SCHEDULING

What to Check Before You Publish the Schedule, and What Publishing Does Not Do

Seven things are worth checking before a week goes out, and the last one is the one almost nobody runs. It is also the one that gets blamed on the software afterwards.

What does pressing Publish actually do?

It makes the week visible and it tells nobody. That is true of XShift AI as well, and it is worth saying on our own page, because the assumption costs people a Monday.

A shift still short of a person stays a draft. So generate a month of 44 shifts, watch 41 come back fully staffed, and the button lights up with 41 while three stay behind.

Before you press it there is a preview: every draft grouped by date, who is on it, a tick on the ready ones and a warning on the short ones.

What are the seven checks to run first?

Seven, and only the first one has a deadline you cannot recover from. Run them in this order, because they get cheaper to fix as you go down.

Checks two through six each have a screen behind them in XShift AI. One and seven are yours, and seven is the one that gets forgotten.

Check one is law in some places, under predictive scheduling and fair workweek rules, and courtesy everywhere else. Oregon requires 14 calendar days in writing from large retail, hospitality and food service employers, set out by the Bureau of Labor and Industries.

Seattle and Chicago also land on 14 days for covered employers, while New York City retail is 72 hours. Your own city decides, and none of this is legal advice.

Where a late change triggers a payment, the Labor Department’s Fact Sheet #56B covers how that payment is treated in the regular rate.

Check three has its own screen. A red panel names everyone whose scheduled hours cross 40 in a single week and by how many hours, biggest week first, and marks the exempt people separately rather than quoting hours over a limit.

I would never publish on a Thursday for a Monday. Two working days is not notice, it is an announcement, and you will spend both of them fielding swaps.

How do you find out why a shift is still short?

Read the reason next to each name rather than guessing at it. By hand this means opening three screens per empty slot and checking each person against each of them.

XShift AI writes the reasoning down while the run is happening, so the record of that run is already waiting. It opens with a plain sentence, then the shift, then the people.

Chloe is named first because she is the smallest change away from being allowed. Four hours of headroom fixes that shift; nothing you can do today fixes Ben’s leave.

I open that panel before I look at the calendar. A calendar tells me a shift is empty. The panel tells me who is four hours away from filling it.

It also catches the two silent ones from check four. People with no location at all, who can never be scheduled anywhere, and people who were given shifts but never entered any availability.

The first of those can be fixed inside the panel: tick the people, pick a site, run it again. It keeps the last few runs too, so two attempts can be put side by side.

How do you actually tell the team it is out?

Say it yourself, in a message, the same way every week. Nothing about making a schedule visible is the same as putting it in front of somebody.

XShift AI will email your team when Auto-Assign All fills the open shifts. It will not do it because you pressed Publish.

The middle one is the automatic path. Every active employee with an upcoming shift gets an email naming the week and how many shifts they have, and the managers get a summary of their own.

Publish in the morning, never at the end of a shift. A schedule that lands at nine at night gets read by four people and argued about by everybody else the next day.

Two pages sit next to this one: a guide to communicating the schedule to your team and the 19 checks a staff schedule has to survive.

What else do managers ask before publishing?