Can AI really improve the lives -- and perhaps the staying power -- of hourly workers, or is the technology not quite there yet?
The question is timely, as HCM platform-maker Kronos Inc. just rolled out an AI-powered tool aimed at the hourly workforce. To put it all in context, workers paid by the hour represented nearly 59% of employed Americans in 2017, according to U.S. Bureau of Labor Statistics, and that number has grown 15% since 1996. And a 2015 study from the Society of Human Resource Management found that hourly workers' turnover rate was 49% annually.
In one sense, the Kronos AI announcement seems to be business as usual. In 2018, a slew of AI and machine learning-based tools arrived in the HR space. But almost universally, those products have been designed to improve the HR workload or are used by salaried employees. If AI had anything to do with the hourly workforce, my cynical thought was that it was more likely to be taking their jobs than actually helping them out.
Kronos AI tries to tackle shift swaps
That's why it was time to look deeper at how AI could make a difference in areas like employee engagement, starting with the Kronos AI tool, Aimee, which is an acronym for Artificial Intelligence for Managers and Employees. Jayson Saba, senior director of product marketing at Kronos, is, not surprisingly, excited about what AI can offer.
"A workplace that uses artificial intelligence is actually a pretty cool place to be," he wrote in an email.
But it was practicality and not the coolness factor Kronos was pursuing with Aimee's first big use case in the Kronos Workforce Dimensions suite: shift swapping.
"When an employee needs to swap a shift, Aimee looks back at the last nine or more months of shift swaps and learn from that history," Saba explained. "Who do you usually swap with? Who usually accepts those swaps? What days do you like to swap on? What days do other people not like to work? Aimee then makes a recommendation about which employees are most likely to want to swap."
Kronos also recognized that not all hourly workers are created equal. A nurse might have a hospital email address and easy access to a computer, while a fast food employee might not. That's why the company teamed up with Workplace by Facebook and IBM's Watson supercomputer to offer chatbot-powered shift swapping on any mobile device, as well as coaching or training options, Saba said.
And, like several other competing products, Aimee offers predictive features that can help managers identify which employees might be most likely to leave.
Hourly scheduling is 'hellishly hard'
Putting Kronos AI to work with scheduling issues is largely a winning combination for HR analyst Katherine Jones, who gives Kronos points for tackling the hourly workforce.
"That's the most difficult part of the market," she said. "What makes Kronos AI and the product suite more interesting to me is they're looking at difficult business problems and trying to use smart technology to figure it out." Shift scheduling and swapping are "hellishly hard" problems to solve, she said, so it makes sense to use a technology like AI "that doesn't get tired or overwhelmed with too much data."
That all sounds good -- if you believe AI can really bring it all home. Brian Sommer, founder of analyst firm TechVentive, was not familiar with the new Kronos AI announcement, but he has looked hard at AI in HR and is skeptical -- at least at this point.
Trevor Whiteanalyst, Nucleus Research
"We're really, really in the early stages of AI and ML [machine learning] in this space," he said. Describing AI as a shiny toy most companies lack the expertise or staff to make work, Sommer doesn't buy the idea that AI can predict turnover, which is a particularly tough problem to solve when it comes to hourly workers.
"We're relying on the correlations where we've seen the pattern before," he said. "An employee has used up all vacation time or starts having more erratic time clock activity or has cashed in stock options. The problem is there is no agreement among firms about what are the correct indicators and no two firms use the same indicators to calculate the [flight risk] score."
Although Jones vehemently disagrees with this perspective -- "It's the easiest thing ever to create an algorithm to predict flight risk" -- she does think the debate is meaningless for most hourly workers. "They'll change jobs right away if they can get [a small raise] somewhere else," she said.
Hourly worker priority: Steady schedule and pay
Trevor White, an analyst at Nucleus Research, goes even further. He thinks the Kronos AI platform is looking to solve a problem that simply does not exist.
"I'm not sure how much the stuff around coaching is really going to make a difference," he said.
Nucleus has interviewed hourly workers, he said, and found their priorities are a fixed schedule and a steady paycheck. Anything outside of that doesn't matter, he asserted.
"It's not like they're going to go on their breaks and get on a mobile phone for a coaching session."
Hiring and employee retention are hard in every single segment of the market today, but for companies fuelled by hourly workers, the challenges are greater and the risks of failure are no doubt higher.
AI feels like a no-brainer for shift swaps and scheduling, but like some of the experts, I need more convincing about how coaching on the go can improve an hourly worker's loyalty or job satisfaction. And I'm still wondering how many hourly jobs AI-powered robots might eventually replace.