Claude Scheduled Tasks Explained: Set-and-Forget Workflows That Run Without You

Article author portrait

By

Jan Meinecke

9

Min

Article author portrait

By

jan Meinecke

9

Min

Most of the value I get from AI now isn't from asking it things. It's from jobs I set up once and never touch again. Every weekday morning a summary of what moved overnight is waiting for me. Every Monday a report is compiled before I open my laptop. I didn't run any of it. Claude did, on a schedule, in a session I was never part of.

That's what scheduled tasks are, and they're the quiet shift from AI as a thing you chat with to AI as a thing that works for you in the background. Here's what they are, how they run, how to set one up, and the jobs actually worth handing over.

What Claude scheduled tasks are

A scheduled task is a job you hand to Claude that runs automatically on a schedule, or on demand, without you starting it. It's a newer layer on top of Cowork, the desktop agent Anthropic launched to work across your files without any coding. Anthropic's own description is exactly this: Scheduled tasks "allow you to delegate work to Claude Cowork by creating tasks that run automatically on a recurring basis, or on demand."

The important word is delegate. A normal Claude session is a conversation: You're there, you steer, you read the reply. A scheduled task is closer to handing a colleague a standing instruction. You write the job once, you say when it should run, and from then on it happens on its own. The feature lives in Claude Cowork, Anthropic's agent that works across your files and tools, and it's available on the Pro, Max, Team and Enterprise plans.

How they actually run

This is the part that makes them useful rather than just convenient. Each scheduled task runs remotely as its own Cowork session, on Anthropic's infrastructure, not on your machine. Anthropic puts it plainly: "Each scheduled task runs as its own Cowork session. You can review the results when they're ready, just like any other task."

Because it runs in the cloud, it doesn't need you or your computer to be awake. The job fires at the time you set, does its work, and leaves the finished result waiting for you to review, the same way any other Cowork task does. You come back to an output, not a request to do something.

Every Cowork task follows the same loop: You brief it, it runs, you verify the result

The schedules on offer are the ones most real work follows: Hourly, daily, weekly, weekdays, or manual for the jobs you want to trigger by hand. So "every weekday at 7am" or "every Monday morning" or "on the first of the month" are all straightforward to set.

How to set one up

There are two ways in, and both start from the Scheduled section in the Cowork sidebar.

The first is to let Claude build it with you. Choose New task, then Create with Claude, and answer its questions about what you want, when, and how. It assembles the task and shows it to you to confirm before anything is scheduled. This is the one I reach for when I'm still working out exactly what I want the job to do.

The second is to set it up manually. Choose New task, then Set up manually, and fill in the details yourself: A name, the prompt Claude should run, the approval mode, how often it runs, and optionally which model to use. Save it, and it's live. This is the faster route once you know exactly what the job is.

Writing a prompt that survives on its own

The prompt is where scheduled tasks are won or lost, because there's no one there to answer a follow-up. A live chat forgives a vague ask, you just clarify when Claude checks in. A scheduled run doesn't get that chance: Whatever you wrote is all it has, every time it fires. So spell out the source, the shape and the judgment calls. "Check Slack" leaves all three open. "Summarise the growth channel from the last 24 hours, group it by topic, pull out anything that names a customer, and keep it under 200 words" answers every one, and it'll answer them the same way on a quiet Tuesday and a chaotic Friday. A good test: If a new colleague couldn't run the task from your prompt alone, Claude can't either.

What to actually run on them

The jobs worth scheduling share a shape: They're valuable, they repeat on a rhythm, and they don't need your judgment in the moment. A few that have earned their place for me and the teams we work with:

A morning briefing. Pull together the overnight email, the calendar for the day and the key channel activity into one summary, ready before you start.

A weekly report. Compile the numbers and the notable changes into the same format every Monday, so the report writes itself and you edit rather than assemble.

Recurring research. Track a competitor, a topic or a search term on a weekly cadence, so you notice movement without going looking for it.

Housekeeping. Sort and tidy files, or clear out and label the week's incoming documents, on a schedule.

The common thread is that each one used to be a small recurring tax on your attention, and now it's a thing that simply happens. If you want the specific overnight version of this, we wrote a companion piece on how to schedule Claude to work while you sleep.

When a scheduled task isn't the right tool

Scheduling is powerful, which makes it tempting to schedule everything. Some jobs are better left as a normal session. A one-off, however useful, is a task you run now, not something to put on a cadence. Anything that needs your judgment partway through, a decision only you can make or a reply only you should send, wants you in the loop, so schedule the preparation and keep the decision. And anything where being a few hours stale would cause a problem, a same-minute alert, say, is a job for a proper integration, not a daily run. The sweet spot is recurring, low-judgment work that's just as useful an hour later as the moment it finished.

One limit worth knowing

Because scheduled tasks run in the cloud rather than on your computer, they work with your cloud-connected tools and the files in your Claude account, but they can't be tied to a folder on your own machine. So a task that reads your CRM through a connector or works on documents in your account is fine. A task that needs a specific folder on your laptop isn't the right fit. Plan the job around what lives in the cloud.

If you're still finding your feet with Cowork itself, our practical guide to using Claude Cowork covers the basics that scheduled tasks build on.

Scheduling is one of the biggest wins you can set up, and setting it up well is exactly what we teach. Claude Cowork for Growth is our five-week live cohort for people in Growth, Marketing, Sales and RevOps who want to be the one on their team who actually makes AI work, not just talks about it. It's practical rather than theoretical: You leave with a real growth workflow you've automated and will keep using, plus an Open Badge to show for it. If that's the year you're trying to have, reserve your seat.

Common questions

What are Claude scheduled tasks? They're jobs you hand to Claude Cowork that run automatically on a schedule, or on demand, without you starting them. Each one runs as its own Cowork session and leaves the finished result for you to review.

Do scheduled tasks run when my computer is off? Yes. Each task runs remotely as its own Cowork session on Anthropic's infrastructure, so it fires and completes whether or not your computer is on or awake.

What schedules can I use? Hourly, daily, weekly, weekdays, or manual for tasks you trigger on demand. That covers most recurring work, like a weekday morning briefing or a Monday report.

Which plans have scheduled tasks? They're available on the Pro, Max, Team and Enterprise plans, with full support on desktop and web and mobile rolling out.

Can a scheduled task use files on my computer? No. Scheduled tasks run in the cloud, so they work with cloud-connected tools and the files in your Claude account, but they can't be tied to a folder on your own machine.

read next


Prompt engineering vs context engineering

Context Engineering vs Prompt Engineering, Explained

Context engineering vs prompt engineering: What each is, why context beats wording once AI does real work, and the moves that make agents reliable.

How to Use ChatGPT Work: Access, Setup and Your First Task

How to use ChatGPT Work: Getting the desktop app, choosing local or cloud mode, setting up a project, and briefing your first whole job for it to run.