Increasing new hire diversity is (or should be) at the top of every TA professional’s mind. Inclusive companies generate 30% more revenue than those that are not, and 69% of executives surveyed in Deloitte’s 2017 Human Capital Trends Report rated diversity and inclusion as an “important” issue.
From powering the decisions of self-driving cars to identifying cancer, software with an artificial intelligence (AI) element is great for increasing efficiency, cost savings, and helping users make more data-driven, consistent decisions. And with more and more recruitment technologies incorporating AI into their solutions, it is important to ensure there is no adverse impact on diverse groups.
What Happens When an AI Makes Screening Decisions?
When Unilever implemented a brand new, AI-driven hiring process for their Future Leaders graduate program, they experienced the same increased efficiency and cost savings you would expect from a sophisticated algorithm.
Compared to their previous process, Unilever reduced time to hire by nearly 90% (from four months to two weeks - in some cases, just 2 days), saved over £1,000,000, and identified the top 800 individuals from a quarter million applications.
This was their new four step process:
- Apply via a simple application (can be filled out from a LinkedIn profile)
- Assess candidates' cognitive traits with games (AI-powered screening step, provided by Pymetrics)
- Assess candidates' soft skills and potential with a recorded video interview (AI-powered screening step, provided by HireVue)
- “A Day in the Life of Unilever” job simulation and onsite assessment
The first three steps of this process were entirely automated, and the screening decisions in the middle two steps were made entirely by artificial intelligence - removing human bias from the equation.
Candidates who were identified as having the most optimal cognitive traits in the second step were automatically invited to a HireVue interview. Candidates who HireVue’s assessment identified as potential top performers were forwarded to a Unilever recruiter, where they were provided with next steps.
With so much of the screening fully automated, it was critical that a full adverse impact analysis was performed before the hiring process went live.
Quick Feedback: An Experience Candidates Want
Since applicants were provided with immediate feedback on their candidacy following each assessment step, they were not left wondering for weeks (or months) if they made it to the next step. They appreciated the speedy decision-making, and of the 250,000 applicants, over 80% gave the process positive feedback.
Not only did Unilever’s AI-driven hiring process not have adverse impact on diverse groups, they created their most diverse class to date (both gender and ethnically): diverse hires were increased by 16%. For an international organization with a pre-existing focus on diversity initiatives, that’s an impressive number. By identifying the traits crucial for on-the-job success, artificial intelligence took unconscious bias out of the equation and recommended the best people for the job.