As organizations accelerate AI investment, the gap between adopting the technology and realizing measurable value is becoming a central HR challenge. The CIPD’s 2026 HR Pulse report found that 75% of people professionals had received no training to apply AI in their roles, while 22% said they had never used AI tools at work. The findings point to a practical readiness problem: technology investment needs to be matched by skills, process design and organizational capability. 

On RPI Tech Connect, host Chris Arey speaks with Stacey Harris, Chief Research Officer and Managing Partner at Sapient Insights Group, about findings from the organization’s 29th annual HR Systems Survey. Harris explains what separates satisfied HR technology customers from disappointed ones, why customer service remains a major factor in vendor satisfaction, and how process effectiveness and data integration influence the outcomes organizations get from AI.

In this episode, Harris shares survey findings drawn from more than 10,000 responses, cleaned and consolidated into data representing 6,038 individual companies. She discusses enterprise plans to replace payroll and other core HR systems, the relationship between AI readiness and business outcomes, and why organizations may need to rethink fragmented HR processes before investing further in intelligent tools.

This series was recorded live at HR Tech 2026 inside the HR Executive studios on the expo floor in partnership with the WRKdefined Podcast Network. 

⁠Make sure you’re subscribed to the full series⁠ and ⁠visit ⁠⁠HRExecutive.com⁠⁠ for the news, analysis, and insights shaping the future of work.

Show notes

What determines whether HR technology and AI investments deliver business value?

HR technology investments deliver stronger outcomes when organizations combine effective processes, integrated data and technology that fits the way customers expect to work. Harris argues that buying AI tools without improving the underlying operating environment often produces only incremental gains.

The discussion begins with customer satisfaction. Harris says the quality of customer care, both during implementation and after launch, repeatedly distinguishes satisfied users from dissatisfied ones. Buyers should speak directly with customers who use the product and understand how they experience the vendor's support model. The technology matters too, but different buyers value different service experiences.

The AI findings point to a larger organizational challenge. Companies that had invested in process improvement and data integration were much more likely to see substantial improvements in outcomes from AI. That requires difficult conversations about standardizing processes that may have been customized over many years. The goal is not simply to add another tool, but to create an operating foundation that allows technology to deliver value across the organization.

Key Takeaways

- Customer service can matter as much as the software itself. Harris says implementation support and ongoing customer care consistently influence satisfaction with HR technology vendors. She recommends speaking with customers who use the product and asking how they experience the vendor's service model. Some buyers want technology they can operate independently, while others value having someone guide them through questions before they even use the system. Buyers should assess the service experience they need, not just compare product features.

- Independent customer feedback gives buyers a different view of vendor performance. Harris explains that Sapient Insights Group's user experience and vendor satisfaction ratings are based on customer responses, not adjusted analyst opinions. The research also validates respondents to reduce the influence of bots or paid reviews. This year's survey received more than 10,000 responses, which were cleaned over two to three months and consolidated into 6,038 individual companies. The distinction matters because companies, rather than individual respondents, are the units making technology purchasing decisions.

- Enterprise organizations are reconsidering core HR and payroll systems. Harris says overall replacement plans declined slightly, with typical application replacement rates generally ranging from 10% to 15%. However, enterprise organizations with more than 5,000 employees showed increased plans to replace major systems. The share planning to replace payroll rose from 32% to 37%, while replacement plans for time management and HRMS also increased. These figures suggest that large organizations are beginning to reconsider complex systems that may have been difficult to replace in the past.

- AI readiness depends on process effectiveness and data integration. Harris says organizations that had invested in improving HR processes and integrating data saw much stronger outcome improvements from AI than those relying on automation alone. She describes a difference between a small improvement, such as a hypothetical 2% increase after investing millions, and the much larger gains observed among organizations with stronger foundations. Her research discussion points to improvements of 30 to 40 percentage points in overall outcomes for organizations that had focused on process improvement and data integration. The lesson is to establish the business outcome first, then assess whether processes and data can support it.

- AI changes workforce needs, but learning and development cannot be overlooked. Harris says organizations with AI in place over the previous two years were less likely to reduce their workforces in actual outcomes and more likely to increase hiring than organizations without AI. She also raises concern that learning and development dropped off the list of top HR technology investment priorities as intelligent tools moved up. New technology can change the work people do, but employees still need the skills to perform those roles. Organizations that reduce attention to training may leave their workforce unprepared for the changes they are introducing.

About Stacey Harris

Stacey Harris is Chief Research Officer and Managing Partner at Sapient Insights Group. She has been part of the HR practices and technology research community since 2007, producing research on enterprise HR technology, high-impact HR organizations and talent management practices. Previously, she held research leadership roles at Sierra-Cedar, Bersin & Associates and Brandon Hall Group. She oversees the Annual HR Systems Survey and is the author of Introduction to HR Technology: Understand How to Use Technology to Improve Performance and Processes.

Connect with Stacey on LinkedIn

Connect with WRKdefined

Website · LinkedIn · TikTok · Facebook · X · Substack




Powered by the WRKdefined Podcast Network. 

[00:00:00] Every organization buys HR software and expects to be happy with it. Some are, and others spend years wishing they had picked something else. If you've got only like a 2% outcome increase by putting in place AI, and you've invested millions in that, are you really getting the out, like you're actually behind the curveball then? If you really want to understand what this technology is going to be like inside your organization, talk to someone who is in their customer service area,

[00:00:27] or talk to someone who is using the product who's like you and how they're feeling about that customer service. We're now the work architects and HR has to really rearrange their thinking and the fact that we're not decreasing or increasing the workforce, we're rethinking what it can accomplish.

[00:00:53] Every organization buys HR software and expects to be happy with it. Some are, and others spend years wishing they had picked something else. Today on a very special episode of RPI Tech Connect, we're digging into what separates those two groups. My name is Chris Arey and I'm the host of RPI Tech Connect and today I'm joined by Stacey Harris, Chief Research Officer and Managing Partner of Sapient Insights Group.

[00:01:19] And her organization has spent a lot of time asking those questions, finding out what drives success, what drives adoption, and ultimately satisfaction with HR systems. They're getting ready to release their 29th annual research report. Yeah. November, is that right? Yeah, November 6th, yeah. And today we're going to be talking a little bit about what some of the findings are from that report. So Stacey, thank you so much for joining me today.

[00:01:45] I appreciate you having me. I mean, we just launched some of the key findings today to a room of, I think there was 400 or 500 in that room, so it was pretty cool. So you are getting it fresh off the, right after the big session, so I'm very excited to talk about it. I have to say, I sat in on that session and I had put together my notes and then after the session I had to go back and make some more notes. So I've got those questions in here too. But at a high level, what does the report say separates those two groups of organizations?

[00:02:15] Those that are satisfied with their HR systems and those that aren't. Is there a... Oh, it's always a tough conversation, right? And I think it's very distinct for each vendor in some case, right? But I will say sort of the biggest difference we see over and over again. There's no doubt that customer care, whether that's how it's implemented or how it's maintained afterwards, right? Like that customer service component of it makes such a big difference. And the one thing I called out on stage was we have many times named individuals on the responses that say, like, you know, Sharon was amazing.

[00:02:45] She's the best customer service person we've ever had, right? And that always warms my heart because I know how hard those professionals work, right? So I always tell people, like, if you really want to understand what this technology is going to be like inside your organization, talk to someone who is in their customer service area or talk to someone who is using the product who's like you and how they're feeling about that customer service. That's a really good point. I'm glad you mentioned it. I follow up question that. It seems then that, like, it's very rarely actually about the software itself.

[00:03:15] Yeah, I mean, I do think there are – so, for example, I mean, a really good – we do see, and I'm going to name some names. We do not support any vendor. This is all based off of the data and information, right? So we know that, for example, Rippling's customers really love the AI onboarding. They love that they can just pick it up, go with it, run with it. As long as when they're done, they can get a hold of somebody at the end of it, right? And so I do think sometimes the technology makes a difference depending on what you want from the service model, right?

[00:03:44] On other organizations, like I solved who won one of the number one – sort of total number of number one user experience of understand awards this year. For them, we know over and over again it is that they can get someone who will walk them through their question before they even have to go to the system to check it out, right? So I think – and that's two very different buyers and two very different expectations, right? The difference between Gartner Magic Quadrant and Sapient Insights HR Systems Report.

[00:04:14] That's a really tough question. I started – you know, I started out by saying basically that all analysts, I think, do their best to be as unbiased and as neutral as possible. And I think – so that's just off the mark there, right? Right. But I do think that – I know Gartner uses in their Magic Quadrant some level of customer feedback, but they also have the Gartner opinion, right? Like it is the analyst opinion. And I think it's very valid.

[00:04:39] We – our research, at least on our quadrants, if you want to call them that, our user experience and vendor satisfaction awards, they're all customer, all voice the customer ratings. I don't touch those. I don't give my opinion about that. I don't actually adjust that number at all for anything. And to be honest, a lot of our data – I think there's also this assumption when you do research like this that, oh, it has to be an HRMS with all these features and all these same capabilities to be able to be evaluated the same. And we don't kind of look at it that way.

[00:05:07] We really look at it is that if you're – as a buyer, you make a decision about the fact that this is my HRMS, you know, even if it's a light HRMS or it's not quite the same recruiting platform as someone else, we're looking kind of more at wallet share, right? Like – and you're voting with your wallet and you're voting with your user experience and you're voting with your vendor satisfaction ratings. So it's very voice of the customer. It's very much focused on how the customer sees that product and where they put that product in their tech stack.

[00:05:34] Then we will come behind that and write some commentary in our report about why we think that might be. Sometimes it's because of the comments that they've written. We've got over, like I said, 13,000 comments this year, right? Wow. Sometimes it's just because we know something that the organization launched or did that makes it sort of like, oh, well, we know they launched this new AI tool and all of a sudden their numbers went up. So there's a connection there, we believe, right?

[00:05:59] So I just think that the difference is that we really start from the customer's perspective, not from an analyst's perspective, right? I really like that. And, like, you're truly independent by just collecting feedback from people who are using these tools. Yeah. And that's what you're compiling. And we also make sure that it's real people answering that. And they're not bots. And, I mean, there's some of that that happens in the market, right? I'll pay you $60 to say nice things about us, right? Where you can't really get away with that in our survey because we require minimums and we require that we know the people who respond. We can look them up.

[00:06:29] They're a real human being. So there's a lot of that, too, that we do. We do a lot of cleaning. And so the report that's getting ready to come out in November, how many respondents? So we had over 10,000 people who responded to the survey. Wow. And then we spent about two to three months cleaning the data. And basically that ended up this year with 6,038 individual companies. So when you think about survey data off time, you hear, well, we had 800 respondents or 2,000 respondents.

[00:06:55] But then when you clean that and when you really take that down to an individual company, and individual companies are what drive the market. Individual companies buy technology. It's not a person, right? Basically, that's how we run our data and run our numbers. That's why our numbers come so close to what actually happens in the market a lot of times. Love to hear that. And so your survey also asks organizations what they plan to replace, right? Was there anything in this year's findings that surprised you or maybe? Yeah.

[00:07:22] And actually, I didn't even get a chance to even talk about it up on stage because there was too many things I was covering this year. But we saw replacements actually go down a little bit. So the percentage overall went down in our overall numbers. So we saw sort of like replacements on average run anywhere between 10% to 15% in the market. So that's kind of an expected like if I have a recruiting platform about 10% to 15%, we're going to be replaced every year, right? But what our data shows this year is that that went down slightly as an average.

[00:07:49] What didn't go down and what increased dramatically is enterprise organizations over 5,000 employees, particularly payroll. Payroll went from like 32% of enterprise organizations planning to replace their HR payroll solution to 37%. So that went up. And time management and HRMS also went up in enterprise replacement plans. So the really big, complex, messy stuff is on the board now for being replaced.

[00:08:18] And I think that hits every one of the big, large vendors, right? Yeah. Those are spaces that we operate. So that's always nice to hear. Yeah. You talked something about timing of AI investment during your presentation. How does an organization determine what's right for them? It's like you're like there's a – I think you mentioned something about like investing so much and then not getting the return they expected. What does that formula look like? Yeah. So it's funny because when we – I run a lot of data. And I'll be honest.

[00:08:48] I'm not – sometimes I'm like, oh, I have so much AI. If I run the data, can I get an outcome that tells me that AI is actually not giving you what everybody thinks it should, right? That would be the great marketing spiel that, you know, our team wants to put out and flash is like, AI doesn't give you anything. And I'm like, no, unfortunately, there is an uptick. Even if you do nothing, you get a little bit of an uptick, right?

[00:09:12] But what – from my perspective is that if you've got only like a 2% outcome increased by putting in place AI and you've invested millions in that, are you really getting the outcome, right? Like you're actually behind the curveball then, right? So there's that. But really what we found, and our data was pretty clear on this, is that it is your HR process effectiveness and efficiency metric. So basically what we call AI readiness and your data integration and data capabilities.

[00:09:41] Those two things together made all the difference in those organizations who went from, eh, we get a little kind of bump up from doing AI because it does more automation to – we actually saw a full percentage, a full sort of like 30 or 40 percentage point jump in our overall outcomes. And that all came from those organizations who had really focused on process improvement over the last couple of years and now they were taking advantage of it. And really focused on that data integration conversation.

[00:10:08] I know it's not sexy and it's definitely not what anybody wants to hear because it's the hard stuff. I'm glad you mentioned it. It's been a recurring theme that I've – in the sessions that I've sat in in the last two days, so much has been emphasizing the importance of like before you even entertain or like think about the tool that you want, start with what the outcome that you're trying to solve for. Yeah. And so it's interesting to hear like the data – it always comes back to data, right? If your data is bad, your foundation is bad. The AI is just going to like automate to bad, right?

[00:10:37] You're not going to solve anything. It is. And I think the other side – because it sounds simple, right? But I think the other side of that conversation is that we as an industry, HR in general, have been over the last 10 years very focused on being the service engine to our organizations, to making sure that if they needed a certain type of HR process or they needed a certain type of talent approach or they needed a certain sort of performance approach, right? We fixed it. We figured out how to do it with a lot of different point solutions. We put together manual spreadsheets.

[00:11:06] We made it work because our job to be strategic was to meet what the businesses needed. Well, at this point in time, what that means is we might have to step back and say it isn't as important that you get this very specialized process or that you get this really focused thing on this one particular sales function in the organization that's doing really, really important work. Now I might actually just make sure that I have standard processes across everyone because then AI can actually look across all that information and give us more insight into what's happening in the organization. And that's a hard conversation to have with any operations leader, right?

[00:11:36] It is. You think about all the different parties that are involved in that area. It's not just a... Right? Like, I know you love your onboarding and we've spent millions on that onboarding and it has all these great cool things in it and a lot of little processes and a lot of little manual environments. But it's now what's restricting us from actually getting an out from the AI you want to spend more money on, right? You have to do that unsexy work, as you said. That's where it all starts. Two more questions for you here.

[00:12:06] During your presentation, I think you had a finding that said, if you implemented AI in the last two years, you are less likely to decrease your workforce. That is a powerful, big statement with a lot of stuff underneath it. What's the story there? Yeah. So, you know, we've been watching the numbers for a while on are you planning to increase or decrease your overall workforce?

[00:12:29] And every year we see organizations over the last couple of years kind of say that about 50% of them, 54% of them plan to increase their workforce and just 6% or 7% plan to decrease their workforce. When it gets to the actual next year, what we find is it ends up a little bit more like 44% plan to actually do the increase of their workforce. And we're more like 22, 23% now who actually do a decrease. Those numbers have been pretty similar for the last five years. They haven't changed a whole lot.

[00:12:54] People are much less aware of the fact that they're going to do decreases and much more optimistic about how many they're going to increase. So that's just a fact. But what we do know is that in the actuals, not in the planning. So in the planning, it's about the same. But once I get to the actuals, people who have AI over the last two years and can show they've had AI over the last two years, they are less likely to actually do the decreases than their counterparts, right, who don't have AI in place.

[00:13:22] And they're more likely, which is actually the more interesting thing to me, more likely to hire more. And that's what I'm really excited about because I think that we've been telling the whole world that AI is going to take all these jobs. It creates as much work as it decreases, right? It's just different types of work. It's different types of work, yeah. We're now the work architects. And HR has to really rearrange their thinking in the fact that we're not decreasing or increasing the workforce. We're rethinking what it can accomplish.

[00:13:52] On the other side of that, there is this training and development going down as a whole, it seems, which is concerning. Like, roles are shifting. AI is creating new work pathways. But, like, why aren't we training people on how to do those things? Yeah. I really expected to see more training and development dollars, right? Like, training and development was one of the top five investments over the last couple of years for a while, right?

[00:14:19] But when AI shifted to the number one place where people were going to increase their HR technology spend, and we just have – we call it intelligent tools. That's the category that we give it. The one thing that came off the list, it wasn't recruiting. It wasn't benefits, which has been up there for a long time now. And it wasn't the HR mess or analytics, which is depending on whether you're enterprise or mid-market. It was learning that dropped off the list. And I was like, wow, that's just – I don't know why.

[00:14:45] Now, granted, some of those intelligent tools could be a learning solution, and that might be part of what we're seeing is that people are thinking, look, I can do more with AI learning than I can with a standalone learning tech, right? Sure. But I think what's really happening is we think, oh, we're going to decrease the workforce or we're going to reduce the amount of entry-level workers in our workforce, so I won't need as much learning and development.

[00:15:10] I'm really going to need more focus on recruiting the right kind of qualified people, and I'm going to need more focus on making sure I have the right benefits to keep them. But learning doesn't seem to be at the top priority. And it's – I think that's concerning. I think that's the wrong thing. But, you know, the market swings to where they think it's going to happen, and then they come back sometimes, right? Very true.

[00:15:29] Do you think that maybe the explanation for some of the drop-off in training investment is that AI is, you know, empowering employees to be more self-sufficient and they can kind of, like, teach themselves? Teach themselves. That could all – that's part of it, right? It could be some of that. And I don't disagree with that. I think the challenge becomes you're getting the same issue that the Internet used to train – could train the world, right? Wait, wait, wait. Is it our processes? Is it the way we want you to see it?

[00:15:58] Is it the policies that we need you to be aware of, right? There's some real risk there, right? There is. There absolutely is. I appreciate you saying that. Well, we're getting close to time. I want to thank you so much for hanging out with me this afternoon. I do believe that, for those of you listening in, we do have a discount code for the research. Yeah, so I think my team will give it to you and we'll have it at the – depending on where you guys post this at.

[00:16:23] So – or you can go to our website and we'll have a – say you listen to the podcast, one of those things, right? And then we'll have a discount code for 25% off the report, which is a pretty big chunk, so. Let's go. Thank you so much. Yeah. And for those of you listening in, we'll catch up with you next time. This is RPI Connect. Thanks, everyone. Bye.