**Anthropic** integrated its work agent tool, ‘Claude Cowork’, into the standard chat interface, and unveiled new beta features: ‘Claude Docs’ and ‘Claude Slides’. Now users can handle all tasks in a single Claude chat without needing to choose a separate mode before starting their work.
Key takeaways
Claude identifies the nature of the query on its own; if it’s a simple question, it responds right away, and if it’s a complex agent-style task, it switches to long-form work.
Claude Docs and Claude Slides are being introduced in beta, and the completed outputs can be downloaded as PowerPoint (PPT) and PDF.
First, it will be rolled out sequentially to paid “Pro” and “Max” subscribers, and later expanded to “Team” and free pricing plans.
Claude CoWork absorbed into the main chat
Anthropic unveiled this overhaul through its company blog on Wednesday. The plan is to first deploy the features to Pro and Max subscribers across all web, desktop, and mobile channels over the coming weeks, and then apply them to Team and free plans afterward. Enterprise customers’ administrators will receive advance notice at least 30 days in advance.
After the integration, Claude brings the functionalities of CoWork and “Claude Design” into a single conversational environment. Using the existing conversation context and stored skills, along with various connectors, it determines on its own what workflow is needed for the task. Even if you close the laptop, the work continues in the background, and users can check the progress on mobile.
Users can co-write documents with AI, automatically generate presentation slides for presentations, then edit and present them themselves—or download and save them as PowerPoint or PDF files. Repetitive tasks scheduled to a timetable are also possible. For example, you can set it to automatically prepare a five-slide weekly report deck for your leadership meeting every Monday.
Also see: Bitcoin mining faces a 50% hash rate gap… while AI siphons up power and facilities
Anthropic takes direct aim at OpenAI
According to Anthropic, users found it inconvenient that they have to decide each time whether to start the task in chat or in CoWork, and they also pointed out that work started in one space does not naturally carry over into the other space—something they consider inefficient. In a customer case the company cited, economist **Andrew Keller** explained that Claude found and read relevant precedents in a legal research database and even organized the results into a folder for review.
This UI redesign eliminates the cumbersome “mode selection” step. Native support for PowerPoint and PDF formats should lower the adoption barrier for companies that heavily depend on the Microsoft and Google office ecosystems. At the same time, the move further clarifies the competitive landscape with OpenAI. In July, OpenAI already strengthened its business lineup by launching “ChatGPT Work,” which combines its coding tool Codex with chatbots.
Similar usability confusion has long been a problem that has also held back OpenAI’s dual strategy of “ChatGPT” and “ChatGPT Work.”
In January of this year, Anthropic first introduced CoWork in the form of a research preview, expanding the agent-style workflow of its developer tool “Claude Code” to non-developer roles. Since then, it has added features such as project-based management, third-party connectors, plugins, scheduled execution, and always-on cloud execution. In April, it released “Claude Design” as a separate tool specialized for website and prototype production, and the standalone Claude Design continues to be offered using the same approach even after this integration.
Next article: Shiba Inu, fixes Shibaarium wallet link error, but… SHIB slips 6% over the past week