Anthropic is simplifying the way people use Claude for larger tasks. The company is bringing regular Claude chat and its Cowork agent-style workspace into one interface, reducing the need to decide which mode to open before starting a job.
The change matters because AI assistants are increasingly moving beyond question-and-answer conversations. People now ask them to research, organize files, create documents and presentations, and carry out multi-step work that can take longer than a normal chat response.
Chat and Cowork are becoming one experience
Anthropic says the unified interface lets users start naturally in Claude and then hand off more complicated work without switching to a separate Cowork area. Existing Cowork features remain available, but the workflow should feel more continuous.
In practice, that means a simple conversation can turn into a larger delegated task when needed. A user could discuss an idea first and then ask Claude to produce a more substantial output without rebuilding the context in another mode.
Claude will show more of what it is doing
Another useful change is greater visibility while Claude works. Anthropic says the interface can show files being created and other progress information during a task.
That is important for agent-style AI. When a job takes several steps, users need to know whether the assistant is researching, creating a file, waiting on something or has gone in the wrong direction. Better progress visibility can make long-running AI work easier to supervise.
What about computer use?
Anthropic says the interface also provides access to computer-use capabilities. These tools allow an AI agent to interact with software interfaces rather than only generating text.
Computer use can be powerful, but it also requires more supervision than a normal chatbot response. Users should review important actions, especially when an agent is working with accounts, external services or files that are difficult to restore.
Why AI companies are simplifying agent modes
Early AI products often separated chat, research, coding and agents into different modes. That can expose useful controls, but it also makes users learn which mode fits each task.
A unified interface moves the complexity behind the scenes. Ideally, the user describes the goal and the assistant chooses the right tools while still showing enough information for the user to remain in control.
What users should expect
The practical benefit is convenience rather than a completely new AI model. Claude users should be able to move from quick questions to larger delegated jobs with less friction, while keeping better visibility into the work being performed.
Source: Anthropic product announcement and current technology reporting, September 16, 2026.