Moving from Experimental Pilots to Proven CX Outcomes - Interviews from Pegaworld 2026

0:00So welcome to the next edition of the Punk CX podcast. This podcast is part one of a two-parter that features a series of chats that I had with Pega executives and one of their clients while at Pega World in Las Vegas recently. In part one of this episode, I talked with Ken Stillwell, CEO and CFO at Pega, and Peter Lacroix, head of low code at Achmia, a Dutch insurer, and also a longtime customer of Pega's. Now some of the things we cover include the big themes and takeaways from the event, the CFO's perspective on token maxing and tokenomics, Pega's response to all of that, Kamea's journey with Pega, how they're using various parts of the platform and also the impact on their customers and business.

0:42So let's get into the first conversation I had with Ken. Welcome to the Punk CX podcast. With me today, I have Ken Stilwell, who is the COO and the CFO of Pega. First time on the podcast. Hello. Welcome. How are you doing? I'm great. Thanks for having me. Thank you very much for being here. We're at Pega World, but I just wanted to say, I wanted to ask you to introduce yourself and tell me a little bit about the sort of the work that you do, because it's actually not common having a CFO and a COO kind of role. But tell me about how that kind of looks on a sort of a day-to-day basis. Sure. So I've been at Pega actually coming up on my 10-year anniversary.

1:17When I started, I had what I would maybe call a CFO plus type role where I had the CFO type functions, but I also had the chief information officer and the chief legal officer and some other functional areas. When we moved to the cloud, we made a big move to the cloud, which came shortly after I joined. And I had run parts of the cloud organization in my previous experience. And so it was a natural transition for me to take on some of the cloud and operational areas, kind of, and also some of the engineering areas that support the cloud. And so that's kind of when I became the COO and CFO. So I think now if you asked me what I, you know, how I viewed my role or what a week looked like for me, I would say I pretty much, you know, touch all the different parts of the business, right?

2:10If I'm talking to, you know, Krim around some roadmap decisions or whether I'm talking to our sales teams around how we're, you know, attacking new territories or whether it's my own teams that I'm working with, I kind of view myself as maybe a glue, so to speak, of putting all the pieces together. Okay. And so we're at Pegaworld. We're now on kind of towards the end of day two. The sort of the, I guess, people are starting to sort of wind down and sort of starting to, I guess, ruminate on some of the things that they've seen and they heard and stuff. I mean, it's been, there's been some great stuff.

2:44But I wanted to hear from your perspective, like what are the sort of big standout highlights for you? Well, I think there's one subtle one that actually wasn't really on the stage, but has been an observation I've had as I've talked to people across the conference, which is PegaWorld has really transformed over the years to be more of a business user audience, right? Like a lot more of the people here are business users, business owners, thinking about the problem, how to solve the problem. There are still lots of technologists here, but I do think that shift is noticeable. When you think about some of the messaging and some of the announcements we've put out, is really kind of what we're really trying to help clients do is using our world-class best-in-breed technology is help clients reimagine and re-envision how they will actually get work done and leveraging AI, using some of the best things that they've done over time and also using some of the new best practices.

3:42And the AI convergence there really hits some of the discussions we've had about tokenomics or the economics of that because, you know, one of the biggest challenges that I think our clients are seeing is how do I use this AI technology in the right way? Like how do I really get value? And I think it has been a largely experimental process here for the first part of the year. I want to dig into the tokenomics kind of bit because Alan riffed on the whole token maxing and tokenomics thing. And it's something that's a concern that I've been having around the sustainability, the economic sustainability of much of this sort of stuff.

4:18And I was also thinking about, and I wanted to get your perspective as a CFO, a bit like going, I've been thinking about it, it's a bit like putting your credit card behind the bar and then going, the bar's open, knock yourself out. And then as a CFO, you're like, I'm not sure how big that bill's in a million. How do I control that? So I I wanted to get your take on that issue that's in this, amongst this AI era that we're in. And coming from a CFO perspective, the reaction to it. Tell me what you're thinking about how you would pursue. Yeah. So this is an interesting one, but because it's a problem we've seen in other areas.

5:03Right. Like we've lived in our lives, for example, we have electric coming to our homes. We understand now, after decades, how to measure the cost of electric. And we understand the inputs and things that we would do that would, for example, if we leave all the lights on, if we turn the air conditioning down to 60 degrees, if we did, if we left the windows open to turn the air conditioning down. We understand there's a cost to that. We may not be able to mathematically calculate it, But we intuitively know that that's not really a thoughtful thing to do. For us to be surprised that when you spend$1.6 trillion on infrastructure to build out all the models, right, to think that those vendors shouldn't get paid for what they've invested in, that's silly for us to think they shouldn't be paid.

5:50So for that us to think that the usage that we have, which is theoretically unmeasured, Even they can't measure them in the way that we could with a meter on your house for electric. We need to understand that we are exposed to that. Now, in an experimental time, that's maybe just a cost of innovation and a cost of learning. But when you get to actually operating, that's where I, as a CFO, really draw the line. If you said to me, Ken, I don't know what I'm going to spend on tokens. It might be a million dollars or it might be two million dollars. But I've got to do it to be I would say well what how long is that gonna persist?

6:26Is that gonna be a month? Is that gonna be a year? Is it gonna be a likelihood? It's gonna be this or this and the worthy and there's a risk so there's a risk management piece of that now One of the ways I might do that is I might say well You know what I'll give you a credit card in the limits only ten thousand dollars and when you run out of time You know there's ways that you can't do that with the tokens right because you get you'll get the bill So I think there's a couple pieces to this. One is we shouldn't be surprised that there's cost to this. The second is in the experimental time, I'm probably more tolerant as a CFO to let people try things out a little bit because you want to encourage innovation.

7:00It's a strategic sort of imperative. And it's boxed in. And it's boxed in. But when you actually get to a point now where we're going to start to operationalize this, that's where I draw the line. I said, I want to understand what the ROI is. I want to understand what the impact is with the variability of cost. I'll accept some variability. We lived through this with cloud, right? We all as an industry, everybody can tell stories about when you would get the AWS bill that was three times what you thought it would be, right? Because there was not great visibility. But guess what? AWS created tooling, third parties created tooling, we all got smarter.

7:35Now there's really good granularity on what your usage is. And there's even techniques that even AWS will tell you. Here's how you manage your costs. There's how you buy reserved instances and how you read. So the market will evolve. But where we are right now, I believe it's just irresponsible to not have any idea what your costs are going to be and whether there's any value. And that's, I think, our key message. But then you've actually taken a very different approach. I mean, PEGA's gone, actually, we understand this, but we're like going, we're sort of like laying the track down and saying, this is the way.

8:10What we're trying to formulate is a structure that can be emulated around how you manage this. And one component, there's a few, one component is design time versus run time. There's a logical thing of when you're going to design something and you're going to do it once, you can spend the time and effort to basically ideate and get that right. When I'm operating 2 billion transactions every year, I don't want that to be creative. I want that to be automated. You're very big in that at all. I want it to be fast and cheap, right? And predictable. And predictable. So that's the second piece of it is when you're thinking about when you use AI, you want to think about what model you use.

8:52Because the frontier models versus some of the legacy models, there's a very big cost difference there. And there's a performance and there's availability and capacity. So you want to manage that. Then the last piece is, or another piece is, there's a convergence of the design and runtime, which is at runtime, there are ways to leverage AI. Things like where you might want to, you're going to call an agent or you're going to automate, you're going to use some type of a logic to do a call documentation write-up. Or you're going to send a letter to some, there's things that there is some level of judgment that isn't really part of that regulated workflow structure.

9:32So, I think that we're thinking about like where does it sit? How do you leverage the kind of the gray area, so to speak, around where AI is used in concert with something like a deterministic workflow? And then thinking about the models. If we're really trying to do at Pega, is we're trying to set a course for our clients to leverage AI, get all the benefit, but manage the actual financial economics so they can get the ROI. I mean, I think it's going to be really smart because actually what you're kind of doing is rather than going, here's a big, it's like whack-a-mole, kind of like with one big kind of like, here's a generative AI.

10:08We've taken a generative AI solution to everything. And you're like, no, there's a toolbox here. And this is a growing set of tools. And sometimes you need to pick the right tools for the right job. And it's like you're almost getting people to lean into it and go, we need for you to have a bit more of a sophisticated understanding of what the problem is and pick the right sort of kind of tool for the job. and the particular part of the job that you're at. And that is why one of the other messages, that is why we don't charge for tokens. Because what we do is we're so confident in our approach that we charge for the outcomes, right?

10:43There's an AI-assisted outcome price, and that is predictable, and that is known, and that is transparent. And we will build that system so that it will really help to manage the AI usage in a smart way, and therefore we don't need to charge for tokens. Yeah. And so I think the other thing I wanted to say is because almost to back up some of this, the troubles that people are having, you released some research, which is called, let me get it right, the truth about agentic AI, so success or non-success. I wonder if you can help me understand some of the, I mean, that's almost like speaks to the landscape, right?

11:19And some of the challenges that you're having, because actually you're saying, here's a big, big problem. Here's kind of our approach. Here's some of the kind of challenges. And you're almost like saying, well, this supports our hypothesis. And we're trying to be the solution as well. So what does the research kind of told you? I think one of the, once again, I don't know that I, I'm not sure that necessarily I'm surprised by what the research tells. But just like I'm not surprised that 1.6 trillion needs to actually have a return on investment. So that doesn't surprise me. I'm also not surprised that because of all the experiment, that without thinking about ROI that most executives have not been able to see ROI.

11:58ROI is elusive. I think that really just, what that kind of does to me is it confirms two things. One, AI is very powerful. And two, if you're going to get the ROI, you got to really be thoughtful about where you use it and how you use it. We shouldn't think that because ROI has been elusive that AI is not valuable. It's just valuable for certain purposes. That's where you need to think about using that. I kind of generically think about AI as anywhere where a human being was going to use judgment, was going to use analysis, was going to try to do something using their own mental and experience capacity.

12:32That's the perfect use case for AI. Because quite frankly, we can't even process data fast enough to make decisions in the speed that we want to. So that's a great when when when systems are built to take judgment away from humans It seems silly to then go use AI to undo those system to give judgment to AI now So the ERP system would be an example There's a reason why your piece very structured because you want repeatable transaction period so your auditors can come in and see that Yeah, so it seems to me it strikes me as we why would we undo something that actually works? Yeah, sure and but within all of that sort of stuff because we are in this You're absolutely right.

13:11I mean, it's like the nature of the evolution of the technology is incredible. It's like profound and it's exciting and it's kind of like offers incredible amounts of possibility. And it's kind of scary in parts. And we are almost like building the road that we're traveling on kind of as we're traveling down the road. Yeah. And particularly when we factor in automation and all those things and the changing nature of work. And a lot of people are like, I think are trying to figure that out. And I'd be interested to, as a CFO and a COO, is like, how are you at Pega thinking about that, the future of work and the redesign of work, using all these kind of tools?

13:53Because it's sort of changing kind of everything. So I have a perspective where I think about the nature of the way that things are done. And this highly is relevant in an engineering example. But there's someone sets a strategy. That's kind of one role. The next role is someone translates that strategy into how should we actually achieve that strategy. The next level is the people that typically are the functions that do the actual execution against that strategy. And another one is the kind of the quality of the validation of the back end. In engineering, it's very simple. I want to solve this problem.

14:29My product managers give me what conceptually how it needs to be done. The engineers actually do it and the testing group. But you could go across all the functions and there's a similar process there. So what I view is that do step, the execution is an opportunity for us to leverage AI because it can speed up some of the judgment calls that people make, say, in writing code or in reading a contract to decide how the revenue should be split. Things like there's judgment involved there. So if you think about that, I think it's going to push the real differentiated skills on people that understand how to architect the problem and understand how to make sure that this final solution actually does what it said it was going to do.

15:14So I think what's going to happen, organizations are going to flatten, management structures are going to flatten. You're going to have more people that are content rich and actually driving and leveraging the models to get efficiency. And a really important step is going to be the validation step on the back end. Because if you can't validate what the models are doing, what the agent's doing, you're really at risk to run with kind of work slop, as they call it, or hallucinations. I have a kind of a slightly humorous one, but a real life one that just happened last week, which is not a hallucination, but shows you what can happen if results are not validated.

15:52I was at a conference last week, an investor conference, and I was scheduled to do a fireside chat. The market was down for the day across tech and an AI agent produced a note and the note said, comments from CFO at fireside chat drive stock price down. But I had not done the fireside chat yet. So it was an AI generated. Now you think about that, you might say, oh, that was a hallucination. I would disagree. It knew I was going to be there. It knew that I was going to present. the stock went down it if it assumed it was not a bad assumption it just missed that the timing was off. It made a causative sort of like link where there was no...

16:34It made a probabilistic guess and it just happened to be wrong. What's the cost of that being wrong right? So like that you know that could be that could be a lawsuit because you know somebody put some note out that wasn't... That's a real-life example. If you're not validating the outputs, if you're not But making sure that AI does what it's supposed to do, you can get really bad outcomes. Yeah. I think that it feels like that's probably replicated in also what you talked about. I saw some of the announcements where you have, particularly with the Infinity platform, the new release, and you have the solution builders, and you've built out this solution engineers, which is to your point about the strategy and how we're going to actually, paired with kind of like to help the builders as well to accelerate that development sort of cycle.

17:21So it feels like it's happening in real time as well that you're almost like what you're thinking about in terms of how the work is getting restructured is it's starting to manifest itself in kind of the products and the kind of ecosystem. Have you ever seen that funny cartoon that shows it's like a playground? and I might not have it exactly right, but it basically shows like, you know, this is the path of the engineer and then this is the path of the user. And it shows like the engineer with this very sophisticated path where it like there's steps that go over a fence and down and around, but then the path of the user is there's just an opening in the fence so they just walk through.

18:01So like there's a gap between like what people envision. What we're doing is we're helping them close that gap because as you're ideating, as you're designing that, You see it right in front of you and you can actually prosecute how you want the work done How you want to design your process as you're actually building it that's with code that is that is impossible to do Yeah, because you're basically waiting and to see what the code produces and then you have to go back to and write code again Yeah, yeah, so maybe we'll kind of like make sure that you can take aim better than you're gonna be on it So couple of further questions.

18:34Well, the first one is what's coming up that you're most excited about? Oh, I think Pega26 and Infinity Studio and how you can take that blueprint experience from the sales and the ideation and the design into the actual build and over to the run and evolve is the biggest transformation that we've seen in the company's history. That overshadows everything for me. I think it's a completely different paradigm of how we want companies to really drive innovation and to evolve and redesign and rethink. And it allows them to do it in a context that is scalable. It reduces variability. And I think that's just a massive change for us.

19:17Awesome. And one final one. If I was to say to people listening into this, they go, we'd like to achieve success in terms of leveraging these AI tools out there to drive both better customer, employee, and general business outcomes, what would be Ken Stillwell's best advice? I think that it's fine to experiment with AI. I encourage it. I think you have to get to a point where you understand what problems you're trying to solve. And then you should use AI where appropriate to go after those biggest problems. And I think the worst mistake that companies make, and this is AI is just one of the examples, is they sometimes take the easy but unimportant things.

20:00Or they go after the things that are very complicated but work really well, and they try to make that better. And I think you've got to think about, like, are those really your biggest problems? Like, I think you should go after the biggest problems. And to me, if you anchor yourself running a business on, like, how do I improve? Where's my biggest opportunity? Or where's my biggest thing that's holding me back? And can AI help me with that? Awesome. Thank you. Awesome. Thank you. Now, I really enjoyed my chat with Ken. I particularly liked his CFO perspective on the whole tokenomics issue, as well as learning more about Peggy's response to that.

20:39Now, let's get into my chat with Peter Lacroix. Head of Low Code at Achmea, where we talk about Achmea's journey with Pega, how they're using their tools, and also the impact they're being able to drive across the business. So, welcome to the podcast. This is the Punk CCX podcast. I'm here at Pega World with Peter Lacroix. Yeah, pronounced perfectly. Thank you very much. Head of Low Code at Achmea. Welcome. Can I ask you to tell me a little bit about yourself and a bit about what Achmea can do, just to give us a bit of background? Yes. I'm an IT manager at Achmea at the Central Operations Department.

21:18And with my colleagues, we are responsible for running all our enterprise application platforms that we have defined in our enterprise architectural. And Pega is amongst one of them. We have 34 named platforms, and this is one of these platforms. platforms and we are because Pega is one of the more important platforms we utilize at the Lachmea. And you've got quite a long history with Pega. I mean, how many years is it since you started working with Pega? 14 years. 14 years? 14 years. Wow. So you've seen lots of developments. We've seen it evaluate from what it was to what it is now. Yes, exactly.

21:55Yeah. So can you tell me a bit more about sort of your journey with Pega and how that first came about, what was the impetus behind it, and where you've come to in your journey, where you use it in the business, and what sort of impact it's been having. I need to start with a little history lessons of Achmeyer first, if you don't. Love that. We are a financial service provider, but we are also a fusion company where we bought many other labels to join the Achmeyer family. And we have different labels, many different labels. We got rid of a couple of them and we have many legacy systems. And that was the time that internet came up and our IT board decided, well, we need to do something to prevent our employees working on those legacy systems.

22:46We need to give them a more modern environment so we can retrain them to more modern environments. And we have to see how we can get the knowledge that is on those old mainframes, how we can open them to other applications. That platform was Pega. Pega was chosen then as a fabric around those old legacy systems to make sure that via Pega we could connect to those systems and then do the maintenance we needed to do, do the transactions we needed to do without doing heavy maintenance on systems where nobody really knew what happened there. That's how it started 40 years ago to help us in the migration and the modernization of our infrastructure.

23:27And year after year that they developed and evaluated to what PEGA is right now. And a very important workflow application platform for us. We use the PEGA automation solution. We use the sales automation. We work on the CDH, the marketing machine, the next best action environment. And we have the PEGA robotics solutions as well. Okay. the VSTS add-on to make sure that our legacy systems can be handled with Pega Robotics as well. So we don't have the full stack, but we are a big user. You're a power user. We're a power user. And what sort of, I mean, because you're talking about it's more in the EFC, you're running a lot of your operations on the Pega platform, but then you're also using the marketing engagement tool around sort of comms and engagement, that one-to-one kind of personalization.

24:18I mean, have you seen any sort of, what sort of, give me a sense of the sort of impact it's had on the kind of the business in terms of the ROI, the improvement, or kind of like some of the metrics that you would kind of like look at in terms of we've done this and this is how it's impacted the kind of business. For the workflow and integration solutions, the more old-fashioned PECA solutions, not talking about CDA, it's just a workflow tool that does exactly what it should do. And it has a perfect audit trail, so it makes it more easy to find out who did what. From a compliance perspective, that's essential, yes?

24:55That's essential. We have a power band that's called Centraal Beheer. and Centraal Beheer finds themselves enormous innovative and that's based on the decision they made in the late 60s in the last century. We will be a call center only insurance company. Okay. At that time, that was a mind shift in the Netherlands. We have an insurance company that you can call and via the telephone you can get a call center that will help you with your claims or your products. So, in the 1960s? It's in the 1960s. Wow. And the claim was even Apeldoorn, Bella. Just call Apeldoorn and we'll fix it. So we tried ourselves with an enormously innovative action.

25:38And then we all get older, new generations get born. And I see that with my children as well. They use the telephone for everything, but not for giving me a telephone call how I'm doing. So they use it for all, but not for what it was invented for. Well, I think they actually talk to each other. They just don't talk to kind of older folk. All the folk, they talk to apps. They won't even tell me what apps they are using because then daddy can check and then I just move on to another app again. But that innovative mindset that got our Centraal Beheren as our most customer-faced brand to think about, well, if they don't use the telephone anymore, what can we think of then?

26:19So we went to the internet. We want to be a digital service provider. And then we decided, well, we need to think of a way to help our customers like the way the big retailers do that. Next best actions. Other customers like they bought this and what. And that's how our marketing engine, our decision hub got in. And it started just as a marketing engine. Right. And it was, well, it's your birthday. happy birthday Font Centraal Beer how are you doing have a nice day just practicing with Pittered Engine and what you see now it's evolved to a much more mature solution where all kinds of marketing campaigns can be programmed and that can be executed to our customer base just to give them some insights some feedback or do them a better solution than they have right now and from Pega perspective we are quite we are quite strange.