How to Make AI Actually Useful in HR
In this episode of our On The Road series, we sit down with Kalifa Oliver, Ph.D. Senior Director of Technology - People Analytics at Lowe's Companies, Inc. to explore why HR needs to stop chasing AI tools and start solving the right business problems.
Kalifa now sits in technology, not HR, leading teams across engineering, product, analytics, and people data. That gives her a very different view of what HR transformation actually requires.
Her message is clear: AI is not magic. It will only be useful if HR asks better questions, understands the problem it is trying to solve, and stops adding technology on top of broken or unnecessary work.
🎓 In this episode, we get into:
Why HR’s role is not being replaced, but radically transformed
How to redesign work before adding AI on top of broken processes
Why leaders do not want more data, they want to know what to do with it
How HR can stop chasing AI tools and start with the problem it needs to solve
Why AI should be treated like a child that needs to be taught, not a fully grown adult
What if recognition was not just a moment of appreciation?
What if it was also one of the clearest signals of trust, performance, leadership, and culture inside your organization?
Workhuman helps companies turn recognition into something much bigger than a thank you. It becomes a source of real-time human intelligence, helping leaders understand who is influencing others, who is strengthening culture, and where meaningful work is happening across the business.
Because the future of work will not be built by technology alone. It will be built by organizations that know how to make people feel seen, valued, connected, and motivated to do their best work.
With Workhuman, recognition becomes more than a program. It becomes a way to improve retention, strengthen engagement, uncover leadership potential, and connect culture directly to business outcomes.
00:11
Khalifa. Yes. Are you following me? Yes. London. When have I not followed you? I remember when you first told me, you were like, I'm going to come to the U.S. Sorry, the U.K. To the U.K. To the studio. And you all didn't believe me. I didn't believe you.
00:27
Also, every time we speak...
00:29
You're doing something new in the best way possible. In the best way possible. You're always challenging yourself. But you know what? I was talking to Shane yesterday and I told Shane. Shane's our co-founder, by the way. Just listen to him. You all were the first podcast that I agreed to do. I still don't know how y'all found me. You were the first podcast I agreed to do. So whatever I am today, I kind of blame y'all.
00:51
By the way, that's a real thing. Because what happens is, well, firstly, thank you for saying yes. But I also felt like the world needed you.
00:59
You know what? I agree. They needed your energy, your personality, your vibe. So I feel like we gave that to the community. And now you're speaking everywhere. We're on your stage and you're all over the world. I said to him yesterday that I think...
01:15
I said yes to you and I still don't know why I said yes, but I said yes. And I always say that you all were the first outlet that I realized I had permission to share and to speak and have fun with it.
01:32
Yes. And then that led to a whole bunch of other opportunities to speak. And here I am still speaking. Speaking. Is that our fault then? So basically, I think, I mean, first and foremost, it's been amazing also just in our time we've known each other to see you, your career. See me? Yeah, progress. What about y'all? I mean, we're just on the ride, isn't it? Y'all grew. Y'all, I'm like, I remember, remember when it was what, five o'clock in the morning?
01:57
And y'all thought I was not going to join at 5am. One of our events. Right. And then...
02:03
Y'all were in the studio. And then I've watched y'all grow and your platform grow too. Yeah, we didn't have a studio when we first met you. Right, right. When I was like, oh, I think you made a mistake.
02:14
We have 5 a.m. here, so you're probably not going to be there. And I'm like, no, I'm here. I was like, why? What is wrong with you? Go to sleep.
02:24
So since we last spoke, a lot's happened, right?
02:28
So when we first met,
02:30
I keep saying Black & Decker, but it wasn't. WPP. WPP. WPP. And what was your role there at the time? I was the head of people listening. And then you went to Black & Decker. And then I went to Message Bird. Oh my God. We don't need to talk that long. And you're right. I went to Black & Decker after. Yeah. And then I went to Messagebird. Is that the right order? Yes. Yeah, I think it is. I love the way I'm telling you about your career, by the way. It's crazy. And then I went to Ford and now I'm at Lowe's. And now I'm in a position where it's almost like my whole career led to this. It was like I couldn't refuse this. This is this fusion between technology and HR in the most perfect way possible. What makes this different though? So you're Senior Director of Technology and People Analytics, but what makes this different to your previous roles? I sit in technology as opposed to HR. Why is that important? Well, one, if you've ever worked in HR, it is very difficult to get the tools you need when you're in technology, right?
03:24
You often feel like, a little bit often, because you're not doing some of the transactional work that needs to be done. But when you're in technology, you're speaking to people who...
03:38
actually understand what you're saying. So one big difference for me is I have engineers. I have a product team, right? I have an analytics team under me across two countries. And we're building out products, which is so fun, so exhausting, but an amazing thing to do. Yes, it's an amazing thing to do. So I love it.
04:02
It was, I feel like my whole career led to this. Does that make sense? Not that I'm taking credit for your career.
04:08
But do you also think a big part of that was the fact that you built your personal brand and that you speak and that you've like, you know, people got to know you and your work? Absolutely. Because, you know, I think I remember when we first spoke, one of the things you would ask me is one of the lessons I had learned. And I said that I would shrink. Right? And it took me a long time to learn to find my voice and be okay, for example, with being smart. Because one of the comments I had received in a performance review was, you know, Khalifa is so smart that she makes people feel stupid, so she should work on that. That's insane. Right? That's the craziest. I remember you first told me that. That's crazy. And if you was a white man, that would not happen.
04:56
Right. And I'm like, why am I on your mind like this? And it took me a while to, I don't have to apologize for this, right? So instead of feeling like I need to shrink, I realized I can speak, I could be smart, I can be transformative, I could be impactful. And that has helped me speak out and run my mouth on LinkedIn.
05:16
You got so excited that you were like, I'm going to write a book. I'm going to write a book. I'm going to write a book. You went from like no brand, just leave me alone. So now speaking on stage all over the world, writing a book. How amazing is that? It's amazing. It's been right. And sometimes I myself am surprised at how much I've done. Like I've literally gone around the world, right?
05:42
I flew to London and said, I'm coming on your podcast. And you turned up. I turned up. yeah so i think it's allowed me to take risks so yeah whatever um you're speaking you spoke twice right or did you spoke yesterday today today tell everyone a bit about one what you was chatting about and maybe some of the takeaways that you took away from some of the sessions Yeah, so a big focus, of course, is AI, right? Everybody talks about AI. But I think we've crossed the point of saying, how do we collect data in HR? Because we've been talking about that for years. Big data, small data, whatever. We have data. The problem we have right now is now we have this technology and we have data. And how do we make sense of it? So a lot of what I've been talking about is, how do we make sure we're ready to go to the next step? And how do we make sense of the data we have? Because leaders don't really care about the actual data anymore. They want to know, what do I do with it?
06:39
And that's why they keep talking about AI. They're not really talking about the tool. They're talking about the outcome. And we in HR don't talk enough about, so how do I get from point A to point B?
06:50
we're still kind of stuck on point A. How much more data can I collect? And how much data can I just spit out at you and tell you, go ahead and make a decision? It's like a safety, like a comfortable place now. It is an uncomfortable place because HR has been teetering into the future very slowly. Now it's been pulled. Now it's being pulled. Legs kicking.
07:12
And you know what? We got a little bit of a head start with COVID, right? HR got shoved into the forefront before they were ready and then slowed down again, right? And after slowing down, here comes AI saying, nope, keep going, right? So we can't afford to be behind anymore. We just, we can't. Yeah. I suppose the question is, though, like, now we have the data, we have AI, like...
07:40
there's a lack of guidance of where do we start. Most people are like,
07:45
Where do we start? I think it's, one, it's where do we start? And it starts with asking the right questions, which we don't like to ask. The way I describe AI right now is AI is now turning into a solution, searching for problems to solve, right? And we have not started to ask ourselves, well, what's the problem? And let's find the right AI solution. So we're working backwards here. The other thing I think we have to get past is the marketing aspect. of AI, right? This idea that AI is this Harry Potter magic pill that would solve all ills and think about, well, which part of this tool can I use, right? Which thing fits which part? Because AI is about how you train that model and what you're training it on. So I'm like, listen, it's not a wizardry, you know, but if you think it could be, that's what they'll sell you. But also it's going to do what you tell it to do.
08:41
It will do. And if you teach it garbage, it will show you garbage. And I'm like, think of AI as a child, right? We're treating it as a full grown adult. It's a child. It needs to be taught. It needs to get information. And the one thing that I stressed yesterday in my talk was it cannot unlearn information.
09:02
I've had to delete entire. The entire model. Yeah. And start again. And start over. I don't think most people know how like, yeah, it's painful. And it's painful. You cannot. And I'm like, because we have to be more careful up front. Can't remove that context once you give it to it. It won't unlearn it. It will not unlearn at all. That's the problem.
09:26
You can try and make it, but it doesn't work. It doesn't. Yeah, I've made some big mistakes recently and had to delete and start over again. But to your point then, that's a different skill now of like,
09:38
what is the right question to even ask right what is the problem you're trying to solve that's the first thing but that's but and also but that's no that's not technical it's not and that's the thing it's a technical tool to solve a practical problem i'm like when you're building a house are you really when when you go to build a house are they saying well what's the technical space that you know what are the power tools we've got to use yeah
10:04
Right? I want this stove. It is up to the people like me who have the technical skills to figure out what's the dimensions and stuff you need in order to build this thing that you want. But I don't go into the store and say, I want a stove that turns left. No.
10:22
I can do that. Your job is to just think of the realistic, practical things that you want in this home. I want a garage.
10:29
I want a big bedroom, right? And then I'll figure out the rest of it. What do you want the experience of the house to be? Exactly. For those people to live in there. That's a good analogy, actually. Yes, and that's what I think we miss. I think we think we have to go in and be the expert at it. No, you need to be the expert of the thing that you're trying to solve. You know more about the problem and let the people who have the data then help you solve the problem. I think where I see people hit a wall though, is like, in order to make this really work, it's not just starting with AI, it's starting with like, completely redesigning work itself.
11:12
Every like process, like from scratch. Right. And the other thing is just because you have the tool, do you need to use it? And just because you can, should you, right? And that's really, really important in HR because you shouldn't.
11:29
Don't do it. Start with your default as you shouldn't. And then just work from there. Right. You know, I was at a conference talk recently and I told people to think of the use of AI in HR the way we think of science.
11:41
Science is not about proving things. It's about disproving.
11:46
Your default belief is that whatever you think is right is wrong, right? So when you're thinking about AI and AI models and you're training it, assume it's going to be wrong. Assume whatever it gives you is wrong and then go back from there. Let you, your job is to prove it right, right? After it proves, you just go, this is going to be wrong.
12:10
I'm going to add data to this and then let me see if this is correct. That's the human and the new part of AI that we don't think about.
12:19
Assume it's wrong and you will save yourself so much tears when you realize that you went into court with a case that was completely made up because AI just made it up. Yeah. But also we're seeing even with the vendors and the HR space and companies get in trouble for, you know, big lawsuits and stuff like that. Massive hallucinations. Yeah. Based on the hallucinations. That scary thing about it, right, is if it messes up, it can do it at scale. At scale. In a huge way. I can't remember what company it was just now where it just deleted the entire... It asked it to clean up, they turned and deleted everything. Yeah. I don't know which company it is. And I'm like, assume it's going to be wrong, so back it up. Assume it's not going to understand you, right? And so it'll save you a lot of heartache.
13:14
What does this mean for the role of HR? Because this is so far away from... I think the biggest thing we have to do in HR is get away from the idea that we're being replaced.
13:29
I don't think we are. I think our skills are being transformed. Are people saying that? Yes. People think it's coming to take your job. One of the things I've talked about with one of my peers was the idea that we can get people in HR to use our AI tools. I said, you can't? Because a lot of them are using it for shopping lists, for everything else in their life. Why do you think...
13:52
They're not using it at work. It's not a skill and it's not curiosity because they're showing you they will use it. They just won't use it here. What is it that we are doing at work that is not messaging the impact of AI tools properly? And it's because we're trying to bring the people to the tools instead of bringing the tools to the people and letting them figure out how it can help create efficiency. They know their job more than the tool does.
14:20
Or we do. And so we're saying, use the tool. And they're like, okay, well, first let me figure out my job. That's the bit I was saying. Because people don't know what to use it for. But they are the best place to do that. Giving everyone access to a co-pilot license doesn't mean anything. It doesn't mean a thing. I got co-piloted to where?
14:40
To do what? To do what? To do what? Yeah, we just expect people to know. The human condition, and this is what HR practitioners need to know in general, the human condition is, I will accept things that will help me solve a problem that I have.
14:55
You don't need buying at that point. They're all the way in. Right. And there's nothing wrong with people saying, okay, I'm going to learn this, but what's in it for me? Yeah.
15:03
That is just a value proposition. There's nothing wrong with it. Customers do that every day. If I buy your product, what's in it for me?
15:12
Yeah, I started this process only like last week with my team where everyone's documenting their work and the tasks every day in a shared document. And then we're starting to identify areas where we can improve and they're seeing it themselves. They're like, oh, wow, I can like... I don't have to do that three hours of admin over here. I don't have to sell that to them. Exactly. Now I just have to help them understand how they set up and how does it work and guidance around what you should and shouldn't do.
15:44
The foundation of HR technology usage is not sexy. That's the problem, right? We're selling this sexy thing and saying, oh, it could do this and it could do this and it could make your life. But the foundation is not sexy. The foundation is trying to figure out what you do on a daily basis. Yeah. That is not sexy. That is annoying. That is documentation. And it takes, there's no quick way of doing it. There's no quick way of doing it. It takes time. Right? And then the ironic thing that I've seen is then people ask AI, what's my job?
16:14
AI doesn't know. That's crazy. So you ask them for a task listing and say, you know, what are the things I do every day?
16:21
I don't know. Unless you're connected to all of your tools. The system we use, we use Perplexity.
16:27
And we have that connected to our CRM, our marketing, our sales platform, our email, our Gmail, our calendars, our Zoom, everything. So it does actually have a really good picture of what you actually do. But imagine somebody sitting in HR answering emails every day.
16:47
You connect it into your email. It says when you answer emails all day long. That doesn't set an impact. That's why I'm documenting with the team now, right? And there's things they're doing I don't even know. I'm like, oh, that's a thing?
16:59
Or worse, are we still doing that? Why are we still doing that? Start with removing things before adding. Exactly. Everyone's so quick to add AI on top. It's just stripped back. You check and you go like, should we be doing this still? Because we don't need to automate this. We need to get rid of it. Completely.
17:20
And I think that's what HR has to get to that point. Yeah. One of the quick questions I asked the team and actually Shane recently, I was like, if we just list out all the things that drain our energy and all the things that give us energy, no surprise that all the things that drain our energy were things that actually could be automated.
17:37
Like no surprise whatsoever there. I like that question. I think I might steal that question. They didn't think about it as like work. Yeah. And it was like easy to digest. Just write down all the things that you just steal your energy. I did it for myself. I'm stealing that. I've got two agents now, assistants that manage my inbox and the podcast, the guests and scheduling and everything. I always save like a day a week. I love it. Oh, I'm stealing that, Chris. I'm stealing it. Can't let you go without talking about the book. Okay.
18:10
Tell everyone a bit about the book. The why and then... The why. So the book is called I Think I Love My Job. And that think is very intentional, right? Because I think we've gotten to a point where people feel like they need to feel the same way about work every single day. And there are days you go, I don't know if I like this. Yeah. Even if you're passionate about it, there is a Wednesday somewhere where you're like, I'm done, right?
18:38
And so that's why I talk through that. It's just like Angela Duckworth said when she was here yesterday. She said, sometimes you want to stop yourself from quitting on a bad day, right? And those are the days where you go, well, I think I love this show, right? I think I do. And it's about...
18:56
helping organizations understand how to create employee experiences, focus on building real culture, using data, while also thinking about business outcomes and people-centered design. So I focus on what is the actual value proposition of your company and how do you really define that and live that, but in a data-centered and people-centered way, because we have a tendency to see these things as opposing forces. I can't care about people.
19:25
And how their experience work and make money and use data. And I am here to say we use all when we should use and. A lot. Everything can be true, but it requires intentionality. And my book is about how do we get to that level of intentionality? How do we do it strategically? Step by step. I feel like the book's become more relevant now than ever. Yes, it has. So buy it.
19:50
QR code right here. Scan it right now.
19:55
It has been. And it's funny. When I started writing it, as you know, like I had changes in my job and stuff like that. When I started writing it, it wasn't something I set out to do. It was just because I would talk to people a lot and they would be asking me these same questions. And you were doing the work. And I was doing it work. And then I was like, well, I guess I could write it down. And I just kept writing. And you didn't even have AI to help you. I did not have. Yes, I want to make that clear.
20:20
I did not have AI to help. There's loads of authors out there now. Everyone's an author. I did not have any AI help. Yeah. Did it the hard way. Listen, before I let you go, because I know you're going to be speaking later.
20:32
Where can people follow you? Connect?
20:34
I am always running my mouth on LinkedIn, so you can always find me for everywhere else. So I'm on TikTok. I'm on Instagram. You're on TikTok? Yes. You need to follow me, see? TikTok, Instagram, YouTube, Blue Sky. I don't know what that one is. What is that? Upwork. Blue Sky is like the Twitter.
20:56
the open source version of Twitter. I'm everywhere. Yes. I'm everywhere and it's at Dr. Khalifa. Oh, that's it. Across every platform. Yeah. I'm easy to find. Maybe too easy. No, no, there's never, there's no such thing. And then the book's available in all of those. The book is available wherever your favorite books are sold. And it's an audio book too, so. Got audio? Yeah. Is it your voice? It is my voice. It is? It is! Oh, I know how long those things take, so well done.
21:21
Let's say I have a newfound respect for people who do audio and narration for a living. It wouldn't be the same if it wasn't your voice. You know, I wasn't going to do it. And that's exactly what- I can't imagine. That's exactly what my sister told me. She's like, you have to do it yourself. yeah it's fun cool well it's good to see you again as always always fun catching up with you so i need to see you what in charlotte at some point 100 okay what we're gonna do and we're doing an event bet with you guys yeah i'm in see you later thank you
Kalifa Oliver, Senior Director of Technology - People Analytics at Lowe’s Companies.