Anthropic merges its chat and agentic products into one AI assistant in push to build a superapp | Fortune
Anthropic is folding its chat and autonomous task-execution products into one Claude interface, while adding collaborative document and presentation tools.
Anthropic Combines Chat and Agentic Tools
Anthropic announced that it is merging its Claude Chat interface with the more autonomous Cowork product to create a single AI assistant. The move follows a broader industry effort to build an “everything app” that can replace a growing array of separate tools for searching, writing, coding, file management and routine work. The concept promises users a seamless experience in which the assistant keeps the thread of a task and automatically selects the appropriate tools, eliminating the need to pick a specialized AI product or manually transfer context.
Benefits for Companies and Users
Industry analysts note that a unified AI front gives companies a tighter relationship with users, encourages longer sessions within one product, and simplifies the bundling of premium features into subscription plans. For enterprises, a single interface helps centralize and control employee AI usage, a priority as businesses look to manage productivity and data security. OpenAI is pursuing a comparable strategy, planning to fold ChatGPT, its Codex coding agent and possibly the Atlas browser into a single “superapp” that would let the AI move between research, coding and explanation without switching contexts. Anthropic’s own coding‑focused offering, Claude Code, stays separate for the time being.
Streamlined User Experience
Anthropic describes the integration as a way to simplify how users interact with its technology. Under the new model, users no longer need to decide in advance whether a request belongs to Claude Chat or to Cowork, the more autonomous task‑execution engine. Instead, the combined assistant can determine the best approach on the fly, while still giving users the option to intervene.
Potential Cost Impact
The shift also raises cost considerations. Agentic products such as Cowork, which reason through multi‑step tasks and make tool calls on a user’s behalf, typically consume far more tokens than a straightforward chat exchange, and token usage is the primary billing metric for AI providers. A study by Stanford’s Digital Economy Lab found that agentic coding tasks used roughly 1,000 times more tokens than simple chat‑based reasoning. If routine queries are routed through the more token‑intensive agentic pathway by default, compute expenses could rise even for basic questions.
User Control and Enterprise Settings
Anthropic assures users that they will retain control over Claude’s level of autonomy. By default, the assistant will request permission before taking any action, though users may allow it to proceed continuously and intervene only when a specific review is needed. Enterprise administrators will also be able to toggle the beta Docs, Slides and Design features for their organizations, deciding when and how those capabilities become available to employees.
The consolidation aims to deliver a more cohesive AI experience while addressing both user convenience and the operational demands of businesses seeking a single, manageable AI platform.
Source: fortune.com · 2026-09-16