Can AI Help Companies Struggling with the Great Resignation?

As organizations face workforce gaps and talent losses, Pymetrics sees its AI platform as an alternative to traditional HR hiring practices. Using AI can help lessen human bias and identify quality candidates, hire equitably, reduce turnover rate, and enhance performance. Pymetrics applies AI-driven assessments to help companies build diverse, top-performing teams.

Image credit: Pymetrics

Since the onset of the pandemic, nearly 57 million American workers left their jobs between January 2021 and February 2022. Some are finding work with a different employer, and others are opting out of the workforce completely. Coined “the Great Resignation,” the number of workers leaving is still on the increase. According to a PwC report, one in five workers globally plan to quit in 2022.

With so much employment activity, employers are turning to AI to help them fill open positions. More importantly, AI is helping find employees that are more likely to stay, that fill diversity gaps, and share traits and characteristics of a company’s current, successful leaders. Pymetrics is one of the standouts in this category for its use of neuroscience data and AI.

Positive Impact of AI in Hiring

Pymetrics gamifies the hiring process, asking candidates to take 12 tests that measure cognitive and emotional attributes. Even before candidates take the test, though, Pymetrics works with the hiring company to build a trait profile of the organization's top performers. Those traits help train the algorithms to select best-fit talent. The algorithms are also audited to remove any gender or ethnic bias.

More than one million job candidates around the world have played Pymetrics' games, and the company has helped more than 100 clients hire a workforce using predictive analytics and AI. At ANZ, a multinational bank, Pymetrics helped attract and recruit a more diverse set of college graduates.

Pymetrics first worked with ANZ to build a model of performance and values that are important to the organization. Applicants were invited to play the Pymetrics games, with candidate names, gender, ethnicity, specific university, and self-reported disabilities excluded from the process.

Using Pymetrics, ANZ increased applications by 48 percent and realized a 99 percent completion rate. It also saw an uplift in diversity with offers going to universities outside the usual categories by 11 percent. On the candidate side, 95 percent were satisfied with the process.

Job performance and ethnicity graph comparing traditional hiring with Pymetrics.

Image credit: Pymetrics

Using Body Language Analysis

One of the biggest success stories for Pymetrics is Unilever. Candidates for Unilever positions first filled out a digital application with a choice of including a resume. Selected candidates then played the Pymetrics games that assessed cognition, problem solving, and risk-taking aptitudes. During interviews, candidates were assessed by video analyzing software using an ML algorithm.

Candidates’ facial expressions, body language, and word choices were captured as they were interviewed from computers or mobile phones. The machine learning algorithm examined the responses and determined which candidates were a good fit through natural language processing and body language analysis.

In only 90 days after adopting the new hiring strategy, Unilever doubled the amount of application to jobs success rate moving from 15,000 to 30,000. The hiring managers also hired their most diverse candidate pool by focusing on hiring non-white applicants and increasing representation to 2,600 universities, as opposed to 840 universities in the years prior. Additionally, Unilever achieved greater gender parity in their hiring, as well as increasing socioeconomic representation by 20 percent.

Hiring managers saw benefits too. The average hiring time decreased from four months to four weeks, saving almost 50,000 hours. Recruiters spent 75 percent less time reviewing applications, and the acceptance rate of offers to candidates went from 64 percent to 82 percent.