Automate Before You Even Think About AI: An HR Leader's Playbook

Every HR team has been told to do something with AI, and almost none of them were handed a plan for it. Kristin McDonald, a self-described HR nerd who has spent years in HR tech, and Dani DeMaio, a senior director of HRIS in private equity focused on M&A, have both lived that gap from the inside. Their take is quieter than the hype: most of what leaders call an AI problem is an automation problem in disguise.

‍

Where should HR start with AI when there is no roadmap?

‍

Start with the business problem and the process under it, before the tool. McDonald argues the foundational work, naming a process and where AI might fit, is exactly what teams skip in the rush to adopt. She puts the moment plainly: "we're being asked to fly this plane or design the plane while flying it. And I think we don't even have a flight plan yet. We don't know who's the pilot."

‍

Where does AI genuinely help, and where does it fall short?

‍

AI is strong at summarizing, drafting, and answering routine employee questions, and weak anywhere the work needs system-specific judgment. DeMaio found that troubleshooting Workday with a chatbot cost her more time than doing it herself, because she had to translate its advice back into the system. As she puts it: "AI does not have the answers for me because it takes knowing operations and a little bit of total rewards and all these different things to be able to actually come to conclusions." A headcount reconciliation that needed four separate data loads was something no chatbot would have flagged.

‍

How should HR measure AI's impact if not by headcount?

‍

Measure it by time returned for strategic work, not by clicks removed or seats cut. McDonald reframes the whole ROI conversation around what the freed time makes possible: "are we making employees lives easier or managers lives easier, allowing each other to be more strategic? And that's the impact, not like less clicks or one less transaction." The freed time is the return, showing up as capacity for the strategic work leaders keep saying they want.

‍

What is the first practical move a team can make?

‍

Automate and document your biggest time sinks before layering AI on top. DeMaio maps where a team's hours are spent, then tracks the time given back on a spreadsheet. At one company that exercise let her tell leadership, "this HRIS team just saved a thousand hours of productivity," with no AI involved. The same exercise documented what each person did, which de-risked the business during private equity transactions and turnover.

‍

How do you handle the fear that AI will take HR jobs?

‍

Treat it as real, then point people toward skills that travel. DeMaio's tough-love version: "job security is never in the company that you're working with," so the move is to deepen the parts of the craft that stay in demand, like payroll, integrations, studio, and advanced comp. McDonald adds that the wave of layoffs has already made this a hard sell, so honesty beats reassurance. Both land on the same practice: carve out time to use the tools until they stop feeling foreign.

‍

The throughline is patience with direction. The teams that come out ahead over the next two years are the ones automating and documenting now, so that when AI is ready to sit on top, there is a clean process for it to stand on.

Kristin McDonald (00:02): We're being asked to fly this plane or design the plane while flying it. And I think we don't even have a flight plan yet. We don't know who's the pilot. We don't even have a...

‍

Dani DeMaio (00:10): What was interesting that I found is I spent more time actually trying to understand what AI was telling me to do, translating it to Workday lingo, and then figuring out in this system. From when I could have just done it myself, from years of...

‍

Kristin McDonald (00:25): Are we making employees lives easier or managers lives easier, allowing each other to be more strategic? And that's the impact, not like less clicks or one less...

‍

Dani DeMaio (00:34): So like, are you going to take my job with...

‍

Logan (00:38): Every HR leader is being told to do something with AI, but nobody's handing out a playbook. On this episode of Pulse by HRBench, I sat down with Kristin McDonald and Dani Woods, two HR tech leaders who live inside these systems every day, to talk about what AI adoption looks like when there's no roadmap. We get into where AI is genuinely useful right now, where it completely falls short, and why most teams should focus on automating before they even think about AI.

‍

If you're trying to figure out where to start without creating more problems than you solve, this one's for you. Let's get into today's episode.

‍

Logan (01:10): We're here to talk about AI transformation for HR folks. And if you didn't hear enough about AI last year, we're gonna keep it going. I'm here with Dani and Kristin, HR tech aficionados. Kristin, what would you like the audience to know about you?

‍

Kristin McDonald (01:25): HR nerd for sure. I would love to say that's me, but I definitely feel like I'm just more of a nerd than anything else. I have been in HR longer than I care to admit, but I love geeking out on all things HR, tech, people, process. Those are the things that keep me going in my career and the passion I have for them. So yeah, I want to talk about that. I'm not here to pour that.

‍

Logan (01:49): Great. And Dani, what would you like the audience to know about you?

‍

Dani DeMaio (01:52): I'm the really, I think, like really bold person on LinkedIn. I poke fun at HRIS a lot in corporate America, but I currently am a senior director of HRIS. I'm in private equity. I really specialize more in the advisory space and M&A. Similar to Kristin, I've been doing this for almost a decade.

‍

And I loved HR. I started in HR, started as an individual contributor, and then worked my way up from analyst, senior analyst, to manager, and kind of progressed from there. I don't think I'd call myself a Workday expert, because the more I realize I know about Workday, I think I know less about Workday. But I absolutely love technology.

‍

Kristin McDonald (02:36): It's funny you say that though, because I feel the same way sometimes about technology. I'm like, every time I feel like I know it enough, I just come back down, because there's still so much to learn. So I hear you on that.

‍

Dani DeMaio (02:47): Super humbling.

‍

Logan (02:48): Let's kick it off. A few of the topics that we're going to cover today. We're going to do leading AI when there's no roadmap. Everybody's being asked to use AI. How do we measure AI when people are defaulting to headcount as the measurement? And then what are some ways that HR leaders can actually get started today with AI initiatives?

‍

For the first topic, I want to kick this over to Kristin. Everybody's being asked to do something with AI. So how should HR folks think about that, in bringing AI when there isn't a clear path, but we're being asked to use AI?

‍

Kristin McDonald (03:25): Yeah, I mean, AI solves everything, right? That's the message. We need AI, and what does that actually mean is very unclear. There's just foundations that we haven't set yet, right? AI is not going to be what solves everything, right? Just like any other tool. There's a need to really sit down and understand what are our processes and where is AI potentially the solution and where it could be something else entirely.

‍

And that foundation, those conversations, I think are missing. So we're being asked to fly this plane or design the plane while flying it. And I think we don't even have a flight plan yet. We don't know who's the pilot. We don't even have a destination of where we're going with all of this. And I think that's where we need to start there first.

‍

Dani DeMaio (04:12): I definitely started like weaving in little things I could do in my day to day with AI. Outside of asking ChatGPT how to make me sound nicer on emails, because that's like my number one thing I need to work on, I will try troubleshooting like calc fields with it. But what was interesting that I found is I spent more time actually trying to understand what AI was telling me to do, translating it to Workday lingo, and then figuring out in this system, from when I could have just done it myself from years of experience.

‍

I think too, you definitely have to be smarter than it to really understand what it's telling you to do, so you can kind of pull pieces out of it. But I've started really small with summarizations of things. Help me write this email, help me write this policy. What am I not thinking of? And prompting it that way, just to, like almost as gross as it sounds, be a thought partner. But never thought I'd use AI as app, but like ChatGPT is my best friend. So I will use it in that way, but there are still so many parts I think of our jobs that when we deal with really intricate things. For example, I'm working on a carve out right now.

‍

Dani DeMaio (05:18): AI does not have the answers for me, because it takes knowing operations and a little bit of total rewards and all these different things to be able to actually come to conclusions on things. Versus I just have it do small little tasks for me instead.

‍

Kristin McDonald (05:31): Yeah, I call my Copilot bestie. Well, I work, but I use Copilot, right? So if I need to summarize an email thread that's so long, I'll be like, okay, just summarize the action plan. So it's so helpful just for that perspective, and also just giving you a starting point. Like, if you're trying to create an SOP or put together a process or something, it can give you such a great starting point.

‍

Dani DeMaio (05:34): Do you use Copilot or chat? Like, what do you use?

‍

Kristin McDonald (05:57): But I agree with Workday, it's just anything that's system specific, it's how do you use it when it comes to that in your everyday job itself and the tools you use.

‍

Dani DeMaio (05:57): Yeah. Have you heard of Mando? I think I pronounced it correctly, but they're doing that where they have basically a ChatGPT specific to Workday that I haven't personally used. I actually should take a look at that, because I think that'd be super helpful to be able to tell us things. But I think if you're really green, you're not going to understand what the guide is telling you to do in Workday, because there's so many catches that you have and troubleshooting that you have to do on your own, that you just do it enough times and you figure it out.

‍

Kristin McDonald (06:39): Yeah, it's the documentation for Workday, I think, is what kind of drives a lot of Mando, right? As the knowledge source, which is amazing. But I think doing that in real life and trying to take documentation of like how to configure something, it's great, but also your process may look very different, or there may be nuances or things that you still need to figure out for yourself as well.

‍

Dani DeMaio (06:45): We had to do like a big head count. Just like with AOP coming up, we had to do like a reconciliation at this one company. What ChatGPT would never tell me to do is that I needed four separate loads to do this. I needed the data. It might help with data formatting here and there, but I did try it to plug. I plugged in the EIB, not with the data on it, but I was like, tell me what fields I really need to use. It was giving me the wrong information left and right. And I'm like, I thought it was fun. I'm like, yeah, we're not there yet.

‍

Kristin McDonald (07:35): Yeah, that one's, there's a lot of, and so many nuances that are just, you still need that human in the loop, right? I think that's the key, is there's still somebody that needs to be a part of finalizing it and doing the configuration or the outcome.

‍

Logan (07:55): Good. So we're getting into the weeds already with configurations. So when there's still ambiguity, I want to turn this over to Dani on kind of how you measure AI. And how should HR start thinking about AI impact when oftentimes, at least right now in the early stages, we're thinking about productivity and headcount? How should we start to think about measuring it outside of just headcount?

‍

Dani DeMaio (08:20): I think you got to ask yourself, is it causing more issues than the problems it's solving? And kind of going back to using these different tools, for example, you want to roll out a brand new chatbot. If you don't have any policies to feed it with information, you're not going to be able to use it. So I will see companies, it can take like two years to go into actual transformations, depending on what maturity level they're at. You have to, AI is really good for those strict employee questions. Summarize this, read this, look at this with HR. But then if you want to keep dialing it in further, you've got to really figure out, is this something that is going to cause more issues than it's actually going to help with?

‍

Kristin McDonald (09:01): Yeah, let's see what business problem is actually solving, or the impact that it's having, right? And I think that is not the starting conversation, which it should be, right? I think knowing where to use AI or any tool needs to come back to what is the business problem and what's the impact that it's having. I think that's the biggest thing. Like, are we making employees lives easier or managers lives easier, allowing each other to be more strategic? And that's the impact, not like less clicks or one less transaction, right? It's where we're opening up time to be more strategic, which I think so many people want. And I think that's where we don't hear enough about that being the ROI, for whatever reason.

‍

Dani DeMaio (09:29): I think like availability too, like with these different bots that we can put in place and employees can go to it for questions instead of relying on the nine to five HR team. They can start and like have tools at their disposal and they can work more autonomously versus relying on us to answer a lot of their different questions. And I think too, at different companies when I've gone in, I will take a look at like all of our foundations. Can you pay people? Like, what is your pre-hire to your first paycheck, and what's in between, and where are we spending the most time? So I will go nitty gritty. This is the amount of hours it is, this is the amount of people that it's hit. Because like you said, you want to hit the actual impact. Just because one person complains too many times about one task does not mean you need to stick AI on top of it.

‍

Kristin McDonald (10:29): Right. And where is that work being done? I think that's where HR could be guiding, is, do you know what your workflow even is? Like, if you go into your business process in Workday for hire or for termination, do you actually know what that looks like end to end and where those opportunities... No, no, they don't. And that's where those opportunities lay, to actually start there and be like, are there points that, in the...

‍

Dani DeMaio (10:46): No, most people don't.

‍

Kristin McDonald (10:56): BP or in this process that we've created that we can automate, versus where do we still need judgment or decision making by a human, and really being specific with where that impact can lie.

‍

Dani DeMaio (11:06): I think we're running into capacity issues too, from what I'm seeing, right? Like, people and companies only move so fast. Like, think about how long it takes us to book a meeting with four people. You're looking at like next week, let alone going through all these processes one by one. And these are all the people... like, I worked for another company. We paid a consulting firm a lot of money to tell us five issues that I already was aware of.

‍

Dani DeMaio (11:38): By the time we got it, we were so exhausted and there was no budget left to fix the issue. And I keep seeing this time and time again.

‍

Logan (11:46): I think one thing to keep in mind, especially on that, the length of time to book a meeting or anything like that. We've been doing workflows with SaaS and other tech for years now, and how difficult and hard it is for companies even to do that. Now to flip the switch on AI and bring AI workflows in is, I would imagine, still just as difficult.

‍

Dani DeMaio (12:09): Even if your tool works, it's still... it's like calendars and it's people's availability and it's all these different fires that are going off during the day. You could have the greatest tool in the world to help you schedule a meeting, it's still gonna take you a week. Like, it just depends.

‍

Kristin McDonald (12:22): Right. Yeah, it's very much not so much a tech problem, this whole conversation, right? Yeah, exactly. Yeah, if it could solve just creating meetings for me and looking at all... like, I love Copilot and Outlook, but it still doesn't give me everything I need. I still need to go in there and look at schedules.

‍

Dani DeMaio (12:29): It's only people. No, because if someone's blocked, if they block their counter, then you gotta reach out to them. But it just goes to show, we have all these mechanics in place that we still have to get over before we can do what we need to do. And I think that's the bigger picture of what's actually slowing it down.

‍

Logan (12:58): So then, I guess, trying not to slow things down, what are some practical ways that we can start to implement AI into companies? Like a tactical thing somebody on the call can do today.

‍

Kristin McDonald (13:11): Well, I do think we can't forget the fact that most of the tech you use has some AI capability in it, or automate, that you can just turn on, right? So there's some quick wins that I think we don't want to forget about. Like, within Workday alone, there's some agentic capabilities that, wanting to turn on and start using. Like, start getting, start using it, start getting your bearings and understand that, while you're working on kind of like longer term, your bigger use cases or potential opportunities, and doing that work to understand what that looks like.

‍

Dani DeMaio (13:48): I would start asking people what they're already using AI for, to, number one, figure out how you could create a process that amplifies it more. Or I would find out what the biggest time constraint is. So like, if you take a look at what your day is, where are you spending most of your time? And that's the piece that I would maybe use AI, or maybe it's more of like an automation. I think in HR, with integrations too, we'll go through a process we need to do and we don't need to layer on AI, but we just need to automate it first, and then we can stick in AI. So I always started with like the biggest time sucks in an entire day, and I would start and slowly bring in AI. Same with HR.

‍

Kristin McDonald (14:29): Yeah, it's interesting. There's not more conversation about like really looking, sitting down with people and understanding what their day to day does look like with those, you know. And doing that work to ask, to answer that question of where can you automate? I do agree with you too, it's a very valid point that automation is always meet AI, right? There's other ways to automate.

‍

Dani DeMaio (14:38): Like, I'm in a Workday implementation right now, and that's the next thing. As soon as we are done and we are like fully implemented, the next thing that we do is we'll start an activity, like going from team to team and saying, where do you spend your most time? I had to do it at my last company too. And that's where we found there was... we were able to give back. Like, we would start tracking how many hours we were giving back for manual tasks, and we keep a spreadsheet of it. And then come performance time, I could go to leadership and say, this HRIS team just saved a thousand hours of productivity. Like, that in itself is why we want to get into the world of AI, but it's not necessarily using AI. It was just automating things that were never thought of, because no one knew what they were doing.

‍

But the cool part with that is I then got documentation of what people were doing. So if you're in private equity, we have a lot of turnover naturally, right? Because there's always transactions to be done, and that just got done. So we would keep a repository of what everyone did, and it actually would de-risk us during transactions as well.

‍

Kristin McDonald (15:50): Yeah, that's key, for documenting, because how much knowledge is in people's heads? And that's true of just even configuration that you do in the system, for even like businesses' usual activities that you have. Performance reviews, annual comp cycles. Like, if that's not documented somewhere, you know, that person leaves or something happens. And that's what also will allow for those conversations to be a lot easier about where you can automate and really have those steps in the process automated.

‍

Dani DeMaio (16:21): How have you stopped people from feeling threatened by that? Because when I was running that project and they were sitting down and thinking through everything, I remember someone saying to me, so like, are you going to take my job with technology? And I'm like, well, obviously that's why I'm hired. But no, I'm just kidding.

‍

Kristin McDonald (16:36): Well, think about the messaging, like all the layoffs have had on people. No wonder people don't trust AI, because companies are laying off employees, you know, and for cost cutting or reduction or whatever the messaging is. No wonder people don't trust AI. Like, think about that. We're already going on an uphill battle for AI adoption. And that's not helping, that people are worried about their jobs being lost or their jobs changing.

‍

Kristin McDonald (17:03): I think the reality is jobs are going to change, right? And that's just... that's not just AI. Like, haven't we seen that since technologies continue to evolve and shift? It's not new, just AI. Yeah, maybe it's accelerating it. But I think, especially if you are in tech, work in tech, that is the reality. Your job's going to change. You know, how do you continue to evolve with that and be open to that, I think is key. And you know, it's probably a really tough sell right now for people, especially if they do like day-to-day configuration. Like, what is this going to look for me long-term? I love just sitting down, you know, SAP, Workday, whatever, doing my configuration. Like, that's what I would just want to do. And they're worried some of that's going to be automated. It's not an easy conversation, and everybody's perspective is different on that as well.

‍

Logan (17:51): What kind of advice would you have for people that are feeling like that?

‍

Kristin McDonald (17:55): It still comes down to, I think, we keep learning. I don't think we should ever stop learning. We need to evolve just as fast as technology and things are evolving. Like, it's not just technology that's evolving. Businesses are changing. So many factors continue to change business and jobs. And I think it's hard to give advice that doesn't feel threatening, because we also have to be real about it. And there's always going to be... go ahead, Dani.

‍

Dani DeMaio (18:25): I was gonna say, I agree with you that our jobs are gonna change. I think, tactically, one thing I've done is I wrote down... so I've taken 20 Workday courses, and I decided I'm gonna redo and go through some other courses. But if I think about it, what are the safe jobs, I think, in HRIS, or the safe modules? It's payroll, it's going to be integrations, it's going to be studio. There's some of the most sought after skills. And as a hiring manager, I see the rates of what the value companies have on these more like the pack consultants, we call them.

‍

So I decided, like, I know integrations, but I'm gonna get deeper into integrations because I think that is going to be like the next gateway with AI. Like, what are we going to integrate into, how, like what platform are we going to use that then calls Workday to do different things? That's where I think we're headed with it. And I agree with you, like the roles are going to change. The admin tasks of like creating sup orgs and things like that, that's eventually probably going to go away at some time. But how about thinking through how we name them, or how we work with the business to consolidate job profiles, job catalogs, is going to be a really big one. If you can learn advanced comp, those are the ways, like that's the way to go. But just basic report writing, I think, eventually is going to go away.

‍

Because if you think about all the tools that are out there, like even like an HRBench, there's tools that are being used that are gonna be easier that integrate right into the system, that give you analytics or that give you different things to look at. So I would focus in on payroll integrations, which I think are the worst areas of Workday. You want a job?

‍

Kristin McDonald (20:02): You mentioned report writing. I hated writing reports. Like, I just wanted to be like, I want these fields, can you just make this happen for me? You know, I'm excited about that stuff, what makes your jobs faster. And I do agree with you, there's always going to be functions within HR, whether you are an HRIS or not, that will continue to also persist and need humans. I mean, that's just the reality, right? It's also the reality of the work that we do.

‍

Dani DeMaio (20:26): Yeah, it's not doom and gloom. I just think, though, this is probably my tough love, but you just kind of have to get over it, and you have to strengthen your own skills. Like, job security is never in the company that you're working with. It's going to be in, like, what can you bring to an organization? And use it as a time to learn a different area, instead of being scared and freezing and saying, I don't know what to do with my life because I can't go into this field. Guess what? Every single field is facing this.

‍

Kristin McDonald (20:53): Yes. Yeah. I think carving out time just to play around. I think where so many people are starting from, it's like, where to start? I don't know what it means, I don't know what this is. And I think you just have to start, and you have to play around with it. You have to start getting yourself comfortable. Like, I carve out AI time to play around, whether that's in ChatGPT or Copilot, or take a course or whatever, just to continue. Because it is still moving very fast as well. And I only feel as comfortable with it now, which I'm very far from being an SME in any of that space, but I only feel comfortable with where I'm now because I've continued to play in it and see how to use it. And I think that's what people have to do. You just have to get into it.

‍

Dani DeMaio (21:31): It's like exposure therapy, right? If you are so afraid to do something... oh, okay, Logan hasn't been to as much therapy as I have. You get closer to your fears of things. So if you have a fear of roller coasters, go stand by a roller coaster, right? Get in line, act like you're gonna do it, so then you start to normalize your nervous system to it. That's what you have to do with AI, expose yourself to it.

‍

Logan (21:39): What is that?

‍

Kristin McDonald (22:00): Exposure... maybe that's the playbook, exposure therapy.

‍

Logan (22:00): I'm probably... could not imagine having to do that with snakes. For me personally, that seems very traumatizing. But anyway.

‍

Kristin McDonald (22:09): Oh, no, we're spiders. Mm-mm.

‍

Dani DeMaio (22:10): I'm afraid of them too. Go live in Texas. I used to step on them, like really close to stepping on them with running. And my fear still hasn't gone away. So I actually don't know how effective exposure therapy is.

‍

Logan (22:26): Well, that's good to know that everybody can feel a little good that it may not always work as much as you want. Good. I mean, staying AI curious... I've spoken with a few people and they've said, you know, just stay AI curious and keep going. And it's new. New tech is always that way. So, kind of a bit of a closing-ish question is: if HR leaders remember one thing, as we meandered through this from the AI conversation, as they continue into 2026, what should it be? And we can go one at a time. Maybe, Kristin, you can go first.

‍

Kristin McDonald (23:01): It's funny, when I think about AI and this conversation, I think HR is not new to being thrown things, right? So I think we need to just take a step back and remember that. Well, one, I always say this: give yourself grace. Like, we're all in the same place. We're all starting down this journey. We are all learning. We're all at different paths or parts of the path for that. But just start, I think, is the key. And start putting together that flight plan of, you know, building the plane. Like, what will it take? What are you looking to accomplish? Sit down with the business, or understand what your business goals are, and start with the business outcome you're trying to achieve. And not so much think about it as an AI conversation, but what are the outcomes we're trying to achieve? And like, where are we at? Where is the org ready for that, right?

‍

So it's kind of like the X and the Y, or greatness versus like the business problem. And I think it's just taking a step back to set those foundations, and maybe start simple, and remember that we're all learning together, and just continue to evolve and learn from that.

‍

Dani DeMaio (24:12): I like that. I'll leave them with what I told the head of IT the other day, as we're in the middle of a carve out, and it's: you cannot lose 40 pounds in 40 days, but you might at the end of the year, if you're not like the 60% that quits by January.

‍

It's so true, that I think what this is, it's really like task after task will compound, or initiative after initiative, that by the end of the year your department might be completely different. But it's not happening right now, even though we're getting the pressures of it. But it might happen by the end of the year, or the following year. Or maybe not. Maybe we'll just have this conversation again next year, that could happen too.

‍

Kristin McDonald (24:52): We had the same place. Exposure therapy didn't work.

‍

Dani DeMaio (24:55): It did work, we're back.

‍

Kristin McDonald (24:52): Well, I don't know why it's like this. We have this roadmap, but there's... what's the end? There is no end. I think it's, you know, it's the same thing with technology. HR technologies continue to evolve and change, and it's the same thing that's going to be with AI. There's no real end. It's just continuing to evolve with this new technology, this new tool.

‍

Dani DeMaio (24:55): There's no finish line.

‍

Logan (25:18): Well, great. That is great closing advice for everybody. Kristin, Dani, I appreciate you both taking time out of your day to chat with us about how we can better use AI in HR and adapt and keep this conversation going.

‍

Logan (25:35): Thank you for joining us for today's episode of Pulse, the podcast that has real conversations with HR leaders linking people to performance. HRBench is the people analytics platform that turns HR data into business strategy. It connects to all your tools: HRIS, ATS, payroll, surveys, whatever you're using. In just a few clicks, you get over 45 metrics instantly visualized in boardroom ready dashboards, benchmarked by industry and company size. Book a demo at HRBench.com, and we'll build you a custom benchmark report using your data.