In Alberta, overtime is all hours worked over 8 hours a day or 44 hours a week, whichever is greater, and it must be paid at at least 1.5 times the employee’s regular wage rate unless a written agreement says otherwise (Government of Alberta). The rule has two tests, and the larger one wins. A manager building next week’s schedule from memory is checking neither.

How overtime sneaks in

It rarely comes from a big decision. It comes from a covered sick call on Thursday, a shift swap that nobody re-checked, and a Saturday add-on that pushes someone past 44. Each change is reasonable on its own. Nobody sees the total until payroll, when the cost is already spent.

Rules first, then the schedule

The principle: check the rule when the schedule is built, not when payroll runs.

Scheduling software earns its place by encoding your constraints as rules and checking every change against them before anyone sees the schedule:

• The daily and weekly overtime tests for your jurisdiction.
• Skills and certifications required on each shift.
• Availability, time-off requests, and minimum rest between shifts.
• A labour budget per day, with a warning when a change would breach it.

Tools such as Deputy and When I Work do this kind of rule-based scheduling. Automation generates and validates the schedule and flags only the exceptions. Gen AI is useful for the human layer: explaining why a shift changed, suggesting who could cover a gap, and drafting the swap conversation.

Check the numbers you are quoted

Vendors publish impressive overtime reductions. Case studies I could find report anywhere from roughly 20% to 40% fewer overtime costs, but they are vendor-published and vary widely by business, so do not plan around any single figure. Plan around your own baseline.

A one-week test

1. Pull last month’s schedules and payroll. Count the overtime hours and note which weeks they fell in.
2. For each overtime hour, ask whether a rule check at scheduling time would have caught it.
3. Pick one team and run the next four weeks with rule checks on every change.
4. Compare overtime hours, schedule-change count, and how many shifts had to be re-covered at short notice.

Boundaries

Automation owns the rules, the calculation, and the audit trail. Gen AI owns the communication. Humans own exceptions, approvals, and anything an employment standards officer would ask about. Confirm the thresholds for your own industry, because some sectors have different rules or averaging arrangements.

Schedules built on memory cost you overtime. How many of last month’s overtime hours were a decision, and how many were an accident?