30 Nov How AI can predict the employees who are about to quit
Artificial intelligence (AI) is now being used to predict how likely it is that potential hires and employees will stay with a company. Read this blog post to learn more.
Tim Reilly had a problem: Employees at Benchmark’s senior living facilities kept quitting.
Reilly, vice president of human resources at Benchmark, a Massachusetts-based assisted living facility provider with employees throughout the Northeast, was consistently frustrated with the number of employees that were leaving their jobs. Staff turnover was climbing toward 50%, and after many approaches to improve retention, Benchmark turned to Arena, a platform that uses artificial intelligence to predict how likely it is that an employee will stay in their job.
“Our new vision is about human connection,” he says. “With a turnover rate that’s double digits, how do you really transform lives or have that major impact and human connection with people who are changing rapidly?”
Since Benchmark started using Arena, staff turnover has fallen 10%, compared to the same time last year. During the hiring process, Arena looks at third-party data, like labor market statistics, combined with applicants’ resume information and an employee assessment that will give them a better sense of how long a candidate is likely to stay in a role.
“The core problem we’re solving is that individuals aren’t always great at hiring,” says Michael Rosenbaum, chairman of Arena. “Job applicants don’t always know where they’re likely to be happiest. By using the predictive power of data, we’re essentially helping to answer that question.”
Arena isn’t interested in how an employee responds to assessment questions, he says. They’re much more interested in how employees approach the questions.
“What you’re really doing is your collecting some information about how people react to stress,” Rosenbaum adds.
For example, if an employee is applying for a housekeeping role, Arena may give them a timed advanced math question to complete — something they may never use in their actual job. Arena then studies how the candidate responds to the question — analyzing key strokes and tracking how the individual tackles the challenge. The software can then get a better sense of how an applicant responds under pressure.
Overtime, Arena’s algorithm learns from the data it collects. The system tracks how long a specific employee stays at the company and can then better predict, moving forward, whether other employees with similar characteristics will stay.
“Overtime they are able to sort of refine that prediction about those that are most likely to stay, or be retained with our organization,” Reilly says. “They may also make a prediction on someone who might not last very long.”
Reilly says he’s been encouraging hiring managers at the facilities to use the data given to them by Arena to take a closer look at the candidates the platform rates as highly likely to stay in their roles. Although it’s ultimately up to the hiring manager who they select.
“Focus your time on the [candidates] that are more likely to stay with us longer,” Reilly says.
For now, Arena exclusively works with healthcare companies. The platform is currently being used by companies like Sunrise Senior Living and the Mount Sinai Health System in New York. Moving forward, Rosenbaum says, they’re hoping to get into other industries, although he would not specify which.
Rosenbaum says Arena is not only focused on improving the quality of life for employees, but also for the patients and seniors that use the facilities. The happiness of patients, he says, is closely tied to those that are caring for them.
“Is someone who is in a senior living community happy? Do they have a positive experience? It is very closely related to who’s caring for them, who’s supporting them,” he says.
SOURCE: Hroncich, C. (15 November 2018) “How AI can predict the employees who are about to quit” (Web Blog Post). Retrieved from: https://www.employeebenefitadviser.com/news/how-ai-can-predict-the-employees-who-are-about-to-quit?brief=00000152-1443-d1cc-a5fa-7cfba3c60000