New Research: Is AI Better Than Humans at Employee Surveys?

November 11, 2019

New Research: Is AI Better Than Humans at Employee Surveys?

In the past few years I have been digging very deep into AI and how it can help (and hurt) the employee experience. What I’ve been itching to do, however, is create an experiment that demonstrates just how powerful AI can be when paired with essential human skills in the workplace.

An Experimental Approach

Augmented Listening ReportThis year I had the opportunity to create an experiment to see how AI can help with employee survey, or voice of the employee, efforts. What I was curious to know is whether AI is better than humans at identifying trends, emotions, and other facets in an employee feedback dataset. Yes, I think about these things because I’m a researcher, but also because I’m highly practical.

When I worked as an HR executive I simultaneously loved and hated when it came time for our annual employee engagement survey. I was excited to see how everyone felt, what their needs were, and how we were doing as an organization. I was terrified at the prospect of combing through mountains of data to make sense of it all, knowing that some people would be left out simply because we didn’t have the time to dig into every single facet of the responses.

What Can AI Do to Help?

Fast forward to today, and AI is being used across the business and across industries to help with everything from transportation to medicine. Surely there are applications within HR as a function of the business world (hint: I wrote about tons of these in the book). What role does it have in supporting these employee engagement surveys that are a core part of how we understand and support the workforce? Is it better than humans? If so, what does that mean for the work we need to be doing?\

That led to an experiment where we let humans and AI analyze some data and then we compared the results. It was pretty astounding, even for me.

The resulting research report (20+ pages) is chock full of details, including how we designed the experiment, what the results looked like, and what practical lessons employers need to consider both now and in the future.

Thank you to Ultimate Software for sponsoring this research and partnering with me on the analysis.