Here’s a New Way to Verify HR Software’s Capabilities
The HR Policy Association has launched a new initiative it hopes will clear up much of the confusion among HR leaders about just what, exactly, all the new recruiting software and tools that have been introduced recently can and can’t do.
The intent was to filter out the “noise” in a marketplace that’s been overwhelmed with new applications, says Mike McGuiness, HRPA’s director of talent initiatives.
“There’s a growing consensus among our members that there are a lot of vendors out there claiming to have AI capabilities,” he says.
The Recruiting Software Initiative features a review board comprised of TA leaders from 20 HRPA member companies. Formed at the beginning of this year, the review board reached out to more than 100 software vendors and asked them to complete a detailed questionnaire. The questionnaire was assembled by a committee chaired by Target’s CHRO, Stephanie Lundquist.
Thirty-two of the vendors agreed to participate in the process. Of that number, 15 were ultimately awarded one or more “badges of excellence” by the RSI for innovative approaches for identifying, attracting, evaluating and retaining talent. Each vendor’s products were evaluated based on ability to adapt, mitigate unconscious bias, comply with employment regulations, ease of integration and ability to justify outcomes, says McGuiness.
The badges do not represent an endorsement by HRPA of each vendor, he says. Instead, they’re intended to validate that the vendor’s product appears to actually do what the vendor says it will do.
The RSI has created a database with information about each vendor that participated, including feedback from members of the committee that evaluated their wares. Although the database is available only to HRPA members, the organization has made public the list of vendors that received badges of excellence (see list below).
Some of the most important takeaways by the RSI’s review committee included that some vendors are stretching the definition of “AI” just a bit, says McGuiness.
Generally, to be considered “AI capable” a machine or software tool must have the ability to process large amounts of data and learn from or recognize patterns in that data. However, the definition of AI has grown fuzzier in recent years as the number of vendors touting AI-based products has grown exponentially.
“A lot of what we saw being advertised as ‘AI’ was really intelligent automation,” he says. “It’s the same sort of matching and algorithms you see being used in apps like Spotify and Waze.”