What it does
Give a coworker a recurring job, like a morning inbox triage or a weekly report, and it starts the work on time without a prompt. The result is waiting when you get in. Schedules live on the coworker’s Schedules tab. Under them, the same tab lists the triggers that hand it work when something happens.Setting one up
1
Pick how often
Every day, Weekdays, or Weekly. Weekly asks which day.
2
Pick the time
In your own timezone, which is shown next to the field so there’s no ambiguity.
3
Write what to do
What your coworker should do each time. The field’s own example is a good model:
Triage the support inbox and draft replies for anything urgent.

A weekday triage. The timezone is shown on the row, so there's never a question about which 08:30 it means.
Writing a good scheduled job
A scheduled job runs without you there to clarify, so it has to stand on its own. Write it as a complete instruction. “The usual report” means nothing at 7am on a Tuesday with no conversation behind it. Say what to do when there’s nothing to do. A quiet inbox should produce “nothing needing you today”, not a report padded out to look busy. Say where the result goes. Otherwise it sits in a session you have to remember to open. Pair it with Talking to people or a Slack connection and it can send you the summary directly.When you’d use one
- A morning briefing on what needs you today
- A weekly report assembled from the same sources each Monday
- A nightly check that something ran, with a message only if it didn’t
- A daily scrape of a portal that has no other way out
- A Friday summary of the week’s numbers
- A recurring reminder that does the preparation, not just the reminding
Triggers that hand it work
Some work shouldn’t wait for a timetable. A trigger hands your coworker work the moment something happens: a form posts to a web address, an email arrives, a file lands in a folder. You add one on the coworker’s agent. Open Settings → Triggers, click Add a trigger, pick what starts it, and under What it runs pick the coworker. Then say what it should do with each event. See Hand it to a coworker. The Schedules tab lists these triggers under your schedules, in a Triggers section:Each one hands Ana work when something happens.
- Each row shows what starts it, when it last fired, and whether it’s listening.
- Change it opens the trigger on the agent. Change what your coworker does with each event there, switch the trigger off, or remove it.
- Add a trigger opens the agent’s Add a trigger with this coworker already picked. With no triggers yet, the section says “No triggers hand Ana work yet.”
In each environment
Your coworker’s agent can have more than one environment, such as Production with your real data and Demo with sample data. Each environment then has its own schedules. A schedule belongs to one environment, and fires only while the agent runs in it. After a switch to Demo, your Production schedules stay quiet and only Demo’s fire.- The tab shows one environment at a time. Pick it under Environment: the one the agent runs in comes first. A schedule of another environment says Off in Demo beside its time while the agent runs in Demo.
- A new schedule goes into the environment shown: “It is made in Demo: it runs while the agent runs in Demo, as you. Copy it to another environment from its menu.”
- Copy to Demo, in a schedule’s menu, copies it to an environment that has none of it yet. The copy fires as you, so it asks first:
The copy runs as you: each run is a task under your name, at its own times, while the agent runs in Demo. Changing it never changes this one.The copy is separate: change its time or what it does, and the original stays as it was. If you can’t change the coworker, the copy keeps its owner and starts switched off. Someone who can change the coworker may switch it on. A coworker can have up to 10 schedules in each environment. Deleting an environment deletes its schedules. Moving the coworker to another agent takes only its Production schedules; the others are deleted, and the move window says so first.
Schedules, self-scheduling, or keep working?
Three different things, easily confused.Good to know
Each run is its own session, so you get a clean history rather than one conversation growing forever.
A coworker can pause its own schedule if the session can’t be done as written — and it tells you why on the schedule row, so a broken schedule doesn’t fail silently every morning.
Coworkers don’t run dynamic workflows for now, so a schedule can’t run a saved one. A workflow that Make repeatable created earlier still hands the job to its coworker, which can’t run it for now.
Triggers that hand a coworker work listen only while its agent runs, too. Retiring the coworker switches them off, and bringing it back leaves them off until you switch them on again.
Common questions
What timezone does it use?
What timezone does it use?
Yours — the one shown when you set the schedule. Editing from somewhere else re-stamps it to where you are then, so check the field if you travel.
Can one coworker have several schedules?
Can one coworker have several schedules?
Yes. A morning triage and a Friday report on the same coworker is a normal setup.
What if a run fails?
What if a run fails?
The schedule keeps its place and the row shows what went wrong. One bad morning doesn’t stop tomorrow’s.
Can I see what a scheduled run did?
Can I see what a scheduled run did?
Yes — each run is a session you can open and read in full, like any other.
How do I stop one temporarily?
How do I stop one temporarily?
The switch on its row. It stops running and keeps its settings for when you switch it back on. To stop every schedule and channel of an agent at once, click Stop in the agent’s header.
Next steps
Autonomy overview
Everything that runs without you.
Self-scheduling
When the coworker decides instead.
Talking to people
Getting the result sent to you.
Managing your tasks
What a schedule produced.
