Most guides to Claude Cowork are written for people who want to see everything it can do. This one is written for the owner of a 10-to-50-person business who wants to know how to use Claude Cowork for three or four things that matter, get them right, and not create a mess along the way. It covers what Cowork is, what it costs, five workflows worth starting with, and the short list of jobs Anthropic itself says to hold back.
What Claude Cowork is, in one paragraph
Cowork is the part of Claude built for handing off work rather than chatting. Anthropic describes it as using “the same agentic architecture that powers Claude Code, with no terminal required,” and the promise is that you “describe an outcome, step away, and come back to finished work.” In practice that means Claude plans a multi-step task, works through it, and produces documents, organized files, or a finished piece of research instead of a reply. On Team plans, sessions can run in the cloud, so a task keeps going after you close the laptop.
Two features make it useful for a business rather than just a power user. Connected folders: on the desktop app you grant access to specific folders, and Claude “can only read and write files in folders you've connected.” Scheduled tasks: you save a prompt once and it runs “hourly, daily, weekly, on weekdays, or manually,” remotely, “even when your computer is asleep or the Claude Desktop app is closed.”
What Claude Cowork costs
There is no separate Cowork price. It is included on every paid plan (Pro, Max, Team, and Enterprise), on desktop for macOS and Windows, with web, mobile, and Chrome access rolling out across those plans. For a team, the realistic cost is the Team plan seat, and the realistic constraint is usage: Cowork sessions are longer than chats, so the people running scheduled work all day are the ones who may need a premium seat.
Five workflows worth starting with
Each of these has the same shape: a clear input, a repeatable output, and a human who reads the result before anything leaves the building. That shape is what makes them safe first projects.
- The Monday client roll-up. A weekly scheduled task that reads last week's email threads and project-tool updates for each active client and writes a one-page status: what shipped, what's waiting on the client, what's at risk. The account manager reads it Monday morning instead of assembling it. Needs your email and project-tool connectors, read access only.
- The morning “needs you” list. A weekday task that triages the owner's inbox into three buckets: needs a decision, needs a reply someone else can draft, and handled. It drafts nothing and sends nothing; it produces a list. That restraint is the point in month one.
- Proposal first drafts from your own past work. Connect a folder of past proposals and a folder of call notes. Ask Cowork to draft the next proposal in the house structure, pulling scope language from the three most similar past documents and flagging every number it was not given. The output is a draft in your folder, not a sent document.
- The monthly reporting pack. Point it at the folder where the month's exports land (ad platform CSVs, time tracking, invoicing) and have it produce the client-facing report in your template, with a short cover note on what moved. Run it manually the first two months, then schedule it.
- Meeting notes to tasks. After a call, drop the notes in a folder and let Cowork propose the task list in your project tool. This is usually the first write action a team is comfortable with, and it is a good one to practice approvals on: Cowork shows you each task before it creates it.
The setting that matters most: approval mode
Every Cowork task runs in one of three modes. In Anthropic's words, Automatically approve means “Claude reviews each action for safety before it runs and blocks anything it determines to be unsafe.” Manually approve is what Anthropic recommends when “the task touches sensitive files, accounts, or sites” or when “mistakes would be hard to undo, like sending messages or making purchases.” Skip all approvals exists, and there is no reason for a business to use it in the first month.
On Team plans there is a useful default underneath this: connectors that can write to another system require per-task approval unless an Owner turns on “Always allow,” which is off out of the box. Leave it off until the 30-day review. The teams that trust Cowork later are the ones who watched it work first.
What to keep off the schedule for now
Anthropic's own safety guidance for Cowork is unusually direct, and it is worth quoting rather than paraphrasing: “Don't schedule tasks that access sensitive files, send messages on your behalf, make purchases.” A few practical translations for a small business:
- Keep payroll, banking, and HR folders out of connected folders entirely. Cowork can't reach what you haven't connected.
- Let it draft client emails; don't let a scheduled task send them. A person presses send in month one, and probably month six.
- Treat anything Cowork reads from the open web as untrusted. Anthropic describes prompt injection as instructions “embedded in external content that Claude reads,” and the risk is highest when a task can both read outside content and take an action. Reading a competitor's site is fine; reading it and then editing your CRM in the same task is not a first-month workflow.
- Stay near tasks with real consequences. The guidance is to “stop the task if something looks off,” which only works if someone is looking.
How to roll it out to the team
Cowork is on by default for Team plan organizations, and owners can disable it from organization settings. Rather than announcing it to everyone, pick two people and two of the workflows above. Give each workflow a name, an owner, and a place its output lands. Run them manually for two weeks, then schedule the one that worked. At 30 days, look at three things: which outputs people actually opened, which approvals they clicked through without reading, and what they started asking for that you hadn't planned. That last list is your month-two roadmap, and it will be better than anything you could have written on day one.
If you'd like a hand choosing the first two workflows, or someone to run that 30-day review with you, that's the work Cedar & Loam does with agencies and small firms. Proudly based in Kansas City. Say hello.
Sources
- Get started with Claude Cowork, Anthropic Help Center
- Schedule recurring tasks in Claude Cowork, Anthropic Help Center
- Use Claude Cowork safely, Anthropic Help Center
- Use Claude Cowork on Team and Enterprise plans, Anthropic Help Center