PegaWorld 2025: Bovemij & ING Accelerating Enterprise Value: How Leading COEs Drive Continuous Innovation with Pega

PegaWorld 2025: Bovemij & ING – Accelerating Enterprise Value: How Leading COEs Drive Continuous Innovation with Pega

All right. Good afternoon everyone. Thank you. Thank you for joining us for our COE panel. My name is Alec Taylor. I'm the senior vice president of consulting at Pegasystems. Uh, and super excited to be moderating this panel today on COEs Uh, it's an interesting time for Centers of Excellence.

Uh, it's a time when we're entering this era of AI and Agentic, and I'm delighted to have a couple of folks with me today who are going to talk about their experiences with CEOs and tell us more about where their CEOs are going. Uh, two great things always happen in June. PegaWorld and Pride month. And I wanted to wish everybody a Happy Pride Month this year. Pega pride theme is vibrant fabric.

And what a great segue into a conversation about Coes, because when we combine all of the threads that we all bring together, we can do amazing things. And Pega absolutely believes that innovation is best when it is inclusive. So thank you again for joining um co Coes. Historically a centralizing function, providing expertise, best practices, governance, ensuring the kind of orderly adoption of technology in an organization. Coe is in an era of AI really about how they accelerate innovation.

And that's what we're going to get into today with our our two panelists. Now, maybe just to kind of give us some context for all of you. We've already established that most of you have come today because your friends and family of Tim and Lex. So if your friends and family of Tim and Lex, let's let's give them a shout out. Let's give them the woo hoo! Okay. Significant. I didn't I didn't realize these gentlemen had significant fan following. This is good.

Um, how many of you have COE in your organizations? Okay. A good number of you. How many of you don't have CEO's but are thinking about whether or not you need them? All right, we got a couple of folks. Fantastic. How many of you are thinking about where your Co is going next? All right. So we've got three topics we're going to cover today. We've also got Mike stands up here at the front. We would like to keep it interactive. I'll moderate the discussion.

But you know the moment something strikes you that you want to talk more about feel free to jump up, grab a mic and we'll keep it conversational. Or we can keep the questions to the end. So with that, I'm going to have these two gentlemen introduce themselves and we'll get rolling. So, Tim, why don't you kick us off? Thank you Alec. So I'm Tim Imls. I work for a wholesale bank, ING, based in Amsterdam, currently working on all the digital delivery within within the scope of wholesale.

And that means 25 squats on Pega Other. Which gives a challenge, of course, to keep them all structured and uniform building on the platform. And that's also the reason why we have done a Co, and I'm really proud that I'm sitting next to Lex as he is a kind of mentor for me. He gave me the example how to do it and I implemented it. And maybe we can deep dive a little bit more later on. Uh, because I did. I picked up the blueprint from Lex, but I couldn't put it in the same way. Chambray.

So there is always an exception or a difference per company that you have to tweak a little bit left and right. But one of the big learnings is don't invent the wheel yourself. Please talk with others, other companies, other colleagues, maybe within your company. We have experience within the Co and learn from them. Yeah, absolutely. Great point. So so to passing it over to the mentors. So Lex. Lex why don't you introduce yourself. Yeah.

So just to complement that, I did the same when I started this five, six years ago. Um, but Lex Ruijter worked for Bovemij. Insurance. Um, before that. That's where most of you know me from, is Lisbon. Um, there. I ran the COE for four years, I guess. Um, I. Live in Harlem, close to Amsterdam. Um, and so you had a question about what insight would you want to kick off with? I think, um, who does some wave surfing here in the audience? Anybody? One, two.

So, you know, when you go out the waves and you have that first moment, so you're going to spot them first and you're going to choose the one you want to take, and then you need to start turning around. And you need to start paddling like crazy. When you see that big wave coming, and then you need to take it and then show off all your skills or fall in the water. That's the little bit that I think the topic that we're touching on here is really interesting.

When you look at, um, innovation today, if you just look at one year past PegaWorld last year until today, the technology change and the enablement, what it gives all of us is just flabbergasting. It's going so rapid. If you translate that back to that wave, I think we as a sea, we need to also prepare not just our Pega community, but also departments like legal, compliance, HR to be ready to to surf that wave.

And it doesn't mean you can do that and do a nice I example in Pega because you're not going to get it to production because legal is going to think something. Compliance. HR you don't have the right people. You don't have the knowledge in the House. So I think just to kick off that from my perspective as a mindset, it's massive. Internet was great. But this is, I think times ten on impact on organizations, what it will bring with us. Yeah.

And we I, we we're totally going to get into that because I think you're right. I mean, we, you know, think big, start small. Figure out how you get all the pieces aligned. But why don't why don't we go back in time for a second and just talk about when each of you knew that your COE had kind of gone from a centralizing governance function, you know, maybe being seen as a little bit of a policing organization to something that was truly creating value. So.

So for Tim, for you at ING, when, when did you know that you guys had gone from governance to value creation as a CCO? And how did you how did you do that? Yeah, the COE is only here for one and a half years, but in that time span we've already found out. So and it always starts with that governance foundation that never goes away. So you still have to do the quality. You still have to do the IT risk part that will be there forever, even with AI. So that's the basis of your CCO.

But after that you get to know, ah, there is much more out there than only doing risk or doing, uh, checks and balances which has been built. So then you're going into, okay, what kind of AI capabilities are out there? Are Blueprint it came out last year, so we thought, let's adopt it first in the COE. We have Lsas and Bas in the COE sitting with me, and they bring it to all those departments and we call it tribes in ING.

And they give them the training, the learnings, the best practices about the Blueprint and then it gets adopted. And then you see great things, because we started with design sessions six years ago with Pega, and it took 3 or 4 weeks in total dedicated. And now we are back to three days. That's that's interesting. I mean, it's almost like you. It's almost like a Center-out COE, Tim, where there's a foundational layer of things that you're making sure are in place so that you can be successful. You're taking the new technologies or the new capabilities. You're trying those out, and then you're taking those out into the organization so that they can drive innovation back in from the edge. Yeah, because as a CEO, you are you have the experts, right? So when new capabilities come in, they don't know it yet. So first we have to learn it ourselves, have the basis and then distributed amongst the rest of the of the company. Yeah.

We should put a pin in that because that may come back when we talk about AI and Agentic and how Coes might play in that. Now, Lex, what about you? And it can be Bovemij or it can be leaseplan. You know, when did you know that your CEO was was adding value? And how did you get to that point? The same what Tim said. I think for sure, you need to get those foundational points and those checks and balances in place.

But I think the reusability, um, to build a reusable component takes you three months, but if the next one uses it, it's for free and it's straightaway available. And so in the beginning, you don't see that spin off that quick. But after a while, when you have like 20 reusable assets, you can see that it speeds up really fast. Um, but I think also with I coming to the table. Imagine you have ten teams and ten teams are working in a different way with different quality levels.

And now you're going to add some new innovation. Let's say GenAI an example. Um, now to get all these teams adapting it, adopting it in the same way, in the same speed, you're going to have a challenge because one is here, the other is there. And if you can, from a COE basic point of view, like you mentioned, if you have that foundation, everybody is working in a consistent and predictable way, adding a new innovation.

It's much easier because then you have everybody speaking the same language and you have processes in place. And I think the checks and balances with Lisbon, we started with pretty rigidly the police side of the story, but we wanted to move to the enabling side. So let's first make sure everybody walks in those corridors and then we want to start helping them. Um, but I think it is more about enabling and getting the people to, to learn and become enthusiastic about it.

Yeah, I think the best moment in Lisbon was when the crew were asking 100 people asking you, when are you going to go to the next release? And it was going to be next week. And that that pull from the developer saying, I want the new release from Pega, that's the magic moment. I thought, now we're winning. Now let's let's get into that a little bit and we can talk a little bit about how you got to that point.

And, Tim, we can talk about, you know, what you've told me is a pretty expansive library of reusable components. Tell me about those early days where, I mean, you you you inherited a situation at least Blend where that foundation was not strong. It doesn't sound like the right component parts were there.

Tell us a little bit about the journey of what you put in place, why you put those things in place, and how you eventually got to a point where, you know, you had created this demand and you were satisfying that demand in near real time. Yeah. So at least from Bovemij to worlds in Lisbon, we started it because we we were seeing the business requests for new projects was really high. And we saw already that there was silos starting to build up and they were building the same thing.

And so that's where we pulled the handbrake and said, well, stop, start rebuilding or building a reusable component in Bovemij. I think the reusable asset sounds really easy. It's an easy word, but it's more to that. So you need to be able from App Studio select it and then reuse it as a as a recallable module. It's not an interface. It's not a call to some chamber of commerce. It's a it's a module which really does something.

And you can select it from App Studio because then your business architect with the business in the DCO session can just click the whole system together with all these views of assets available and sometimes they don't work. You need governance on the reusable. You have that already set up. Is that where if there's a change on the reusable asset, you need versioning. You need to have all these teams be aware that there's a change coming. Be careful. Retest it.

Um, but I think that's that's where you see the real spin off start because, um, having six teams do the same, get customer data. It's a simple thing. Every company will see the same thing, get your customer information. Or if you get all these teams building it or you build it once. And how did you organize yourselves on reusable components is. Hold on a second, Tim. I'll be the moderator. Sorry, you can't resist. No. Go ahead, go ahead. No.

So on reusable components, I have a dedicated squad working on reusable components and I know they also have a backlog. And then the contribution model kicks in. Um, is there a difference between Bovemij and Leaseplan? Yeah, I personally we thought about centralizing it in Lisbon, but it's, uh, what we created there. What we what we were seeing happening is that you have too much handovers. So when you're in a product team, you basically know what you want.

And your business requirement, the outcome you want to accomplish is really clear in that team. So if I need to, um, if I get a new requirement to get, uh, whatever Chamber of Commerce information, and then if I need to hand that over to another team, I need to do a knowledge transfer. And I have all these communication to check if it's right and if I can keep the knowledge where it's where the demand is, then I can accelerate just by the business side, effort wise and governance.

I would definitely see if you can centralize it because that's where you get the numbers counting. And from a governance point of view, never decentralize that. But for example, if I have an example, an existing module today and I get a request for changing it, then the governance is with the COE. But the effort is in the new team. So the person who wants something to change, they change it. But under governance and guidance from COE. Got it. Okay.

So you're keeping it because I was going to ask you because you come in, you see that everybody is doing the same things in their silos. You say, well, there's a better way to do this, centralize it. You create a robust library of reusable assets and you can literally just start putting together applications. There's got to be a point at which, you know, you've got to stop people. I'm an artiste. I mean, like I want to I want to build it.

Like, how do you how do you govern that part of it where you're driving kind of adherence to. No, no, the way we work is about time to value. It's about out of box. It's about using the components we've got. Did you guys ever run into situations where you. Had definitely in the beginning, nobody wanted to do it because I have my timeline, my budget. And why does the CEO want to say something about what my agenda is?

But after I think the first 5 or 6 were built, then they saw the advantage because now they had a new requirement and they could just pick it from the from the App Studio and it was done. That saved them three months of work. And then you see that the momentum starts. But from a the business and the product owners, they're the mostly they say, oh forget them. Just build it. And then in the technical review we say you're building something which is nonsense because we have it.

And then you have the governing and the police role that you need to say. But then you're always kind of focused on showing them that you can provide value. I think it's a learning curve for me. The starting point was totally different. I started off with a centralized team, so all the squads were reporting into me, and then there was a scale up from five squads to 25. And then I said, okay, I don't handle that anymore on my own. I will push them away.

But the starting point was that the mindset of those squads was all the same. And loving the usability, loving the governance we had and that was spread around. So that was really powerful. And since then, I don't have that much of, uh, discussions on, on the, on the governance or the uniformity. Right. Got it. Very good. So, I mean, it's so many similarities between what both of you did. Right.

Like centralizing coordinating function that sees all of these, um, you know, kind of duplication of effort across the organization, centralizing that, winning those battles. Of overcoming kind of hey, use the library by showing that you can get to value faster, you can get more throughput.

Um, and then once you kind of have that mindset and that capability starting to move it back out to the edges of the organization so they're more self-sufficient while the Coe continues to introduce new capabilities. Um, probably a good segue to how you guys are thinking about AI and Agentic and what that means and what the role of the Coe is going forward. I mean, tons of hype, tons of executive interest in this. Everybody wants it. Yesterday we saw a bunch of really cool stuff this morning.

That's real as well. Um, but, you know, how are you guys thinking about the Coe in an era of of AI? Shall I start? Yeah. Yeah, yeah. So it's really exciting to be honest. But there's so much out there. So it's also difficult to choose what's best for us. So I think it always starts like you do with new technology you first discover. So first see what is a match for us for ING to use.

And then from there, if we start off, however, it's not only Pega who's offering all those nice capabilities, all the other technologies within ING, they have the same. So it's a kind of web where you have to navigate through. And as a CEO, you play a crucial role there to help all those parties and already talked about talked about it, legal, compliance, all those stuff divisions.

You have to take them and hold their hands and help them getting the same mindset at the end because that wave is coming. And if that wave is is at the doorstep, then you're too late. And I think every big company has to invest a lot of time in AI now. Yeah. So I. Two follow up questions on that. I mean, you must have really liked Allen's chess analogy this morning. Yeah. You know. Right. Right. Tool for the right thing.

Are you finding that you're kind of collaborating with a different set of stakeholders as you think about AI? You talked about legal. You talked about compliance. I don't imagine you dealt with them as much historically. No, no. The main focus is more on operations because my Co sits in operations, although I deal with lending, financial markets, trade and all those other business lines. But it's more of getting the efficiency and the cost reductions.

So then we look at okay what capability fits best to reduce those costs. As soon as possible. Right. And now increasingly kind of bringing in those lenses of Compliance and traceability and and all those things. For a highly regulated industry like it, it will require some different considerations and competencies on. Yeah, you have some hard discussions with them. And yeah, of course there are a lot of bureaucracy out there filling a lot of paperwork.

But, but, but but in the end I feel that's my role as a front running CEO, that I have to push it through and get get everyone on board. Yes, I love that. A front front running co, I love that. Now what about you? What's what's next in an era of AI? Yeah I think like my starting sentence was also I think the impact is so much bigger than Pega. It's organization wide. You need to when we're having a case about claims, it's just one of the departments.

But if you think about what the impact is, you need to look at the organization of the department. You don't have 40 people anymore. Maybe in the future, maybe have a different scenario. Um, and also from a sponsorship point of view, I think CEOs need to lift up. You need to have a direct line with the strategy and the vision of the company to have that short line and that sponsorship. When you go to a legal department, that they are also on the same page that it's you can't say no anymore.

It's not a matter of, no, we don't like that. It's dangerous security, compliance. Let's not do it. It's the question how can you? So also sponsorship needs to say like a CEO needs to go to the head of legal, say, well, we know it's tricky, but come up with reasons, ways how you can do it. And that sponsorship is something that enables us to do our job because you can push a lot, but at a certain point you will be blocked from every corner of the company.

You need that sponsorship to move forward, and I think it's like it's so that's the organization side. But I think also from the process side, every company has like um, a funnel, how to take a new initiative, a new idea somewhere pops up in the organization, whether it's a burning fire, whether a new business idea, and it goes through a funnel of prioritization.

And it's natural that a bigger team, if they get a question to do a new, um, claims process or the Blueprint claims and they follow the same pattern their mind is taught them in the past years, and then they'll have a you open the claim, you validate it, you get a pricing. So your mind is set to do that in stages and steps in Pega. And then great, everybody's happy. It's three months you go live. And what I think we should start doing is when the idea pops up, the first design is ready. Stop.

Now what? What could we do with agents? What could we do with if we put AI on certain parts of that process? And how can we make it even more fit for purpose for this use case? Because I think it will be mind blowing how much things you can now already take out of the the step side and put agents behind it. Absolutely. Yeah.

And I mean, we as we prepared for this panel, that was one of the things we were talking about is agentic and and agents across kind of our workflows, um, and the ability and Blueprint to kind of put in those agentic moments. Yeah. Um, but we've we've got a question before we get to the question, just to kind of summarize what you guys are saying going forward. Um, different set of considerations, different set of stakeholders, really important to have executive sponsorship.

Um, and to the riding of the wave analogy, if you think you're going to just show up on a Saturday morning and surf a 20 foot wave, it ain't going to happen. You know, start small, start building up your capability so that you're ready. But we're going to go to an audience question now this is fantastic. We actually said success for us is if we got a question during our presentation. Tick in the box. So we're very excited. Hi. Glad to make you successful here. Um, my name is Jeff.

I'm from Intuit. So, um, we're pretty new at integrating with Pega, and, um, we've been on the AI track for a while now. Um, as a matter of fact, we're building AI. Everything's AI, native AI first. Um, so much so that now that we're looking at integrating the Pega, we don't really have a center of excellence built up yet, but we're having to understand how we can use Intuit's responsible AI guidelines and our own guardrails that are in place without necessarily utilizing Pega's AI.

So we have to think about how we use our language models, our integrations first, but also everyone's you know, we have what's called global engineering days. A hundred different Agentic AI solutions came out. They're all very flashy. Everyone's very excited about them. And they say, we want this in Pega now.

But how do how do you look at building an orchestration layer between your center of excellence and making sure that you're integrating this smart from the beginning, and not running into those problems with everyone having their own solution? To me. Sorry, I. Didn't. Either one of you. I think, um, yeah. Pega uses their own set of models behind. Um, but if you look at under the carpet, it's basically calling a model with an API.

So you could basically do any model you want and you could configure that in a different way. Um, I think staying out of the box brings your time to market easier, because if you go to that level, then you. Yeah, you have a new layer in between where you could even say for if you have a chat bot, you can have a question. A could better go to that model. And question B could better go to that model. So that but I think from a SEO point of view I think it's.

Like somewhere ten minutes ago we said, I think if you have the foundation ready that everything is consistent. And you know what, how Pega projects are like 90% in your organization set up, then I think that question is very easy because you have a consistent design.

Your Layer Cake is well set up so you know where to position the model, um, like reusable assets in the Layer Cake in our in the Layer Cake sense, it's above the enterprise and you could put it one layer up, but I think there's a separate repository of our ML models. And you have to find a way how to call them.

So I think looking at it from a COE how to put that architecture down first time right is going to save you loads of time afterwards, because otherwise you're going to get the same thing in a siloed project. I'm just going to choose what I know, and I'm going to choose what sounds really logical to do today. But it won't look at the enterprise how to do it in a big picture. Does that answer the question?

Yeah, I think, um, part of the way there, I think what we're trying to understand is we don't have that center of excellence. So, you know, how do you tame everyone? Like, hold on. I know this all looks really exciting, but, you know, we're we're not even adding to it allowed right now to start teams. Do you have. Um, so, uh, across into it. We're integrating two teams right now. But now we are forward looking and we have 20 to 30 that want to onboard in the next year.

So we very quickly need I think. Yeah I think at least on I saw when we had two teams, we already saw the benefits of having a Kiwi because I said this joke many times, but if you have ten losses you're on the front row. And I would give them the same use case and give them an hour to build it, you'll get ten different solutions. And that's the problem with Pega. You can do anything with it.

And but you don't want to get everything out of it because you want everybody to build it in a consistent way. And I think that's where you straight away get your benefit, is that if all these teams start working in the same guardrails, so that's your selling point. Um, and you can talk to probably any customer, which is now in PegaWorld, and they're all acknowledged that you get really quickly.

You you end up having teams doing their own thing because there's a different LSA who has its own history, and with all the right reasons it's doing it in his way, might not be the right one from an enterprise point of view. Maybe in a silo if you just have one project, fine. but from an enterprise point of view, it's terrible. And if I can add a COE sounds really formal already. Just start simple.

Putting some agreements on paper in which you work together on and agree on, and then start from there and building that up and scale it when you are increasing the number of teams. Absolutely. I think that's a good point. I think maybe for the for the conversation, there's there's not a run book with a, with a Blueprint and this is what you're going to put down day one. It's something that organically needs to grow at least one. I had a totally different organization set up than in Bovemij.

It's a different organization, different politics, different position towards architecture. Really important often in your conversations. Um, do you centralize things or do you leave it in the teams? So you have to adapt it to your organizational model and never begin with big, small first routine, um, Introduced testing. Unit testing is if you ask 100 developers, probably 99. Don't test do it. I start introducing those things. Train them to do it because a lot of people don't know. Fine.

But it's an opportunity to get everybody on the same page. Excellent. Thank you so much. Yeah. And then related to the AI we already discussed before, I treat them as a kind of reusable component. Yeah, yeah, yeah. In certain steps you need that component, that AI agent, and then you get something back from that. Yeah. Um, and of course also write down the purpose of that and when to use it. So the other teams also know how to apply it. Yeah. It sounds like it sounded like from that question.

I mean, there were some things, some principles within the organization that are going to be very, you know, that can be leveraged here. You know, there's there's a series of agents, there's a responsible framework. You know, how might you know, It's a it's a library of reusable components, so to speak. How might we do something similar around Pega and where might we get that started? And you know, how do we agree on that.

And then kind of build it up from there so that these two things meet in the best possible way? Um, what's the superhero of the future look like, guys? Like, not in life, but like, in Pega in Pega land. Like. Because, I mean, things are changing so quickly.

And you talk about Lsas and the variety of different ways that lsas have kind of taken advantage of the world's greatest box of Lego over the years, and how you've kind of brought people into more kind of consistent, repeatable, scalable ways of working so that you accelerate time to value. It's easier to stay current, all that good stuff. And now comes this wave that is bigger than any of anything any of us have ever seen.

That seems to be changing at an exponential rate that many of us are having trouble keeping up with. Like, what's your what's your COE superhero of the future look like going forward? I think still, um, looking at the maturity of the community I see and I get in contact with and I see it in my own company too, is I think we still need to focus on getting everybody into App Studio Constellation module way of designing. That's still something we need to work hard on.

Still on the mission of what the what Vodafone was saying today. Like no sprint without a sprint. And then you want to import that into App Studio. You want to build it out of the box with Constellation. And on top of that you get, I think, uh, three years ago in Lisbon, we shifted from LSA driven to LDA driven. So the BA owns the case type. The LSA does not own it. And that was a I got so much pushback on that one because that is the role.

The back to your question, what's the rock star look like? I think the B and the S is maybe something from Pega five years back. I think we're more looking at like a configurator, knowing how App Studio works and whether you have an S background or a B background doesn't really matter. Just that that's the answer to the question who takes it further? So in technical level, um, and then if you add AI to that, basically you need somebody either that five legged sheep doesn't exist.

And if you find them he's probably really expensive or he doesn't have time for you. Um, and I know one sitting here somewhere, but that that's not going to help you. It's not sustainable. So we need to have different roles. And I think an AI specialist and probably different angles and an agent is a totally new perspective. Yeah. So you will need specializations in a team how I'm trying to do it. But I'm a really small company in ING.

It might not work, but I think I want to start like this POC engine, which is just going to take in use case by use case maximum two weeks. Spit it out. Works, doesn't work, shift it out. And once the POC is proven, bring it back to the business. Where where a Pega team is. Add knowledge to it from outside and get these people to learn and get HR involved that they need to search for this profile because we can't it's not scalable.

And I think that was my challenge in the beginning of Lisbon is COE. If you if you jump in too much in the projects, your scalability dies. Because then I have three consultants in COE and they go out to project A, B, C and then D starts and oh, how am I going to do that? So you need to get the teams enabled as soon as possible that they have the skills set in their own team. Otherwise you don't have that enablement power from your seaweed.

So it sounds I mean, there were three things in there that were that are very interesting. One is people who design the application and that that could be listening to the business, interpreting the needs, opening Blueprint, showing them something, dropping it into App Studio, taking it as far as they can in terms of the design of the case, and then the configures the builders, the people that work on kind of the gnarly things that exist in any of these organizations.

And then a third dimension around just AI savvy and learner's mindset around AI. Tim, I'd love to know what you think of that, but you also reacted when Lex talked about switching the s, the solution architect ownership to the business architect ownership around the case. That's what I'm also currently doing. I'm transforming that from a more business driven Within design instead of the technology. Who's leading it? So all my experts are certified business analysts, architects.

And and that works out fine. And you see that engineers are struggling with that moving from dev to App Studio. But that's the only way forward because then they have the same language, the same style. And you can be efficient, as efficient as possible. Yeah. And then related to the Rockstar I think in the end all simplified simple work will be gone. So there are only there are multiple rockstars in the COE. So it's the business architect, it's the LSA.

And then indeed the we have to come up with a new name, the lead AI specialist, and that that these three together are forming the band. Yeah, yeah, yeah, I think we were just before this session, we were talking about, Um, we have a manager who does the claims department. He has 40 people working in his team. And now that was maybe still today, but for sure ten, 20 years ago. So this manager is putting people to work. Simple work, medium work. Complex work. And then type of claims.

So he organizes the work with people capabilities. Now fast forward let's say optimistic one year from now. And now we have ten people 23 agents and an orchestrated process with all AI influences. What's this manager going to do? What is the the ten humans left over? What are they going to do? What's their skill type? Who's going to manage all these agents? Which AI agents.

And now if you put that back to Pega and AI and Co, I think you need to add a skill set in your CEO who also looks at the correlation. Okay. We have a Pega design Layer Cake enterprise. And now. Okay I the model repository. How do we keep that in sync. We all know that ChatGPT started with version one. Now it's into the four or whatever. So the version change doesn't mean it's better. So we still need to check again. COE on delivery. It's also the operations part of it. Yeah.

So it's multiple dimensions. Yeah yeah yeah 111 thing we haven't figured out is apprenticeship, right? Like what? Because if the model, if the traditional kind of pyramid model and the way we've built projects with discrete roles that do various things and they kind of apprentice on a path to mastery gets replaced by designers and builders and experts in I. And you're kind of left with your high performers supercharged by AI and Agentic. How do you become. What's your path to mastery?

To become a high performer? That's fascinating. Yeah, yeah. But to the question, I think the gentleman has more experience with ML models than I do. But it's also, I think, as an agent also starts as an apprentice and then it becomes a junior. And then people, the people, the humans start trusting it more. Then it becomes a senior and you start letting it go, and then it becomes an expert and you don't even look at it anymore. You just have to keep an eye out if there's going to be an update.

And I think maybe that's the corner where we need to start looking at. Also, humans can become play a role in you take care of all the juniors, you take care of all the mediums, and that's how you grow your skill set. As a Pega AI I agent developer, right? I don't know, it's just thinking out loud here. Yeah, yeah. And, well, and because you've got the workflow and, you know, we've got a, we've got a set of guiding principles and practices. We have reusable components.

We have a way we build we build workflows. We introduce agentic across it. Maybe there is kind of a maturity model of that agentic where we pilot it, you know, because we're going to be able to build these solutions so quickly and get them into market. We pilot it with, you know, a little bit and then, okay, we trust it more. And then, you know, we start seeing the traceability works and there's no hallucination.

Maybe maybe that becomes one of the functions of the CCO is that maturation of how we use Agentic in the organization and make sure that it's, um, well utilized across the workflow. Yeah. And the more the more companies are going to use cloud, the more data is available as well, the more mature the LLM models are going to be. So I think it's a growing path approach and we are at the beginning. It's just just started. The wave is building. Yeah. Yeah. So it's. Massive.

We, uh, if anybody's got any questions, feel free to come up to the microphones. We've we've only got about four minutes left here. Um, I'm going to put you guys on the spot, and, Lex, I'll do it to you first because you just took a sip of water. Thank you. You can do the Kerim thing and spew water all over me. Um, what's the one piece of advice you would give to. Well, different pieces of advice. You've got no, Co. What's the one piece of advice you'd give those folks?

You've currently got a Co and you're trying to figure it out next. What's the piece of advice you'd give to those startups. Just how to start. How to start if you don't have one, and what to do next if you've got one? I think you had it spot on. Start small. Just look at the different teams. Get all the Lbas and the lsas Together and start looking at how do you do it and then have a vision for yourself.

I think the biggest mistake I've seen is that we as a company, yes, as Bovemij, ING and all the clients from Pega, I really think you should have your own vision. How do we want to use Pega? What is our vision on it? If you ask all the integrators, they all have their own opinion. But what's your opinion? Because that that is key to guardrail where you're heading with the platform. And then which partner is going to help you should satisfy that need.

I think that's my biggest learning in the first years of Lisbon. And also here Bovemij it's really get that straight and then start small. I well said. And I, mayor, you and I were talking about a month ago and you said to me that your, your vision at Bovemij Is I'm I'm going to do everything quickly and show it to the business, but I'm going to do everything quickly, and then I'm going to pause and I'm going to think about how I could use AI in this. Yeah. Right.

And but then I'm then I'm going to build it and I'm going to show it to people and I'm going to keep going. And so I love that kind of, you know, pace but also that pause. Yeah. I think that's that's a really smart idea. Um, Tim, what about you? Yeah, of course I echo it. Start small. And then I also started off today. Don't do it alone. So we are here all together at PegaWorld.

So if you like to start the COE, come to us, have a chat, do a cup of coffee, and we can help you and and give some tips and tricks. That's what I also did. And that's maybe now a good time to shout out to, uh, to Jan-willem from Osprey. He helped me, uh, from BPM company also with my journey to get the co implemented in ING. So I took the help as well. But first indeed I had my own vision around it. Yeah yeah yeah I think it's really important. Clarity of vision which is a big idea.

Start small, build momentum, trust and credibility. And do the same with AI, I think. Yeah, have a vision, have a strategy and don't just go as a. Absolutely. Chickens without heads. Yeah yeah yeah absolutely. Well said. Well gentlemen thank you very much for joining me up on the stage today. I really appreciate. It. Pleasure. Thank you. Thank you. And thank all of you for attending this afternoon. And we'll stick around for a bit.

If you've got questions for Tim or Lex, I'm sure they'd be happy to field them. Enjoy the rest of PegaWorld. Thank you everyone. Thank you, thank you.