Anthropic Makes Unified Claude Chat and Cowork Live for Enterprise

Anthropic has made its combined chat and Cowork experience generally available to Claude Enterprise customers, letting a single conversation hand off longer jobs to an agent that runs in its own cloud virtual machine. The company announced the change on 23 October, a month after the same setup reached Pro and Max subscribers in September.

The idea is simple. Every conversation starts as a chat. When a request needs code to run or files to be built, Claude carries on in a cloud session and comes back with the finished document, analysis or slide deck. Anthropic says the work keeps going after a laptop is closed, and can be picked up again on the web, desktop or mobile apps. Existing chats, projects, connectors, skills and plugins carry over.

For businesses already using Claude, the date that matters is 16 November. Anthropic’s help centre guide for administrators says the feature “will be automatically enabled on November 16, 2026” for organisations that have not made a choice.

Each agent task gets a sealed machine

According to Anthropic, agentic work happens in cloud sessions, “each running in its own isolated virtual machine with network egress off by default for Claude Enterprise and connector credentials kept outside the sandbox.” The help centre adds detail. Each conversation that needs a machine gets one, managed by Anthropic. After a few minutes idle, the machine stops and its disk is saved as an encrypted snapshot. The sandbox expires 30 days after it was created.

Administrators can set network access to off, package managers only, a custom list of domains or all domains. Connector policies are enforced by Anthropic outside the sandbox, on every connector call, the guide says. The setup builds on single sign-on and Anthropic’s Compliance API, and the company says data is encrypted in transit and at rest and is not used for model training by default.

Access follows roles. The standard User, Admin, Owner and Primary Owner roles get the feature by default, while custom roles get it only if it is included. “Access is checked on every request,” the guide says. Anthropic recommends a security review, a pilot group and two to four weeks of monitoring before expanding “in waves”.

Admins get Efficient mode and spending caps

Pricing for Claude Enterprise stays usage-based, Anthropic says, and it concedes that “agentic work can use more tokens than a quick chat.” Its answer is a set of controls: model access by role, spending limits at organisation, group and member level, and pooled budgets for groups.

The new one is Efficient mode. When switched on, the help centre says, members can use only Sonnet and Haiku models in the unified experience, and new sessions start on Claude Sonnet 5.5 at low effort. Other Anthropic products are unaffected.

Anthropic also makes a cost claim of its own. Billed through usage, it says, “Claude Opus 5.5 costs 40% less to run than Claude Opus 5 on typical workloads, and Claude Sonnet 5.5 costs up to 30% less than Claude Sonnet 5 for most work.” TSN has not tested those figures.

Not everyone can switch on. The help centre says the experience is not yet available to organisations using zero data retention, the HIPAA-ready configuration or customer-managed encryption keys. “HIPAA support is expected after general availability; CMEK support is planned for a later release,” it says. Anthropic’s announcement describes those customers as staying “on their current configuration”.

Three customers are quoted in the announcement. “Bringing them together has made everything simpler, both for our users and administrators,” said Daniel Padilla, chief technology officer of LabX. Harley Barnes of Quantium said the change “answers what clients kept asking us, ‘which Claude should I use’ and ‘which one does what’.”

Anthropic is pushing deeper into the enterprise stack

The launch fits a busy month for Anthropic’s business offering. It made Claude Code Review generally available for Team and Enterprise plans, added an Anthropic-run Claude option on Google Cloud, and opened a waitlist for Claude Duet, a proactive agent for Pro and Max plans. Services firms are lining up too, with HCLTech setting up a dedicated Anthropic business unit.

Default-on puts the burden on compliance teams

The opt-out design is the detail that deserves attention. Anthropic is giving administrators a pilot-first playbook while switching the feature on for everyone who has not answered by mid-November. For a regulated firm, that turns a product launch into a change-control deadline. The defaults are sensible, with egress closed and credentials held outside the sandbox, but any organisation with custom roles, data-retention rules or vendor-risk sign-offs will want to decide deliberately rather than inherit a setting.

Retention is the subtle trap. Conversations follow an organisation’s custom retention policy, yet the guide says Cowork cloud sessions a member starts directly sit outside it. Records managers who assume one policy covers every Claude surface may find that it covers less than they think.

Cheaper models will not cap the bill on their own

Anthropic’s cost pitch has two halves that pull in different directions. Per-task prices may fall, by its own figures, but agents that run code and work through files consume more tokens than chat. The net effect on a monthly invoice depends on how far staff lean on the new mode. That is why the governance tools carry more weight than the headline percentages. Efficient mode and per-group budgets let finance teams set a ceiling before usage patterns are known, which is the prudent order of operations.

The next marker is 16 November, when undecided Enterprise organisations switch on. The open question is when HIPAA-ready and CMEK customers, whose settings exist precisely because of strict data rules, will be able to follow.

Sources

• Anthropic, “Claude Cowork and chat come together for Enterprise”, 23 October 2026
• Claude Help Center, “Set up and manage the new unified Claude experience for your Enterprise organization”

Related stories

• Anthropic Makes Claude Code Review Generally Available for Team and Enterprise
• Claude Gets a Second Google Cloud Option, Run by Anthropic Itself
• Anthropic Says Its Claude Duet Agent Will Message You First
• HCLTech and Anthropic Sign a “Global Premier Partnership”, and HCLTech Sets Up a Dedicated Anthropic Business Unit