Work that repeats should not depend on someone remembering it
Opening checks, weekly reports, monthly reconciliations, annual renewals. Set the cadence once and TaskIt creates the task each time it is due, assigns it, notifies the right person and records that it ran.
Included on every plan term — there is no separate automation tier.
The problem
Repeated work is the work that quietly stops happening
One-off tasks get chased because somebody is waiting for them. Recurring work has no requester, so when it is skipped nobody notices until the consequence arrives.
- The reminder lives in one person’s phone, and they are on leave.
- A shared checklist is copied by hand each week, so half the weeks have a slightly different list.
- Nobody can prove the Tuesday check was done, only that nobody complained.
- The person who has always done it leaves, and it takes two months for anyone to notice it stopped.
- A monthly job scheduled by "the 1st" quietly moves to the 3rd, then the 5th, then whenever someone remembers.
How TaskIt handles it
A schedule that produces ordinary tasks
A recurring schedule is a reusable task template plus a cadence. What it creates is a normal task — with the same lifecycle, assignment, handover, notification and audit rules as anything else.
- Daily, weekly, monthly, yearly and custom interval cadences, each with its own time zone, start date, optional end date and optional maximum occurrence count.
- Weekly rules can select weekdays; monthly and yearly rules can select a day of the month.
- The browser previews the next ten occurrences using the same recurrence engine the minute-level scheduler runs, so what you see is what will happen — daylight saving included.
- Pause, resume, skip one occurrence, duplicate a schedule, or cancel it. Cancelling never deletes the tasks it already created.
- A unique schedule-and-occurrence constraint plus a run ledger make generation idempotent: overlapping scheduler ticks cannot create the same task twice.
What you get
The routine survives the person
The schedule holds the cadence, the description, the assignee or department and the fields. Somebody leaving no longer takes a recurring process with them.
You can prove each occurrence ran
Schedule history records every attempted run, the task it produced, retries and error detail. "Did the Tuesday check happen?" is answered by looking, not by asking.
Generated tasks are real tasks
They can be handed over, blocked, reviewed and reported on like anything else — and each one links back to the schedule and occurrence number that produced it.
Cadence without a calendar argument
A time zone per schedule means a monthly job at 09:00 stays at 09:00 through a daylight-saving change, rather than drifting an hour and then back again.
How it works
In the order a person using it meets them.
- 1
Write the task once
Title, description, priority, form fields, and whether it goes to a named person or a department. This becomes the template every occurrence is created from.
- 2
Choose the cadence and the time zone
Daily, weekly with selected weekdays, monthly or yearly on a chosen day, or a custom interval stepping by days, weeks, months or years.
- 3
Check the preview before you commit
The next ten occurrences are computed by the same engine the scheduler uses. If the dates are not what you expected, the rule is wrong — not the preview.
- 4
Let it run, and adjust when reality changes
Pause it over a shutdown week, skip a single occurrence for a public holiday, duplicate it for a second site, or cancel it when the process ends.
A salon’s opening and closing routine
Two schedules replace a laminated checklist and a WhatsApp reminder.
- 1
Opening checks — daily, 09:00
Raised for the floor department in Single Owner mode. Whoever opens accepts it; the rest of the team can see it was done and by whom.
- 2
Closing checks — daily, 20:30
Same pattern, different fields: till reconciled, stations cleaned, stock noted, doors and shutters.
- 3
Stock count — weekly, Monday
Raised Collaborative, because two people split the shelves. Each logs their own progress.
- 4
Public holiday
The Monday occurrence is skipped once from the schedule. The series carries on unchanged the following week.
- 5
Six weeks later
The Recurring Execution Success Rate report shows which routines are being completed and which are being created and ignored — usually a more useful signal than the checklist itself.
The result
A missed closing check is now a visible overdue task with a name against it, rather than something discovered the next morning.
Cadences in full
Every named frequency implies its calendar unit; a custom interval names its own, so "every 3 weeks" and "every 18 months" are both expressible without inventing a fake weekly rule.
- Daily — every day, or every N days as a custom interval
- Weekly — one or more selected weekdays
- Monthly — a chosen day of the month
- Yearly — a chosen day, for renewals, licences and annual filings
- Custom — any interval stepping by day, week, month or year
Pause, resume, skip, duplicate, cancel
A schedule has a lifecycle of its own: active, paused, completed when it reaches its end date or maximum occurrences, cancelled as a permanent stop that keeps history, and failed when something needs a human. A duplicated schedule starts paused, so a copy made for editing cannot fire before you have finished editing it.
Execution history and provenance
Every attempted run is recorded with its outcome — pending, success, failed or skipped — the task it created when it created one, retry attempts and error detail. From the other direction, a generated task exposes which schedule and which occurrence number produced it, so a question about one task never requires searching the schedule list.
Why generation cannot duplicate
An occurrence row exists the moment an occurrence is claimed, and a unique constraint on the schedule-and-occurrence pair means a second tick attempting the same occurrence cannot write a second task. Two overlapping scheduler runs are safe by construction rather than by timing luck.
Frequently asked questions
Nothing. Cancelling or deleting a schedule never deletes the tasks it has already generated — those are ordinary tasks with their own history, and removing them would erase work that really happened.
Yes. A schedule can raise its tasks for a department in either Single Owner or Collaborative mode, so a daily routine goes to whoever is on shift rather than to one named person who might be on leave.
Yes — the browser preview calls the API and uses the same daylight-saving-aware recurrence engine as the minute-level scheduler. If the preview is wrong, the rule is wrong.
No. Each occurrence is claimed in a run ledger under a unique schedule-and-occurrence constraint, so overlapping scheduler ticks cannot produce duplicate tasks.
Where to go next
The capabilities either side of this one, and the problems it is bought to solve.
Schedule the routine you keep forgetting
Pick the check that gets missed most often, give it a cadence, and look at the execution history in a fortnight.