AI Sefa Aydın 5 min read Claude guides

What Is Claude Cowork? What It Does, Which Plans Include It and How to Use It (2026)

What Is Claude Cowork? What It Does, Which Plans Include It and How to Use It (2026)

Short answer

Claude Cowork is the multi-step work mode inside Anthropic's Claude. You give it a goal, Claude makes a plan, works with the files and connected tools you allow and delivers a finished report, spreadsheet or presentation. It comes with the Pro, Max, Ultra, Team and Enterprise plans and runs on desktop, web and mobile depending on the plan.

Claude Cowork is the mode inside Claude that carries out multi-step work for you. In a chat, you ask a question and get an answer. In Cowork, you give Claude a goal: it makes a plan, works with your files and connected tools and hands back finished work. In the corporate training sessions I run, Claude Cowork is a feature people often ask about, because it goes one step beyond chat. In this article, I explain what it is, who can use it, how to start and where it is safe to use at work.

What is Claude Cowork and how is it different from chat?

Cowork launched as a research preview in the desktop app in January 2026. Anthropic introduced it as "Claude Code for the rest of your work". The idea is the same: you describe the result, and Claude plans and carries out the steps. The difference is that Cowork focuses on knowledge work such as reports, spreadsheets, presentations, contract reviews and file organization instead of code.

This is the much-discussed idea of agentic AI applied to everyday work. The model doesn't just produce an answer; it runs a task from start to finish.

FeatureClaude chatClaude CoworkClaude Code
What you give itA question or requestA goal and filesA coding task
What you getA text answerA finished file or taskChanged code
Access to your filesOnly what you uploadFolders you allow on desktopThe project folder
Who it is forEveryoneTeams doing knowledge workDevelopers

If you are curious about the coding side, I cover it in What is Claude Code.

What can you do with Claude Cowork?

According to Anthropic's official getting started guide, Cowork can:

  • Run multi-step tasks: work runs in the cloud, and the session continues even if you close your laptop.
  • Use local files: in the desktop app, it can read and write the files you allow.
  • Act in a browser: it can take steps on the web with the built-in browser or the Chrome extension.
  • Run scheduled tasks: with the /schedule command, you can repeat a task on a cadence you choose.
  • Use sub-agents: it can split a large task into parallel workstreams.
  • Build spreadsheets and presentations: it can deliver the output directly in those formats.

The examples on Anthropic's product page are business focused too: preparing the same report every month and comparing it with the previous period, flagging variances across regional spreadsheets, reviewing a stack of contracts against company rules as a first pass and collecting objections from sales call notes in one place.

Which plans include Claude Cowork?

Cowork is available only on paid plans: Pro, Max, Ultra, Team and Enterprise. It isn't sold separately; it's included in the subscription. Where you can use it depends a little on the plan:

  • Desktop: on macOS and Windows, and in beta on Linux, on all paid plans.
  • Web and mobile: on Pro, Max, Ultra and Team (in beta on Team), and on Enterprise once an owner turns it on.

I compare what each plan includes in Claude pricing. For the current price, check Anthropic's pricing page.

How do you use Claude Cowork?

Getting started takes four steps:

  1. Open Claude on the web or in the desktop or mobile app.
  2. Select Cowork in the message box. If the unified experience has reached your account, skip this step: Claude brings in Cowork's capabilities on its own when a task involves your files.
  3. Describe the task: which files, what output, what format.
  4. Review Claude's approach and let it run if it looks right.

For work with local files, the browser or computer use, the desktop app has to be open and connected. You need an internet connection throughout the session.

Chat habits are not enough when you describe a task. My rule is simple: state the file name, the format and the audience of the output up front. "Merge the three spreadsheets in this folder, highlight regional variances in red and add a one-page summary for management" works far better than "prepare the report".

What does "Claude Cowork is now just Claude" mean?

Anthropic is turning Cowork from a separate mode into part of the main Claude experience. In the words of the product page, "Claude Cowork is now just Claude": you ask for what you need, and if the task involves files, Claude uses Cowork's capabilities by itself. This unified experience is rolling out to Pro, Max and Ultra first, with more plans to follow. Until it reaches your account, Cowork still appears as a separate option next to Chat.

Is Claude Cowork safe, and what should you watch out for?

Because Cowork is an agent, it carries different risks than chat. It can reach the internet, write to files and take actions in connected tools. Anthropic states this risk clearly in its guide. When companies roll it out, I look at three things:

  1. Approval mode: there are three modes. Manual asks you before actions that need approval. Auto runs actions through an automated safety check. Skip runs no automated check and only pauses for tools you set to "Always ask". Use Skip only when you fully trust every action, connector and file involved. I recommend that new teams start with Manual.
  2. File scope: give Claude access to the folder the task needs, not your whole drive. It already asks for permission before permanently deleting files, but a narrow scope is the stronger safeguard.
  3. Outside content: hidden instructions in a web page or email can steer an agent the wrong way. This risk is called prompt injection. Always review the output of tasks that work with untrusted sources.

How should companies start with Claude Cowork?

A common mistake I see in my corporate training sessions is turning Cowork on for everyone at once and saying "try it". The result is scattered. A better order works like this: pick one recurring task with clear rules (a monthly report, a proposal document, a first-pass contract review). Write one task description for it, let a few people use it for two weeks and check the outputs. Only then roll it out more widely.

If you want to set this up with your team, in my corporate AI training we work with Cowork on your own files and workflows.

Frequently Asked Questions

What is Claude Cowork?

Claude Cowork is the mode in which Claude carries out multi-step work for you. You give it a goal, Claude makes a plan, works with your files and connected tools and delivers a finished report, spreadsheet or presentation.

Is Claude Cowork free?

No. Cowork is available only on paid plans: Pro, Max, Ultra, Team and Enterprise. It is not sold separately and comes with the subscription. Check Anthropic's pricing page for the current price.

Which devices does Claude Cowork run on?

On desktop, it runs on macOS and Windows, and in beta on Linux. On the web and mobile, it is available on Pro, Max, Ultra and Team (in beta on Team), and on Enterprise once an owner turns it on. The desktop app must be open to work with local files.

What is the difference between Claude Cowork and Claude Code?

Both are agent tools that carry out work step by step. Claude Code works on code for developers. Cowork focuses on knowledge work beyond code, such as reports, spreadsheets, presentations and contract reviews.

Does a Cowork task stop when I close my computer?

Tasks that run in the cloud continue even if you close your laptop. Steps that need local files, the browser or computer use require the desktop app to be open and connected.

Want to learn to do this with AI?

In individual training sessions, live on your own project, and in corporate training for teams, I teach step by step how to do everything covered in this article with AI. Write in and I will get back to you the same day.

Message on WhatsApp See individual training Corporate AI training for your team
Sefa Aydın

Sefa Aydın · AI Trainer & Consultant

An AI trainer and consultant who has worked on the Turkey projects of world-famous luxury brands. He teaches, hands-on, how every kind of work is done with AI: design, video, branding and vibe coding.

About Sefa Aydın →   LinkedIn ↗

Related articles

Want to apply this to your own work?

Tell me what you are working on. I reply the same day with next steps.

Your details are used only to reply to you. See the privacy notice.

Message on WhatsApp