Anthropic is combining Claude Chat, Cowork, Code, Docs, and Slides in one workspace. The Artificial Intelligence (AI) system is supposed to decide whether you need a quick answer or want to delegate a larger assignment. The change initially affects Pro and Max subscribers who want to handle reports, documents, presentations, or coding tasks without moving between separate areas.
What does the new Claude bring together?
Regular chat and Cowork, the area intended for larger assignments, were previously separate. According to Anthropic, many people struggled to decide which one to use, especially because their capabilities overlapped. Under the new design, both operate as one Claude: you describe your request in a chat, and the system determines which features it needs.
The merger of Chat and Cowork also includes Artifacts, meaning content that can be created and edited inside a conversation, and Claude Design for websites and prototypes. Claude Docs and Claude Slides are new additions. Together, they move Claude beyond a conventional question-and-answer tool toward a workspace where you can create, revise, and distribute finished material.
The immediate benefit is less about a single new AI capability and more about simpler navigation. Simon Willison describes the move as a step toward a general agent, meaning a system that can break assignments into steps and carry them out rather than merely answering. His assessment of the unified Claude also adds some caution: figuring out which features and surfaces are active in a given situation may still take work.
The rollout begins with Pro and Max subscriptions on the web, desktop, and mobile and is expected to take place over several weeks. Free and Team access should follow later, while Enterprise administrators are due to receive at least 30 days’ notice. The reports provide no specific prices, regional restrictions, or firm schedule for every account.
How does delegated work run?
The agentic approach means Claude organizes an assignment on its own. “Agentic” describes AI that divides a goal into steps, uses tools, and continues working over an extended period. A task can keep running in the cloud, meaning on the provider’s remote computers, even after you close your laptop.
One everyday example taken from the reports is a document due at noon. You can start the assignment on your computer, close the device, and check its progress later through the mobile app. That is more convenient than a session that depends on an active browser and computer, although it does not guarantee that the finished report will be accurate.
Delegation goes further in Claude Code Projects. The previous version of a project was essentially a folder containing files and one chat; the redesigned beta turns it into an ongoing conversation managed by a coordinator. That coordinator creates threads, meaning separate work streams, which run as parallel cloud sessions and, according to MarkTechPost’s account of the new Projects, work on their own versions of the project.
The threads share goals, memory, and a library of uploaded files and generated results. You can open each one individually or manage them from the main chat. In a workplace example, one thread could change program code, another could run tests, and a third could prepare a pull request, which is a proposed change submitted for review.
Parallel work does not eliminate conflicts. If several threads alter the same code, their overlapping changes still have to be reconciled like ordinary software contributions. The Verge explicitly notes the potential for merge conflicts, so an AI coordinator does not remove the need for supervision or expert approval.
The redesigned Projects feature is launching in beta for selected Pro and Max subscribers who use Claude Code cloud sessions. Team and Enterprise access is expected later, along with integration into Cowork and regular Claude chats. Anthropic says support for local tools and locally stored code is coming, but the reports indicate that it is not available yet.
What do Docs and Slides add?
Claude Docs and Claude Slides are intended to make generated results reusable without moving them into another tool first. You can create a document or presentation from a chat, request revisions, or edit the material yourself. The reports say presentations can be downloaded as PowerPoint files or Portable Document Format (PDF), a fixed format commonly used for distribution.
A concrete workplace example is preparing a presentation: you can ask Claude for an initial slide deck, comment on unclear sections, and open the result later on your phone. In a document, you can draft sections with Claude, ask questions about completed passages, and leave comments. Documents and slides can be shared through a link, and Anthropic says other people can edit documents collaboratively at the same time.
This puts Claude closer to established office suites. The Verge compares Docs and Slides with Google’s productivity offering: Google Gemini already has close ties to Docs, Sheets, and Slides, while Anthropic is now adding similar workflows directly to Claude chats. The claim that Anthropic’s models are generally more advanced is an assessment, not an independent comparative test documented in the supplied reports.
Docs and Slides are launching in beta, meaning they are still test versions rather than finalized products. Editing, collaboration, and export could be useful, but they should not yet be assumed to match the reliability of a long-established office environment. Important reports and client presentations still require a manual review of figures, claims, formatting, and sources.
Pros and Cons of the unified Claude
Pros:
- Fewer interfaces – You no longer have to decide whether an assignment should begin in Chat, Cowork, or another area.
- Background work – Larger assignments can continue in cloud sessions after you close your laptop.
- Reusable results – Documents and presentations can be edited, shared, and exported in common formats.
- Parallel execution – Multiple threads can handle parts of a larger assignment at the same time under central coordination.
Cons:
- Reduced visibility – Claude makes more decisions about how work is divided, which can allow errors to develop less visibly.
- Beta software – Docs, Slides, and the redesigned Code Projects have not completed testing.
- Unclear cost – The reports provide no prices, while parallel agents increase token use, the processing units consumed by AI requests.
- Cloud dependence – Active assignments and uploaded files are handled in a provider environment whose privacy details are not explained in the sources.
The commercial incentive deserves attention. The Decoder notes in its coverage of Code Projects that more agents and more autonomous work also mean higher token consumption. Convenience and provider revenue are moving in remarkably compatible directions.
What does this mean in practice?
For beginners: A sensible first step is a clearly limited assignment with a result you can easily verify. You could ask Claude to turn existing information into a draft document or short presentation and then inspect every section yourself. This lets you explore the unified interface without handing a critical business process to a beta feature.
For advanced users: You can gain more value by structuring larger assignments around results that can be reviewed and by monitoring individual threads. In a code project, for example, you could run changes and tests in parallel, then inspect conflicts, test results, and proposed changes separately. Shared memory and the file library may simplify recurring work, provided the stored context remains current and correct.
For users in Switzerland, the main confirmed issue is currently plan-based availability: Pro and Max come first, with Free, Team, and Enterprise access later. The sources mention neither special Swiss restrictions nor the quality of German-language output, data locations, or applicable privacy arrangements. If you handle confidential business, education, or personal data, the announcement therefore does not provide enough information for a full assessment.
The unified interface described by TechCrunch addresses a genuine usability problem and connects creation, collaboration, and mobile monitoring. The larger change, however, is delegation: Claude is meant not only to generate text but also to coordinate and continue work over time. How reliably these workflows perform outside selected beta accounts has not been independently verified.
The new Claude therefore looks mainly like a cleaner and more capable workspace, not a fully autonomous specialist. Docs, Slides, and parallel cloud sessions could shorten real work processes, while review and responsibility remain with you. The unresolved risks are hard-to-see errors, rising resource consumption, and an unclear approach to sensitive data.
Sources
- Anthropic verschmilzt Claude Chat, Cowork und mehr in einem einzigen Claude-Produkt – The Decoder, 2026-09-16
- Anthropic Launches Claude Code Projects in Beta – MarkTechPost, 2026-09-17
- Claude Code relaunches Projects to manage multiple AI agents in the cloud – The Verge, 2026-09-17
- Anthropic baut Claude-Code-Projekte um: Entwicklung wird noch automatischer – The Decoder, 2026-09-17
- Claude Cowork and chat are now one Claude – Simon Willison’s Weblog, 2026-09-16
- Anthropic merges Claude chat and Cowork in one interface – TechCrunch, 2026-09-16
- Claude comes for Gemini with its own take on Docs and Slides – The Verge, 2026-09-16
Image: KATRIN BOLOVTSOVA via Pexels