Anthropic merges its chat and agentic products into one AI assistant in push to build a superapp
The company is folding its chat and autonomous task-execution products into one Claude interface, while adding collaborative document and presentation tools.
Anthropic is merging its agentic tool, Claude Cowork, and its chatbot, Claude Chat, into a single product. The change, which takes effect Wednesday, is also happening alongside the introduction of two new beta products, Claude Docs and Claude Slides, alongside a wider rollout of Anthropic’s Claude Design tool, which will now be available inside every conversation rather than as a standalone product.
It’s also an attempt to court enterprise customers as businesses attempt to centralize and control their employees’ AI use. OpenAI has been pursuing a similar strategy and plans to fold ChatGPT, its Codex coding agent, and potentially its Atlas browser into a single “superapp” meant to cut down on fragmentation and let the AI move between researching, coding, and explaining within one session. Anthropic’s own coding-focused product, Claude Code, remains separate for now.
Anthropic frames the change as simplifying the user experience. Under the new setup, users no longer have to decide upfront whether a task should be given to Claude Chat or to Cowork, Anthropic’s more autonomous, task-execution product.
But the shift may also carry cost implications. Agentic products like Cowork, which reason through multi-step tasks and make tool calls on a user’s behalf, tend to consume more tokens than a simple chat exchange, and token usage is what AI companies bill on.
For example, a Stanford Digital Economy Lab study of coding tasks found that agentic tasks consumed roughly 1,000 times more tokens than simple chat reasoning tasks. If simple queries inside “one Claude” end up routed through more agentic, tool-calling machinery by default, that could push up the compute cost of answering even straightforward questions.
Anthropic said users will retain control over how autonomously Claude works. By default, Claude will ask before taking an action, though users can let it continue working and check in only when something needs closer review. Enterprise administrators, meanwhile, can decide when to enable the beta Docs, Slides, and Design features for their organizations.
