AI-Powered HR

The corporate HR department has typically been the butt of jokes and scorn, and HR executives often struggle to make their voices heard on top management teams. In the age of AI, huge amounts of people management data — from employee surveys to performance reports — can be fed into natural language processing systems to glean insights that could help organizations work smarter and more efficiently. That will make HR executives important drivers of workplace transformation. 

This episode features the insights of two HR experts who are also Smith School of Business alumni. Heidi Klotz is vice-president, total rewards, at AtkinsRéalis (formerly SNC-Lavalin).). And Dagmar Christianson is head of workforce transformation at ProFinda. The two discuss how HR leaders will have to prepare employees for the technology disruption to come. They are joined in conversation by host Meredith Dault. 

Transcript 

[Music playing]  

Meredith Dault: Imagine submitting your resumé for a job opening and having a chatbot review your credentials with you. You land the job and things seem to be going pretty well — until performance review time. Though you haven’t necessarily been aware of it, an AI-powered system has been tracking and analyzing your performance throughout the year, considering your project involvement, as well as any feedback from your colleagues. Another AI platform has been analyzing your skills and career ambitions and recommending training courses, webinars and workshops. Helpful, right? Of course, there’s that nagging feeling that every keystroke and email is being monitored, which is maybe not so helpful. 

There will be a little bit of everything in the AI-powered workplace of tomorrow, and the next few years will largely determine just how useful, or how intrusive, it will be. Will rote HR functions be automated, freeing managers for more high-impact work? Will employees have the power to manage their employment journey? Or will AI be used for control and surveillance? 

Welcome to this episode of AI Reality Check. I’m your host, Meredith Dault, a journalist and media producer at Smith School of Business, and today we’re talking about AI-powered human resources. We’re in very capable hands with our two guests, both of whom are Smith alumni. Heidi Klotz is vice-president, total rewards, at AtkinsRéalis (formerly SNC-Lavalin). Heidi also earned her Master of Management Analytics at Smith School of Business. Dagmar Christianson is head of workforce transformation at ProFinda, an AI skills-matching tool helping clients to align their business demand and strategy with their employees’ skills and aspirations. Dagmar holds an MBA from Smith. Welcome, Heidi and Dagmar. 

So, Dagmar, let’s start with you. I know you’re helping a lot of organizations to introduce AI-related technology to their talent management processes and to tap into the ton of data that HR generates. In broad strokes, how far along are we at this point and how is AI being used? 

Dagmar Christianson: Even though it feels like we’ve been hearing about AI in HR and in the HR hype cycle for years, I think we’re actually relatively early in the curve for AI applications in HR. This depends on, you know, the organization, its leadership and the type of AI. But I think if you look at Gartner Hype Cycle as an example, they put HR AI applications around the five- to 10-year mark before they reach what they call the plateau of productivity. So, I think HR leaders are exploring, they’re experimenting, they’re understanding the use cases, but many aren’t ready to dive headfirst into a big financial investment as of yet. I think if we use self-driving cars, the analogy of where we’re at with AI tech applications in the HR world, we are very much still building the sensors for this car. It’s going to take a lot of work, a lot of time, to get us in a place where we can take our hands safely off the wheel. 

The HR area that I think AI has already permeated quite well is talent acquisition and recruitment. There are so many tools that I’m sure many listeners have had some first-hand experience with at this point, either on the candidate side or on the HR manager side: video interviewing AI-enabled candidate matching. At ProFinda, where I work, we’re using AI to help match people to work internally within the company. We take into account skills, preferences, experience, accreditations and availability, to drive users to make more objective data-driven staffing decisions. So, I heard a great stat from one of our big four clients the other day on their most recent exit surveys: about 75 per cent of leaders cited lack of choice or visibility of projects as their primary reason for changing companies. 

We think about AI as augmenting a human manager who isn’t able to gather and remember and analyze even a fraction of what the machine can, especially all the variables for best fit project assignment. AI skills matching and tailored career recommendations quickly become a differentiator for your employee value proposition and your retention strategy. Otherwise, they can’t be achieved at all without AI tech. So, to come full circle on this question, the technology itself: innovating at pace and it’s fair to say HR leaders are tuned in and paying attention, but full-scale adoption of AI is a bit more sluggish. Many, dare I say, most, of these tools sit unused and I think unadopted. This is mostly because people are resistant to change, even if that means giving up some of the really automatable admin that they’re already complaining about in their existing job. So it’s a friction-full process, one that needs to have really strong executive sponsorship, whole-function, buy-in and ongoing change management investment in order to drive real business impact via the adoption of AI apps. 

05:16: MD: Great. And Heidi, how does this match up with your experience at AtkinsRéalis? 

Heidi Klotz: Yeah, that’s a really great observation and really something that I’ve seen in several organizations that I’ve worked in. So, being more on the practitioner side of HR, there is definitely an interest and a push, especially from leadership. So exactly like you were saying, Dagmar, you know, there is a lot of executive buy-in, there’s a lot of interest in moving things forward, but the adoption and the skills even from the people within the HR function are really just at their infancy. So for me, I see that as a really great opportunity, and that’s one of the reasons that I decided to go and do my MMA at Smith. I really wanted to be able to work in that space and be one of the resources, one of the people that could champion, but also understand, what exactly AI and analytics could bring to the HR function. There’s just so much opportunity. 

There are a lot of tools that are available. I’ve worked for an organization that was a people analytics cloud platform, and the overwhelming challenge was adoption and again, change management and really removing the fear that AI is going to take our jobs and replace the people in our teams. There’s that real fear. So in a large organization like ours at AtkinsRéalis, we’re 37,000 employees across the world, and we’re really just at the beginning of our HR people analytics journey. We have started with dashboards in the past year, and that’s, you know, we’re in 2023, and that’s how we’ve decided to kick things off. But we are looking now at partnering with different organizations and consultants to see how we can take things to the next level. 

07:05: MD: Right. And how, in your day-to-day, where’s AI coming into play in your role now? Like some specific examples? 

HK: Yeah, from a total rewards perspective, so total rewards is all about compensation, benefits retirement programs. We’re being asked to make more intelligent decisions or help the business actually make more intelligent decisions. So, things like understanding our population, things like how are we going to reward our people smarter based on the projects that they’re working on? And also, since we have historically been the people in the organization who have that more analytical mindset, our HR colleagues within the HR employee relations or change management, learning and development, are really looking at us as the data custodians or the people who are the ones who understand what the data is that we have in our systems to help them leverage and make better decisions and implement better programs. So, AI, like Dagmar said, I think it’s really kicked off more in the talent acquisition space, and we’re still really looking at how we can bring that into our space and really help HR make better decisions about how we’re going to attract, retain and motivate our teams. 

08:24: MD: Got it. Thank you. And Dagmar, you alluded to this earlier, but can we sort of dig in a little bit about what generative AI, in particular, has to offer when it comes to recruiting and onboarding as well as growing those opportunities, as you mentioned about aligning employees with the right projects and finding opportunities for them? 

DC: Yeah, sure. So gen AI is probably one of the fuzziest technology terms around right now. I suppose just to start off, for those who don’t know, gen AI is like intelligent Google on steroids. You can ask it to summarize information or “write me a poem” or answer a complicated question, or generate an image in a certain style, which we’ve been seeing a lot of. Fair to say that most HR leaders are actively exploring potential use cases and opportunities when it comes to gen AI. I think exploring is the key word here because many people are having a play around with gen AI technology, but for them to overcome the mental barriers and use it on a daily basis, it’s a huge culture shift. 

I think gen AI is definitely starting to proliferate in the natural language processing space. Think virtual assistants, chatbots, unstructured data processing, all very commonly used HR tools continuing to gain popularity, but now in a much more meaningful way than before due to gen AI technology advancement. They weren’t very good before, as anyone who’s interacted with a customer service chatbot can attest to. I know lots of us have probably had a little play around with ChatGPT. I know many people who have begun to use that in all sorts of professional capacities already. But the challenge is that there’s really limited consistency or any type of standardization right now in how it’s being used or the guidelines around its use. Success with gen AI, in my view, will require equal parts training and technology. So in order to get the adoption of gen AI at scale in your business, it really needs to be embedded into the workflow and the tooling. 

10:41: MD: Right. And I actually think that one issue, and we’ve heard this a lot in the podcast so far, is that people are not sure they can trust the results with some of the gen AI stuff. 

DC: Yeah. Absolutely. And I think it’s really important to find a practical application that you can then test and evaluate and get to that level of comfort. One that comes to mind is job descriptions. There’s so much bias existing in language already. Very few people are able to accurately describe the skills needed for the role. They’re really long and tedious to write. Anyone who’s been in that hiring manager or HR position is nodding their head emphatically as I say this. It’s a really good use case for gen AI. And I think a nice way to start learning how to write prompts is to focus on a task like generating better JDs [job descriptions] or just kind of picking one small area that can start to get you comfortable. 

If you have talent acquisition or a resourcing tool in your org or on your radar that has this built into the workflow, as I was saying, for defining job requirements, it’s a great way to start familiarizing yourself with gen AI and a helpful time saver. And it’s something that we’ve been working on in our product, and something we’ve recently integrated, because we’ve been getting feedback for years that no one bothers to write a description. It often just sits empty, which is extremely unhelpful for the person trying to fill that role. And to your point, if the options are it sits empty, or maybe I have to spend a little bit of time on it because I have to do a little bit more tailoring because what was spit out by ChatGPT is not perfect, it feels like an OK compromise. 

12:27: MD: OK. I see. That makes sense. I mean, it doesn’t seem surprising to me that we’ve been using AI to some degree. Because I feel like robots have been sifting through CVs now for some time. Is that true? 

DC: Absolutely. Yeah. 

MD: Yeah. So this is just sort of a natural development, I suppose. 

DC: Definitely. 

12:44: MD: OK. And so, and with that in mind, and you just touched on it, Dagmar, but Heidi, I know diversity is an issue that’s on your mind often. And one thing we’ve heard over and over again is that bias is being built into AI just by human bias being recreated and reproduced. Do you think AI-enabled systems or platforms can help advance the DEI agenda? Or do you think they’re going to work against it? 

HK: Well, I think with anything related to AI and the analytics space, it’s really important to use that as a guidepost, but not as your final and be-all-end-all. I really love the fact that with all the data that we have sitting in HR, the structured data, unstructured data, we can go beyond just looking at the number of men and women, visible minorities … We can really go and look at the individuals who are working for our organization: what’s important to them, what is keeping them in the organization, what’s keeping them motivated, why are they leaving? And then look at if there are any trends related to specific demographics. That’s where I think we have the value add. And I think that’s where it’s going to really open the door and help us understand more about how we can make some headway in the diversity space. 

Like anything with AI, we have to be mindful of bias. We have to be, you know, there’s always going to be a human component that’s needed with AI, with the robots, with the machines, right, to validate the information. And that’s a lot of the education that I’ve been sharing with my team, based on my experience. And what I learned during my MMA was it’s really important to question, it’s really important to understand what the algorithms are doing so that you can really make better decisions, but also question what the information is that’s being given to you. And that’s — going back to what Dagmar was saying about job descriptions and using generative AI — we have to make sure that people understand what the results are that are being given to them and have them be able to question it. And it’s really there to either support or debunk any hypothesis that we already have. So, in the diversity space, it’s still a really hot topic. I think it’s sort of falling out of the top priorities maybe in the past year or so, but it’s still something where we can make a big impact. It’s just a matter of educating people and having the right skill sets and the teams to be able to leverage the information we have. 

15:19: MD: Right. I mean, we’ve all heard about the … I think most of us have heard about the situation at Amazon where it was screening out female candidates. The bots were screening out female candidates because they had different backgrounds. It didn’t mean they weren’t qualified for the jobs, but they hadn’t followed the criteria or the, sorry, the pathways that maybe some of the male candidates had, so they seemed unqualified. How do you deal with that kind of thing? 

HK: I mean, again, you have to really understand and look at what the algorithms are doing. And it’s really going to be a challenge because we can’t just a hundred per cent rely on AI right now. Because at the end of the day, it’s still humans who are programming it or, you know, AI can sometimes look at past trends to forecast the future, but maybe that’s not really the right answer. It’s really looking at what we’re trying to do, what we’re trying to change. 

You know, in our organization, we’re really looking at increasing the number of females in our engineering fields, which are typically male dominated. So how? Just looking at historical data within our organization is not going to be enough. We really need to think outside the box and try to understand why there are not more women graduating from engineering programs. And what can we do to influence that? It’s really taking it to the next level and taking it to the next step, and not just relying on the past historical data that we have. 

16:37: MD: Right. So, I guess awareness is the most important first step. 

HK: Big time, big time. And that’s something I actually encourage a lot of the people on my team who haven’t had the exposure to AI, or more in-depth knowledge, is to just start asking questions, start educating themselves. There are so many webinars now in this space to help understand and explain, and I really encourage everyone to just start being more curious and asking more questions. 

17:03: MD: OK, that makes sense. There was an article in an academic journal recently that spoke about a new type of AI-augmented leadership that’s emerging. And it’s complete with leader support dashboards, which you mentioned, that reveal patterns drawn from employee surveys and electronic communications, along with algorithms guiding leaders on potential problem areas, like which team members might need more attention. The article speculated that in the future AI will not only support, but substitute, human leadership in some areas. What do either of you think about this kind of futuristic view? 

DC: Yeah, I’m happy to take this first. So, I think managers spending less time manipulating Excel sheets or BI dashboards to pull out people analytics insight, equals more time to spend proactively managing relationships or development experiences or wellness. So there’s no doubting augmentation via AI has value. And Heidi was just speaking to this very well as thinking about AI in a way that means that you are having additional hours in your day to put towards innovation, innovative thinking. Not just perpetuating what you used to do because you had to is a really great way to think about it. We’ve all had experiences with bad managers, and making the talent management experience more data-driven can only be a good thing. I see these more as nudges within a workflow or data at the point of decision for the manager to ultimately take the wheel and have those human conversations. 

Think of a future in which you can identify a flight risk within your team and be provided with a tailored list of interventions. The human’s still the one having those empathetic conversations, but the technology can help them understand where to target their energy and time. And I think advanced pattern recognition and intelligent nudges are the obvious evolution of AI for HR. With one caveat. No one should be implementing AI tooling that is a substitute decision-maker, period. The human always needs to be able to take the wheel. I’m highly skeptical generally that this would ever fly from a compliance perspective and standpoint as well. With the introduction of legislation like the EU AI Act, more roadblocks are going to come up against AI as a replacement versus an augmentation of humans. 

19:43: MD: I’m happy to hear that. But what you’re suggesting is a leader gets a notice that “hey, Maria is not thriving. Maybe a human intervention is required to find out what’s going on with her before she leaves,” and that costs our company money because we have to replace her. Is that the idea? 

DC: Exactly. So, Maria has been passed up for similar opportunities X times in the past six months, and it’s worth now having a conversation with her to see if she’s thriving and see where you can intervene, et cetera. 

MD: Right. Okay. So then, that’s something that could easily get missed in a busy, busy work life. 

DC: Every day. 

20:18: MD: Got it. Okay. Heidi, you earned a Master of Management Analytics at Smith, as you’ve mentioned, which is kind of an unusual thing for an HR leader, and you’ve written about the possibilities for natural language processing for Smith Business Insight in the past, as I mentioned in the introduction. Do you see this skill set involving analytics and AI as essential for your people management colleagues going forward? 

HK: I think it’s absolutely going to be more essential. It’s again from the understanding perspective, and that’s really what I loved about the MMA program. You could take what you wanted out of that program. You could go full-on technical. You could go full-on business. But really the goal of the MMA program was to sort of be that translator between the data side, the very technical side and the business side. And the more that I’m speaking with people in the organization, the more I’m realizing how critical that translator role is because there is a huge gap between the business, and what the business needs and wants, versus the technical people, the data scientists, the analysts, and really being able to bridge that gap to make the business case and sell things moving forward. Even things like we were saying before about putting together a new tool or a new system that’s going to help screen CVs, or generative AI that’s going to write job descriptions. The data scientist can build that and come up with this great tool and the business says, this sounds great, but what does this actually mean at the end of the day? And how do we use that? I think the more that HR is moving into the analytics/AI space, the more we’re going to need people who really understand what that means and how we can apply that to the work that we’re doing to support our colleagues within our organizations. So that’s one of the things that really drove me to do the program. 

22:09: MD: I wanted to ask, has it made it easier for you to champion AI within your organization given that you have this skill set? 

HK: One hundred per cent. I mean, I have the credibility now. I was just on a call this morning where we were partnering with a university to bring students in to do some consulting projects for an HR initiative. And although people analytics is not part of my scope of role, it’s something that people look to me to provide insight and guidance and to make sure that we’re on the right path. So, it really gave me a step above some of my other HR colleagues. And I just like being involved in those projects. And that’s what I think is really nice about the people analytics space. A lot of people just want to learn together. We just want to help grow this together and we’re all willing to collaborate. So being part of that network or being part of this evolving and really exciting field, is just a great experience. And the MMA degree really gave me the ability to do that. 

MD: Right, that makes sense. But it doesn’t mean you’re spending all your time convincing people that robots are not coming for their jobs, right? Because I know that — we laugh about it — but there are people legitimately sort of anxious about needing to arm themselves with skills to protect themselves for their future employment. What sorts of conversations do you think actually need to be happening in offices and on shop floors to address people’s concerns around their future employability? 

HK: I think it goes back to education and awareness and really making sure that people understand that while all these great tools are coming in place, it’s not going to replace the human connection and the human decision-making. At least not in maybe my lifetime. Maybe in 50 years things could be very different, but even going back to the point of having AI support leaders and leadership within an organization. And it’s great to know that based on certain criteria, you may have an employee that is a flight risk, and you want to spend time with them. But what about the other employees on your team? They need just as much reassurance and support and empathetic conversations. And the algorithm can’t really pinpoint that. 

I think we have to still ensure that we’re building relationships with the people on our team, supporting them and trusting our intuition. And AI, it hasn’t been to that level of sophistication just yet. I think that’s going to be something that we can continue to reassure people with. 

DC: I’d love to build on that question briefly. So, a couple of things. I wrote an article a few years ago in Smith [Business] Insight actually, called How Leaders Can Quell Employee Anxiety Over AI. And I think, the basic strategic change management 101 principles apply, so I won’t spend time talking that through. But anecdotally, what I’ve found is the most effective, is building excitement through the “what’s in it for me?” I work with firms that are trying to shift into skills-based organizations, and it’s not really until we explain to employees how they’ll be able to use the tool. You know, they’ll be able to see options for what they can work on next, or what upskilling it takes to get there, or where their skills are transferable, and what career doors this opens up in other areas of the company. That’s when we really get them logging in. And that can be a really challenging message to communicate and bring to life. 

I think people also struggle with conceptualizing role evolution. Almost everyone’s role is going to change. I spend a lot of time talking through with clients what this role evolution could look like with assisted AI tooling. And this is something that I don’t think gets enough airtime. So you’re really missing out on value if you are thinking only in terms of “automation equals saved hours equals headcount reduction,” which is the business case I see most commonly by far. 

For example, I will save 10,000 hours annually via AI automation versus my previous manual process. Thus, I can realize 20 redundancies. This kind of single dimensionality is one of the reasons employees have a fear around AI. It’s a missed opportunity to elevate your back and middle office. They could be focusing on enhancing those distinctly human and relational skills they could be practising, and a little bit of creative thinking is needed here, or you’ll be leaving money on the table. 

26:57: MD: So if people do want to think more positively and want to start to act proactively now — and you’re saying they should be — what advice do you want to offer people? What should people be doing now to proactively prepare for this AI future? 

DC: Heidi touched on this earlier, but to employees, I think ask questions. In its purest intention, AI for HR should be accomplishing the goal of making work better for humans and humans better at work. I’d encourage employees to get more involved in the ownership of their own data. So, understand how it’s being used. Nerd out a little bit trying to wrap your head around machine learning. So I know that Heidi was touching on webinars. You know, if your company offers a beginner course on AI, give it a try. Knowledge is power. Explore the plethora of free resources on YouTube, Coursera, whatever other learning catalogue of your choosing. 

For HR leaders, my advice to those who are maybe feeling a little bit overwhelmed with where to start their AI transformations, is that now is the time to go through the data cleanup and data transformation work you’ve been putting off. AI is not a magic wand, but those who are most prepared will see the most value in five years. 

You can try choosing one or two high-impact use cases, running a proof of concept or a pilot with a couple of vendors. So, in my view, this should be the blue-sky thinking version, where you’re testing a workforce transformation hypothesis that maybe feels a little radical for you. And you’re really de-risking and maximizing your initial investment by testing the waters before you give it that green light to scale. So, as an example, you could try piloting an opportunity marketplace to drive your mobility and retention strategy only between two specific business areas with transferable skills. You don’t have to launch this to your entire company, but you can test it out in a focused area at low risk and controlled cost. 

29:24: MD: OK, that makes a lot of sense. And did you want to weigh in, Heidi? A few final thoughts? 

HK: I mean, I agree with everything that Dagmar said, and my advice would just be to embrace it. This is not going away anytime soon and like she said, see the value in it and see how you can make a difference or how you can bring that into the way that you work. It’s just going to make things easier, hopefully. And to me it’s just something that’s really exciting. And I really try to share my passion for people analytics and what we can do with AI and HR with others, and hopefully demystify a little bit about what it can do for us. 

MD: So, in short, you both sound really positive and excited. 

HK: Yeah, absolutely. 

DC: Absolutely. 

MD: That’s great. Let’s leave it there. Thank you both so much for your time today. It’s been great talking with you. 

HK: Thank you. 

DC: Thanks so much for having us. 

[Music playing] 

MD: And that’s the show. I want to thank podcast writer and lead researcher Alan Morantz, my colleague Julia Lefebvre for her behind the scenes support and Bill Cassidy for editing support. If you’re looking for more insights for business leaders on AI and many other topics, check out Smith Business Insight at smithqueens.com/insight. Thanks for listening.