'It can become quite counterintuitive without appropriate monitoring': Canadian HR leaders discuss how 'robot sidekicks' free up time for more strategic work
The rapid advancement of artificial intelligence is fundamentally altering the landscape of human resources, evidenced recently by IBM’s replacement of “hundreds” of HR employees with AI automation tools – only to restaff in other areas shortly thereafter.
In an interview with the Wall Street Journal, IBM CEO Arvind Krishna confirmed the company’s use of AI to automate the work of several hundred HR employees, with it's AskHR agent now handling 94 per cent of routine HR tasks such as pay statements and vacation requests.
However, Krishna added, the trimming of HR staff has left room for “investment in other areas”, telling WSJ that IBM has been able to increase hiring in programming and sales, and more in “critical thinking” roles such as marketing and software engineering.
From routine to non-routine: automation’s expanding reach
While automation has traditionally targeted repetitive, routine tasks, AI is now encroaching on non-routine work as well. As Ozgur Turetken, professor of IT management at Toronto Metropolitan University, explains, even tasks requiring judgment, analysis and creativity are increasingly subject to automation.
Essentially, knowledge workers – previously thought to be safe from traditional automation – are now faced with existential questions, he says, and they're questions with no clear answers.
“Automation is more appropriate for routine tasks, but now non-routine tasks can be automated as well,” Turetken says.
“That is true for HR professionals, that is true for legal professionals, that is true to a degree for medical professionals, it has always been true for accounting professionals. These trends are existential, and nobody quite knows exactly how much of it can be taken over, or at least be, to a very great degree, done by AI.”
This uncertainty is compounded by the way AI works, which Turetken explains is foundationally unreliable. As AI systems become more sophisticated, HR professionals will need to grapple with new uncertainties around the limitations of automation and the inconsistencies of AI-generated decisions.
“We never will know with 100% certainty what AI is going to suggest or predict or come up with,” he says, adding that this unpredictable nature of the tools is where many users are currently being challenged.
“This is the aspect many people are struggling with, because they say, ‘Oh, it's a machine. People have programmed it, and therefore people are always in charge,'" he says.
"That's not true. People have generated a platform where this agent can be trained, but based on the training, what it's going to do is anybody's guess. No developer can know how the model that they train is going to behave, and that's where this is different from traditional automation.”
Automation in action: HR leader at Nelson Education
Canadian HR leaders are already leveraging automation to handle administrative burdens, freeing up time for more strategic work.
Samara Cazes, senior people manager at Nelson Education, describes how assigning administrative work to “robot sidekicks” is freeing up time for more strategic HR functions on their team.
“We have an onboarding workflow that we use, that I built out in our HRIS system, that automatically sends emails to the hiring manager when a request is in progress for them to help with onboarding their new hire,” says Cazes.
“Automatic emails to prepare them for onboarding their hire on their first day, what they can do for their schedule, what things can look like. Just that back end administrative work ... automated to go to them right when they need it, whether it's in the recruitment process or in the onboarding process.”
This approach allows HR teams to focus on higher-value activities such as talent development, employee engagement and strategic planning, Cazes explains. However, the human element remains essential, she adds, with interviewing, orientations and interacting with candidates still performed by human team members.
The challenge of customization and continuous improvement
Automation is not a set-it-and-forget-it solution, Cazes warns – in fact, the need for continuous improvement and adaptation has emerged as one of her team’s biggest learnings in the process.
“It's a lot of continuous iteration and improving, because we want to bring, as much as we can, a really personalized approach when we're dealing with candidates and when we're dealing with managers,” she says.
“A one-size-fits-all is the easiest to apply when you're building out an automation, but it is not going to give you the results that you need. It actually might create more problems, if people are getting a template canned email giving them instructions of how to use a platform, but that doesn't apply to their role, or that platform is not something that is specific to them. Or they need to be using it as a manager, not as an employee.”
Cazes stresses that front-loading investment into creating automated programs that are “as fine-tuned, personalized, and customized as possible,” will save time in the long run, and having a human on board constantly taking feedback from users and creating new iterations of tools is essential.
“If you don't, then things are going to fall off and you're going to lose the theme of what you were trying to do by making it more productive,” she says.
“It can become quite counterintuitive without appropriate monitoring.”
Integration, adaptability and the evolving HR role
For Sarem Yesus, executive director of people and engagement at Nelson Education, the biggest challenge is customization and keeping pace with change.
“We do have to change the structure of our organization quite a bit to fit into whatever's happening at any one time,” she says.
“So that means a workflow that we have set up and an automation that we have set up maybe three months ago is no longer relevant. So we have to either iterate or scratch the whole thing and restart it.”
Another challenge that has come up involves getting various AI systems to “talk” to each other, Yesus adds – making reporting to different areas of the business tricky.
“From a reporting perspective, we do have some challenges there where we do need … an analyst or business systems, we work with them to help us make sure that they're all talking to each other, so that in the reports that we need to present to the business, that's easily accessible.”
Sarem Yesus (Submitted)
The integration of automation into HR processes is not just about efficiency – Yesus sees it as an important opportunity for HR professionals to embody a more strategic organizational role. By automating repetitive tasks, she explains, organizations can empower their HR teams to contribute more directly to business outcomes, employee satisfaction and organizational culture.
“I think that HR has had a lot of changes over the years,” Yesus says, adding that rather than being afraid of AI edging them out of their roles, HR professionals should seize the opportunity for expansion.
“Initially, [HR] really was just administrative. If we were still in that space, I think this would be a concern,” she says.
“But the truth is, in the last 30 years, especially the last 10 years, we're business partners. We don't see ourselves as a cost center. We don't see ourselves as individuals or groups that just deliver the decisions that the company makes. We really are at the top and center of the conversations of the business” - and automation through AI tools is aiding that process.
Navigating the unknown: testing, collaboration and academic research
As organizations experiment with AI, Turetken recommends a cautious, research-driven approach; particularly important, he stresses, is ensuring AI programs are tested and ready before they are deployed.
“You have to build that confidence by experimenting with it. But that experiment should not come at the expense of people,” Turetken says. For example, when testing AI functions and processes, strategies such as using “digital twins”, while initially seemingly costly, will prevent disruption to employees and other stakeholders.
He also advocates for closer collaboration between organizations and academic researchers as mutually beneficial relationships. This approach can ensure organizations benefit from the latest research and best practices, while minimizing risks and unintended consequences.
“That kind of collaboration, with people who understand AI and its applications, I think, is something that should be on everybody's radar,” he says. HR’s role in facilitating these collaborations and supporting employees through technological changes will be more important than ever, Turetken adds.
“HR, or other functions, but HR really, directs the organization as to the training … in that sense, HR can take a role in this, as opposed to trying to protect or avoid the technology. You need to kind of embrace it, because avoidance is not going to work.”
