Anthropic brings chat and Cowork together in one Claude for enterprises
IA en un minuto newsroom · Editor: Jon Elgezabal
In 30 seconds
Every conversation in Claude Enterprise now opens as a chat and, when code has to run or files have to be handled, Claude takes the task to the cloud without the user switching to Cowork. Pro and Max plans got it in September, and in companies that already had an organization it will switch on by itself on November 16. Anthropic keeps usage-based pricing and adds the Efficient mode and spending limits to keep consumption in check.
On Friday, October 23, Anthropic said that Claude Enterprise, its plan for businesses, now offers the experience that merges Claude's chat and Cowork, the mode designed for handing over complete jobs, into one. Pro and Max got it in September; it now reaches companies on the web version, the desktop app and mobile.
With the change, users no longer decide whether to open a chat or a Cowork session before they start. Every conversation opens as a chat, and a simple question is answered as before. If the task calls for running code or handling files, Claude carries on with that same conversation in the cloud, inside an isolated environment run by Anthropic, and delivers the result (a document, an analysis or a presentation) for the person to review. Since the work happens in the cloud, it continues even if the laptop is shut, and it can be resumed from the web, desktop or mobile. Chats, projects, connectors, skills and plugins move over to the new experience.
Each session runs on its own virtual machine, with internet access switched off by default on Enterprise and connector credentials stored outside that environment. The controls the company already applies to chat, such as single sign-on (SSO) or the Compliance API, remain in force, and Anthropic notes that by default it does not train its models on that data.
Administrators manage it from the organization's settings. According to the admin guide, in organizations that already existed it arrives switched off and will turn itself on on November 16, 2026, unless an admin has already enabled or disabled it, in which case that choice is respected. Access can also be limited to certain roles. For now it is not offered to organizations that run with zero data retention, that use the HIPAA-ready setup or that manage their own encryption keys, which stay as they were. Current Enterprise customers will get an email about the rollout in their organization.
Claude Enterprise is still priced by usage. Anthropic acknowledges that, depending on the task, agent work can consume more tokens than a quick query, and it gives administrators four controls: the new Efficient mode, which limits the experience to the Sonnet and Haiku models (sessions open by default on Claude Sonnet 5.5 with a low effort level); permissions deciding which models are open to each role; spending limits per organization, group and person; and monthly budgets shared by a group. In reports, this usage shows up on a line of its own, "Chat and Cowork unified", with no change to the way it is counted or billed.
Why it matters · analysis and opinion
The change goes beyond the interface: by merging chat and agent work, any employee can end up launching long tasks in the cloud without having decided to, and that shifts both consumption and what needs watching. For IT managers, the time left until automatic activation is the window to decide which roles should have it, which connectors stay open and whether it pays to start with Efficient mode. The move also says something about the market: Anthropic is pushing its corporate customers towards more agentic use without touching the pricing model, so the bill will depend on how it is used. Organizations with zero data retention or their own encryption keys are left out for now, so their decision is postponed.
Official source: Anthropic · Written with the help of AI: how we make the news
