edits, leave comments for Claude on individual slides, and continue collaborating without switching workspaces.

πŸ€– Claude Can Continue Working After You Leave
#

Anthropic is also emphasizing asynchronous task execution.

Users can delegate a complex task to Claude and close their computer without necessarily ending the workflow. Claude can continue working on the task, allowing users to check its progress later from another device.

This changes the interaction model from:

Ask β†’ Wait β†’ Receive

to something closer to:

Delegate β†’ Claude works β†’ Review β†’ Refine

If Claude encounters something it cannot resolve independently, it can ask the user for clarification.

This approach is particularly relevant for lengthy reports, presentations, research tasks, and other projects that would normally require several rounds of manual work.

πŸ“… Recurring Tasks Can Be Automated
#

The new workflow can also extend beyond one-off requests.

For example, a user could instruct Claude to prepare a business report every Monday.

Instead of manually issuing the same prompt each week, Claude can proactively begin preparing the report according to the configured schedule.

This effectively turns Claude into a recurring work agent rather than simply a conversational assistant.

The distinction becomes particularly important for workflows involving regularly updated information, such as:

  • Weekly business reports
  • Recurring presentations
  • Pipeline summaries
  • Team updates
  • Periodic research
  • Repetitive document preparation

πŸ“± One Shareable Workspace Across Devices
#

Anthropic says content created through Claude Design, Slides, and Docs can be consolidated into a single shareable link.

That link can be opened directly from a phone or another device, allowing users to review and interact with their work without returning to the original desktop session.

Visual content can also be modified through direct interaction.

Users can select an element and move it themselves, or describe the desired change to Claude and allow the system to perform the edit.

The result is a workflow that combines traditional manual editing with natural-language instructions.

πŸ›‘οΈ Users Retain Control Over Claude’s Actions
#

Anthropic is also providing different levels of interaction and confirmation.

By default, Claude asks for confirmation before performing actions that require user approval.

Users who prefer a more autonomous workflow can instead enable a mode where Claude continues working and only interrupts when human review is required.

This creates two distinct interaction styles:

Confirmation-oriented mode

Claude pauses when an action requires approval.

Continuous-work mode

Claude proceeds through the workflow and contacts the user when a decision or review is specifically needed.

Anthropic emphasizes that users retain final control over the process.

πŸ“ Example: From Weekly Report to Management Deck
#

Consider a weekly business-report workflow.

A user could ask Claude to:

  1. Review changes in the business pipeline.
  2. Identify delayed items.
  3. Write the report using the company’s established format.
  4. Condense the findings into a five-page management presentation.
  5. Prepare the files for review and download.

Claude can use the same conversation context to generate both the report and presentation.

If colleagues have left notes or additional information in the working materials, those details can be incorporated into the document and reflected in the presentation.

The user can then review the result on a phone, modify a sentence, comment on a particular slide, or download the completed presentation as a PowerPoint file.

The workflow no longer requires users to decide where each individual stage should take place.

πŸ—‚οΈ Existing Cowork Content Will Remain Available
#

Users who already work with Cowork do not need to rebuild their projects from scratch.

Anthropic says existing Cowork content will remain available, including:

  • Chat history
  • Projects
  • Artifacts
  • Connectors
  • Skills

Users can open Claude and continue working from their existing environment.

Those who primarily use Claude Chat do not need to change their workflow either. When they encounter a task that would normally require more extensive work, they can simply ask Claude to handle it.

πŸš€ Rollout Starts With Pro and Max
#

The unified Claude experience will initially roll out over the coming weeks to Pro and Max subscribers.

The rollout covers:

  • Claude web
  • Claude desktop
  • Claude mobile apps
  • Existing users
  • New users

Anthropic says users do not need to manually enable the feature.

Support is expected to expand later to Team and Free plans.

For Enterprise customers, administrators will be notified at least 30 days before changes affecting organizational settings.

Claude Design will also remain available as a standalone experience for users who prefer to use it that way.

βš™οΈ From Separate Tools to an Agentic Workspace
#

The broader significance of this update is less about the individual Docs or Slides features and more about how Anthropic is changing Claude’s interaction model.

Previously, users had to understand the boundaries between different Claude products and select the appropriate environment.

Now, the user can describe the desired outcome while Claude decides which internal capabilities should be used.

That is closer to an agentic workspace model, where the interface becomes less important than the task itself.

A single request can potentially span research, analysis, document generation, presentation creation, visual editing, and recurring automation without requiring the user to manually transfer context between specialized tools.

πŸ”­ Anthropic Expands Claude Beyond Chat
#

The merger of Cowork and Chat marks a broader shift in how Anthropic positions Claude.

Claude is no longer being presented solely as a conversational interface where users ask questions and receive responses. The new workflow is designed around delegation, asynchronous execution, document creation, presentation generation, visual editing, and recurring tasks.

The practical goal is straightforward: users describe the work, and Claude determines how to complete it.

Whether this approach ultimately replaces the need for separate AI productivity tools will depend on reliability, output quality, integrations, and how much control users want over individual steps.

For now, Anthropic’s strategy is clear: instead of making users choose between Chat, Cowork, and Design, it is bringing those capabilities together into one Claude experience.

Related

NVIDIA Restricts Claude Usage as Enterprise AI Data Risks Grow
·2082 words·10 mins
NVIDIA Anthropic Claude Enterprise AI AI Security Data Privacy AI Agents Nemotron Sovereign AI
Dario Amodei Calls for AI Pacing as OpenAI Delays 2026 IPO
·1689 words·8 mins
AI Safety Anthropic OpenAI Dario Amodei AI Agents AI Governance AI Alignment Frontier AI
Claude Formalizes Fermat’s Last Theorem in Lean in 11 Days
·2192 words·11 mins
Claude Anthropic Fermat's Last Theorem Lean Formal Proof AI Mathematics Mathlib Proof Assistants AI Research