As human resource leaders deal with the Great Resignation—waves of employees quitting and many more looking to leave their current jobs—HR leaders and technology experts say that data issues hinder artificial intelligence’s ability to predict employee flight risk accurately right now. That said, they are optimistic that AI will play a role in employee retention.
To identify potential flight risks at Citrix, Donna Kimmel, executive vice president and chief people officer for the web infrastructure and networking solution provider, and her team gather data from Citrix’s 10,000 worldwide employees to gauge their job satisfaction and employee sentiment and measure that information against business intelligence data it has culled over the past five years. AI doesn’t yet play a role in the essential task.
Calling her data “really good and robust,” Kimmel says her firm’s next step is collecting and analyzing the data for employee turnover. But ultimately, it’s the verbatim responses to the monthly surveys that tell the real picture—and not AI.
That’s why Kimmel and other experts say AI still has a ways to go before it’s able to accurately—and on its own—predict employee flight risk. She has yet to see an artificial intelligence tool that can adequately predict when an employee will leave but calls that “a wishlist area” for her.
For now, when an employee submits their resignation, getting a sense of why largely comes down to one-on-one conversations between an HR business partner and the employee. “I would love to be able to do more in the AI space around resignations or are they leaving for [reasons that can be] prevented,” Kimmel says.
The devil is in the data
Although there are AI-powered HR tools that measure employee sentiment, some HR tech experts think something is missing. It all comes down to the data, says Jason Averbook, CEO and co-founder of HR consultancy Leapgen.
“Artificial intelligence, which I like to call ‘augmented intelligence’ because I don’t really think it’s artificial, is basically a tool that can be used if an organization has good data,” says Averbook, a Top 100 HR Tech Influencer who will be speaking Oc.t 1 at the HR Tech Conference in Las Vegas. “Organizations call us every day and say, ‘We need help with AI.’ The demo [of a potential AI tool] looks great and [then we ask], ‘What data do you have to feed AI?’
Related: Learn more from Jason Averbook at HR Tech. Register here.
“Because AI without data is nothing, right?” Averbook adds. “It’s truly an empty closet—and most organizations don’t have data.”
Instead, Averbook’s firm and his clients perform weekly pulse checks of employees, asking how they are feeling with a choice of three colors: green meaning “great,” yellow for “not so good” and red being “not good at all.” If a cluster of yellow or red responses appear in a specific location or division, this might mean a management or leadership concern.
“If I can identify patterns like 18 people with a specific leader are red or a pattern that says people in the 40- to 50-year age group are suddenly asking questions on the HR portal about leave … those are the types of data points that AI can use to alert HR to potential opportunities,” says Averbook.
“I hate to say it, but most people don’t have that data.”
Another risk with AI is the inherent concern about the accuracy and bias of such tools.
When it comes to measuring employee flight risk, George LaRocque, founder of WorkTech, believes AI may not be able to differentiate between an employee who is genuinely dissatisfied with their work and another employee dealing with issues outside their job.
“I’m really concerned about flight risk indicators because an employee who takes a bunch of vacation days and might’ve had a [poor] performance review would jump out analytically to an AI algorithm,” LaRocque says. “Employee A may be someone taking time to interview for [another job] but then Employee B could have all those same [red flags] and could be going through a grief process or dealing with a loss in their family or a parent who’s elderly and being the primary care person.”
AI’s role in empathy work
AI does have its place, of course. According to Greg Pryor, senior vice president and people and performance evangelist at Workday, AI has a role to play in gathering employee sentiment and being able to address their needs.
“I think we’re at the beginning of organizations ultimately [asking] ‘How do we automate and augment so that we can elevate essential human capabilities?’ ” says Pryor. He adds that machine learning and gathering data from multiple sources enable individual leaders to do what he calls “empathy work.”
For example, it can help leaders learn “this is what’s important to our teams, here’s where we may be falling short of those expectations [of empathetic leadership] so we can become really focused on what’s important to our people,” he says. “So, I think there will be increasingly a use of [AI tools and ] machine learning in that employee experience journey.”
In the meantime, companies like Citrix will continue to look at their internal data. The company partnered with Culture Amp, a provider of an employee sentiment solution, to conduct annual and mid-year global surveys of its employees. Culture Amp then parses the survey results and presents Kimmel and her team with analysis pertaining to factors such as the ages and location of their workers. For now, Kimmel will continue to pay close attention to the verbatim responses in their monthly employee surveys.
“There’s a lot of really rich data in the verbatims that we read and when you read those verbatims, that’s where you hear the voice. Using AI around the verbatims would be something fabulous,” says Kimmel. “That would be incredibly helpful.”
Learn more about AI in HR and more at the HR Tech Conference, Sept. 28-Oct. 1 in Las Vegas. Click here for more information and to register.