When a day care worker told Journelle Clark that her five-year-old daughter often cried because she missed her mother, Clark realized something was really wrong. The 32-year-old works unpredictable hours arranging window and wall displays at an Ann Taylor LOFT in Manhattan, never knowing what her schedule—or her paycheck—will be. She could get 15 hours one week and 40 the next. Sometimes she works all weekend, sometimes she doesn’t. She’s been called in to work at the last-minute and occasionally gets sent home early from a shift she’s already started.

“They expect me to be available all the time, 24/7,” Clark says.

Clark is part of a growing segment of the American workforce whose schedules aren’t designed by people but by sophisticated scheduling software programs. It used to be that retail and hospitality workers were scheduled for standard shifts whether business was busy or slow. Now these jobs have become increasingly tied to fluctuations in customer traffic. Using “workforce optimization systems” managers can ensure that stores have exactly the right number of workers, not only by the day, but by the hour.

The algorithms, built on data including last year’s sales trends, economic indicators, changes in weather and political pressures, are intended to create a perfectly predictable schedule and then provide real-time information for managers to make adjustments as these variables change.