Anthropic Merges Claude Cowork Into Chat, Adds Docs, Slides And Design

Alongside the merge, Anthropic launched two new products, Claude Docs and Claude Slides, and extended Claude Design so that it now works directly inside ordinary conversations rather than existing only as a separate, standalone tool. In practice, that means a person can ask Claude for a document and write it collaboratively inside the conversation, or ask for a presentation and have Claude draft the slides directly, without switching to a different interface to do either. Work produced this way can be edited directly inside Claude, presented straight from the conversation, or downloaded as a PowerPoint file or PDF depending on the format needed. All three products, Docs, Slides and Design, are currently in beta on paid plans, and Enterprise administrators retain control over when their organisation gets access to them. Claude Design also continues to function on its own for anyone who prefers to use it as a standalone tool rather than inside a chat.

Anthropic's own description of how this is meant to work in practice centres on a fairly specific everyday scenario, someone with a weekly report due at noon asking Claude, before leaving for the day, what moved in a particular pipeline over the past week, then asking Claude to write it in the format the team normally uses, flag anything that slipped, and turn the highlights into five slides for a leadership meeting. If any part of that request is ambiguous, Claude asks a clarifying question rather than guessing. The person can check progress from a phone on the way into the office, and by the time they sit down at their desk, the finished report is waiting as a document, complete with a teammate's notes if the document was shared, and a matching slide deck ready to open, adjust and download, generated from the same underlying conversation so the two pieces of work stay consistent with each other. From there, a person can fix a line in the document directly, leave a comment for Claude on a specific slide, or select and move an individual design element, and once everything is finished, share the whole set of outputs through a single link that opens on a phone as easily as a laptop. The same workflow can also be scheduled to repeat automatically, so a recurring report gets started by Claude on its own each week without needing to be manually requested again.

Anthropic has also given users direct control over how much autonomy Claude exercises while carrying out a longer task. By default, Claude checks in and asks before taking an action rather than proceeding unprompted. People who would rather let Claude keep working through a task uninterrupted, checking in only when something genuinely needs closer human review, can turn that behaviour on instead, while retaining final say over anything Claude ultimately does or produces.

For most users, the practical change amounts to removing a decision they previously had to make upfront, since anyone who primarily uses Claude for quick chat interactions does not need to change how they work at all, and can simply try handing over a larger report or presentation they would normally have built manually the next time one comes up. The full list of specific capabilities included in the rollout is documented in Anthropic's Help Center for users who want a more granular breakdown before relying on the new workflow for something time sensitive.

The move reflects a broader pattern across AI assistant products this year, as providers increasingly move away from separating quick conversational use from longer, more autonomous task execution, instead building a single interface capable of scaling from a simple question to a multi‑step piece of work spanning documents, presentations and design without forcing the user to choose a different tool depending on the size of the job. Whether merging Cowork's more autonomous task handling directly into everyday chat changes how heavily people actually rely on Claude for larger, multi‑step work, rather than continuing to treat it primarily as a conversational tool for smaller requests, will likely become clearer as the rollout reaches the wider Team, Free and Enterprise user base over the following weeks.