HR AI is here — the panacea to solve all your recruiting and talent management challenges!
That’s what the HR AI hype seems to be saying. The reality is this: Even as straight-shooting vendor reps and analysts expound on the advantages that AI brings to HR technology, a number of them caution that the idea of incorporating true AI into HR software is a long-term proposition.
In fact, many professionals believe the state of AI in HR should be viewed in context. Today, they contend, machine learning and other areas of AI are primarily used for advanced data analysis and automation. In the long term, they have no doubt AI will have an enormous impact on human capital management (HCM) but predict its evolution will be both lengthy and complicated. “What we see now is what’s possible,” said John Sumser, principal analyst at HRExaminer and a widely known observer of HR technology. He made the remark during a conversation at the 2018 HR Technology Conference and Exposition, where everyone from big vendors, like ADP, to more specialized firms, like time-and-attendance tool developer BioGrp, were touting the integration of AI into their products.
But Sumser also said this about AI’s effect on HR: “The impact is going to be big. You can’t avoid it, and you better get prepared.”
So, although the HR AI hype is real, the promise is too. The trick is balancing the two.
Here are four ways to do just that.
1. Understand HR AI hype vs. today’s reality
Currently, AI in HR focuses on implementing more powerful mathematical tools, analytics and automation that help the HR function do its work more efficiently, according to Sumser and other experts.
At its simplest, Sumser observed, today’s AI usually comes into play when systems integration forces practitioners to spend their time doing what he called “stupid stuff.” For example, “there are teams of people whose job is effectively to copy information out of this piece of software, copy information out of that piece of software, then put it into a spreadsheet and tell a story.”
Such chores can be simplified through robotic process automation, he said, adding that RPA “is really just sort of advanced scripting with conditions built into the scripting.” That aside, he sees such technology appearing anywhere where there are those ineffective tasks, such as having a human being manually copy data from one system to another.
However, that technical environment is quickly changing, Sumser said. Not so long ago, programmers worked to preserve storage and minimize processing capabilities to optimize a machine’s performance. Today, technology has advanced to a point where developers don’t have to worry so much about making tradeoffs among processing power, memory and performance. As a result, “things that people used to imagine but always thought were impossible are possible,” he explained. Currently, those possibilities are mostly being realized in advanced mathematical search. “So, what’s in the market today boils down to complicated regression analysis expressed as probabilities of this happening or that happening.”
2. Prepare for a complex transition to HR AI
Increasingly sophisticated people analytics is sure to lead to a huge challenge in managing not just data, but the models it generates. Five years from now, Sumser believes, HR departments will have to manage millions of data models. “Every employee is going to have 15 or 20 data models pertaining to various aspects of their work and personality, and those models won’t necessarily come to the same conclusions,” he predicted. “The impact of that will be extraordinary.”
Not only will organizations need to maintain “data model farms” containing all of its models, they’ll have to develop alternative models as data capabilities and the workforce evolve, an effort Sumser compared to changing a flat tire while a vehicle is in motion. As the role of AI in HR grows, he said, “the business of managing the AI will probably be the most complicated bit of it.”
In addition, “one of the most important things to understand about AI today is that, where machines used to deliver facts, they now deliver opinions,” Sumser said. What he means is that technology such as analytics that predicts future scenarios or sensors that can take action — such as “decide” to call for help based on manufacturing flow — are essentially offering “opinions” that may or may not be accurate. In such cases — and for most scenarios in the foreseeable future — humans will still need to make decisions.
Users of AI tools must learn to work with a machine’s strengths, even while being aware of — and knowing how to utilize — its weaknesses, Sumser said.
3. Cut through HR AI hype with savvy questioning
Such dynamics led a number of people in the industry to caution against being drawn into AI hype, particularly with the buzzword-driven talk that, ironically, may often reflect a lack of knowledge about AI’s true meaning, potential impact and possible unintended consequences.
“Anybody who you can talk to sensibly about artificial intelligence is going to be less than fully informed,” Sumser said during his presentation on AI at HR Tech 2018. While the discussion around AI indeed has depth to it, he said the industry is still at the point where “almost anybody who insinuates that they’re an expert is overstating the case.”
In other words, AI may have such potential that even “experts” don’t know what they don’t know.
“Anytime we have a new technology, it tends to get really noisy. That’s inevitable,” added Jonathan Goodman, the San Francisco Bay Area general manager for The Starr Conspiracy, a marketing agency that focuses on HCM. Because there’s so much AI hype and because AI isn’t a software category in and of itself, he encouraged HR and IT leaders to drill down into how AI makes a particular product different. They need to understand how one vendor’s application of AI truly differentiates its product from its competition’s, Goodman said.
4. Understand how small HR AI efforts add up
Today, AI in HR is often discussed in terms of relatively narrow areas, such as screening candidates, analyzing the employee experience, predicting workforce needs and delivering more effective learning. However, many analysts believe that, in the end, the greatest impact will be more widely felt. “I actually think the biggest promise of AI has more to do with the entire employee lifecycle rather than one specific area,” said Dani Johnson, co-founder and principal analyst at RedThread Research.
It may be that such wide-ranging impact will occur when narrower efforts come together. Stacia Garr, Johnson’s co-founder at RedThread, sees AI as having an exceptional impact on diversity and inclusion. “If designed and used properly, AI can help reduce biases that impact how people assess each other — be it during candidate selection, performance appraisal or leadership or high-potential assessment,” she said.
The cumulative result generated by such tools “is much larger than improving individual processes,” she said. “The AI would instead be enabling the creation of a workforce and organizational culture that can drive better business outcomes across the entire company.”
Also, bear in mind that organizations have begun to think more holistically about their workforces, Johnson said. Where once learning and development, career planning and performance were three separate discussions, today, they’re addressed together. “The promise of AI is not just more data, but better information about any one employee and her goals and preferences, as well as how those align to organizational goals and direction.”