Hi, everyone, I'm Pinjalim Bora. I cover software at J.P. Morgan. Welcome. I'm delighted to have here our Alan Trefler, CEO of Pegasystems, and Ken Stillwell, COO and CFO of Pegasystems. Guys, welcome.
Thank you.
Maybe you can start with a brief introduction about yourself or a little bit about Pega, since we have Alan today, who founded Pega, what, in 1983?
Frightening to say. Yes. Yes, it feels. Well, it is a long time ago. That's true. But, you know, what's interesting about that is, I describe myself as founder and CEO because CEOs come and go, but founders, they, they really can't get rid of you that easily, I figure. But it's, it's been quite a journey over those years. We've gone through five major shifts in the underlying technology landscape. I mean, moving from DEC VAXes and CICS in those early days through, let's just say, some pretty cataclysmic shifts. And now I think we're on the verge of another one that we're actually deeply engrossed in, and we're pretty excited about it. So it's been, it's been a good journey, and, thank you for coming.
Yeah. So I've been with Pega for, actually, I think it's been, it's coming up on eight years. And I think what we've seen happen over the last probably around six, was this big shift away from the business model being largely around, kind of episodic purchases and the perpetual model where the clients manage the application. And that has really shifted to where now the majority of our growth in the business is coming from Pega Cloud. Our Pega Cloud business, you know, which was probably $20 million or so when we started this process, is now, you know, approaching $600 million of ACV. So, I mean, that's just been a, a tremendous...
You know, to see the leveraging of the technology and how it evolves and expands the use cases and really the clients, you know, adopting in a really massive way, our cloud offerings.
Yeah. So, Alan, since we have you here, and you talked about this, you went through a lot of transformation over the last four decades. But it's amazing to see that Pega still continues to grow, call it low double digits, and seems like your foundation is actually probably even more relevant today for GenAI, the foundation architecture, right? Maybe talk about that. What is different that made Pega survive for forty years and still growing low double digits?
Well, I think, you know, the reality is our growth rate has been one of those things that I think is an unfinished piece of business for us to address. I think Ken and the team have done a terrific job in putting the business on a solid financial footing, get us through the subscription transition. But I do think that what we've built is tremendously applicable to the challenges that organizations are gonna face in the next 3-5 years. And in particular, we always were geared around a model-driven architecture. That means how do you create a model of the types of work you do, the types of decisions you make, whether it was basically doing the original workflows that the company started with or the introduction of AI, we always had this concept of a metadata-driven model.
And when GenAI hit the scene big, and we really jumped into it pretty hard, close to 18 months ago, what we realized is that we could plug GenAI into our model so that the GenAI could inform the model and take the traditional work we did with statistical AI, which we've been doing for about 12 or 14 years now, and have a very large business in, and complement that by using generative AI to literally improve the processes themselves. I think we have a big advantage there because we're not, like, generating code. We're actually generating into a model, which then can be looked at by business people and IT people, can be understood, is incredibly reliable because it's the same concept that we've been using through all our generations.
And so I see us being able to plug generative AI in in very exciting new ways, and we're beginning to see an uptake of that with new capabilities we've offered in just the last 5-8 weeks. This thing called Blueprint, which we might want to talk about, which is available to well, frankly, anyone, and really represents, I think, a game-changing approach to using GenAI in the enterprise.
Yeah. So maybe, maybe, pulling that thread a bit, right? Architecturally speaking, what lends itself more so into GenAI now? You were, you were trying to talk about it, but it's... What is the difference, right, how Pega is approaching GenAI versus any other company in this space?
Well, let's talk a little bit about the two types of GenAI use. You know, I think one of the parts of the massive hype and BS that's accompanied AI is that we see three different classes of AI, and people kinda just, like, mush it all together and confuse things. The first is statistical, which is how do you look at patterns, how do you look at data? How do you determine statistically how to do machine learning and apply AI in that mechanism? Right, that's pretty tried and true and extremely valuable and complements the other forms of AI.
And that's something we do with our Next Best Action capability, that came out of a company we purchased, going back now over 12 years, called Chordiant, which we were able to bring in, and obviously that's all changed massively over the years as well.
But that statistical piece, I think, is a key element of what we do. Then there's generative AI for features, and when I say features, I mean, like, summarization or, you know, being able to write a personalized note or a letter and the other types of things that, are gonna add a lot of value. But I see our customers taking some of that from us, some of that, you know, they'll build themselves. And these features, I think, are really powerful, but not necessarily game changing. And then the third type or application of generative AI, which I don't see other people doing, is to use generative AI to actually train the system in what you want the processes to be.
And that's where we were advantaged because we always had this model that people would go, and they'd have to work with, and they'd have to work together with business and IT to kind of model out a loan process or model out how you do know your customer in a bank, or model out how you go search for new leads. Anything you wanted to model, you would model, and then our system, from the model, would write the code. Well, now we've got this incredible power of the Internet, which we've coupled with our historic best practices, to be able to let you blueprint what you want your model to be in record speed. And candidly, it's really, in many cases, much better than what people come up with.
So instead of people spending two weeks in a conference room at a whiteboard trying to sketch out what they want their process to be and capturing it on Post-it notes, they can sit and have this interaction with the Pega GenAI, available right on pega.com. You can actually go try this yourself. And what it comes up with is remarkable. When it's something that Pega already knows a lot about, we're able to add a lot of value there. But even when it's something we don't know anything about, we had one of our customers, we wanted to know what it would take to build a football league in Germany, which is not a source of, you know, any experience, Pega there.
What the system came up with, explaining how to do that, was, according to people who know something about football leagues, remarkably good, and on top of it, it was in German. So the whole game of how you get started, how you execute, and how you stimulate design thinking, and creativity, I think that's changed from its inception right through to execution in ways that we're just beginning now to unpack.
I think just something to add to what Alan said. If you think about the concept of a blueprint versus say ideating without any type of structure the blueprint has, you can imagine if you were designing a house, and you said, "Let's build a house." And you took people that had built homes, and you just used what their memory was in terms of how they build a home versus actually blueprinting out what that home might look like. The consistency of the quality, the repeatability, the predictability of what's built, the leveraging in a structured way what you've done in other homes, is tremendously valuable in terms of the end product.
And it's—it doesn't take away the creativity, but it does take away some of the variable art forms that happen when people build applications that end up, you know, maybe not being scalable, consistent. That's a challenge in enterprise application. So I think it's really a wonderful way of taking a body of knowledge, some level of structure, but allowing the design thinking aspect of it, but not having it be so free form that you end up with something that may not be optimum that you build, and quite frankly, takes a lot longer to go through the traditional process.
What's a good framework to think about ideating without blueprint versus with blueprint, right? What's— How many hours is it saving? Any way to think about that?
Well, you know, we used to have an approach where we would sit people in a room for a couple of days to sometimes a couple of weeks to ideate, and it would lead to good outcomes. They would actually use our tool to capture their thinking, and they would put it in. They would very often go through that whiteboard and post a process that I think a lot of us have probably seen. Now, it's completely different. We talk about design thinking with the customers for a few minutes, and then they articulate what they want their processes to do, that they want to go after lending to a particular part of the market in a certain way, with, you know, these sorts of competitors that we need to be sensitive to.
Just from that sort of textual description, right, or, you know, you can even have ways to feed in pictures of your existing processes if you happen to have them, and then challenging the blueprint technology to make it better, to make it so it will work in all the different channels, to make it so that the decisions can be automated everywhere they can be, and have that really built in to the underlying nature of blueprint. You literally go from what would have been a week or two to a morning, and it's enormously high quality.
I mean, it still needs to be reviewed and paid attention to, and it's not looking to obviate the role of experts, but it puts them in a completely different place, as opposed to having them generate and argue and candidly, you know, in your house design example, which I think is a terrific one, wouldn't it be awesome to have all the homes in the Internet?... to be able to draw ideas from and have them come in. I think that really brings together a lot of power. So we're super excited about this, and we think it plays to our design in a way that it won't play to other people's. So we feel good about that too.
So, just one add to that, which is, which is not only is it a saving to the process that you would traditionally go through, it also allows our clients to be able to go deeper into their potential project list, right? Because many times what happens is a client might have 50 different things that they might want to do. They have to first pick five, that they're actually even gonna go through the multiple week, multiple month, multiple quarter, however long that takes. Whereas this faster process allows them to throw a net over a bigger part of the potential opportunity and maybe end up making different decisions on what they actually pick and what they go after first. So that's another benefit of it.
So, the success that you're seeing on kind of the efficacy of that, you're basically training the LLMs on the existing best use cases that you have, the best practices that you already have.
Yeah. So I want to, I want to be careful about the word training, because I, I think there's a lot of confusion about when you do training and when you use some other techniques. So the approach that we've taken after having done some work, looking at what it took to train and retrain and models, et cetera, is, is we decided that's not the right way to do it for the types of things we're doing. So we're able to use a foundation model. You know, now we're using GPT-4. We've got... You know, we can interface with other models that are out there, et cetera. But what we do is we complement that with something called retrieval-augmented generation.
So we've taken all of our best practices, we've encoded them using LLM technology and make it so that we can retrieve the right bits of those at the right moment, feed it into the LLM, and ask the LLM to bring the best of the internet together with our relevant best practices when we have some, and then we, of course, have to weave that all together at the end, and that's part of what our Blueprint technology does. So I'm just a little sensitive to the word training because some people use it very casually, and the decision to train a model is a very, very critical decision. It makes sense in some circumstances, but in, I'd say, almost all business circumstances, it really doesn't make sense.
Yep.
-um, here.
Yeah, completely understand. That's a fair point. So you're not directly monetizing Blueprint, right? So what's... you're basically trying to get the consumption of Pega through Blueprint. Is that fair to say?
Yeah. I'll say you'll know that we've been successful with Blueprint if our growth rate goes up. So it won't be hard to see. Every single process project that we're now doing is starting with a Blueprint. And by the way, we also have customers who are using Blueprint to record processes that they might never put in Pega-
Mm.
to just use that as a way to understand and think and be challenged. But that doesn't break my heart either. If a large bank were to decide that they wanted, you know, Pega Blueprint to, to be the stimuli for how they thought about process optimization, I figure that will turn into good stuff down the road.
Yep, understood. So you have, let's say. Somebody's thinking about idea, you ideate using Blueprint. You reach to a point where you have to now create the application, and that feeds into Infinity, let's say. Could you use. Now, obviously, coding is a big part of GenAI, right? That's one of the most. I mean, it's basically proven that it's the highest efficacy GenAI use case, right?
You mean to use GenAI to write code?
To write code.
Yeah.
Right? Could you use that further to kind of do an end-to-end application from ideation to code?
Well, I mean, so of course, what we're looking to do when we bring the Blueprint into our Infinity engine. So we wrote the Blueprint application using Infinity, which is our core engine and product, but we wrote it as a SaaS app on the web so that it was super easy for customers and prospects to be able to get to. They can just go there, sign up, use it, no charge. When you wanna actually make that a real app that's talking to your SAP system, using your single sign-on, able to book accounting entries and send funds transfers, that's obviously not gonna run on this little SaaS app.
That's going to be something that's part of your Infinity infrastructure, which may be on the cloud or may be on-premise, depending on the generation and type of customer that it is.
We allow you to export the Blueprint into Infinity, and then it will create the artifacts that Blueprint has defined to radically improve the ability to build out and finish that app. We don't require you... In fact, we don't want you to write any code because we think the model, and we, and we've built, you know, a $1.5 billion business on the idea that these models, the model, not a language model model, but the model of the processes and decisions you make, is far better as a model as opposed to code, where it's really hard to understand, see what it's doing, manage change, et cetera.
So we think that by having a model in hand that already can do all the stuff that Infinity can do, that gives us the perfect thing to plug Blueprint into, as opposed to generating code.
You know, the hard thing here is how you figure out how to change stuff and understand it, and how you figure out how to pass an audit. You know, passing an audit when you have to tell the auditors they should go look at the code, that's a pretty uncomfortable situation to be in than a regulated business.
Yep, understood. So a lot of these GenAI capabilities are being launched in Pega Cloud. So maybe this is a question for Ken, but do you expect to see GenAI drive or accelerate the migration to cloud from your client cloud on-premise solutions?
So we actually, we allow and support our clients to leverage, the Pega GenAI capabilities, like Blueprint, but also our GenAI, kind of accelerators, whether on Pega Cloud or not. We actually have a technology called PremBridge, which basically is the ability for clients to connect their privately managed applications to our GenAI capabilities. That said, I think GenAI is just another motivator for clients to move to Pega Cloud. They've been... You know, when we first started this journey, we were, I would say, more agnostic in terms of saying, you can, you know, you can Pega Cloud, and then there's also Client Cloud.
And clients were, for a while, it was kind of 50/50 in terms of, you know, the growth of clients on Pega Cloud and the clients on client. If you look over the last year and a half, you see that number starting to get upward towards 75%, and I think that number will continue to grow. And in addition to that, you have clients that are selectively deciding to move and to migrate onto Pega Cloud. So I definitely think this will just be another motivation of why. And to be honest with you, you know, many verticals, many industries, had regulatory hurdles to move to cloud. Financial services is one that's been classic. I mean, they probably would've moved faster in many use cases if it wasn't for some of the regulatory hurdles.
I think those are now not entirely broken down, but I would say in a large way, and I think you're now starting to see countries and regions that traditionally had more concern around cloud, actually, you know, embracing it. So I definitely think we have this, a lot of momentum around cloud in general. Pega Cloud, tremendous momentum. Our sales teams, our partners, et cetera, are all very much behind it. GenAI is just another one of those motivators.
I have to say, I think our clients have a way better experience running this on the cloud, where they're getting outcomes as opposed to having to do the, you know, maintenance of the database and the application of patches and other types of things that are... You know, frankly, the reason people move to the cloud is that it's operationally so much more effective.
Yeah. Ken, maybe, maybe let's talk about macro for a bit. Seems like it's still a topic of discussion in among software companies, and we still see kind of a little bit sluggish, maybe, maybe not, worsening macro environment. But I would love to hear what you're seeing from your discussions with customers around the spending environment.
I'll start, and then Alan, chime in. I would say, look, during the pandemic created this artificial kind of spending layer, right, that happened, where people were reacting to things. And so growth rates in IT went up for some period, and then they've kind of settled back down a little bit. That's not a Pega comment. That's just in general in the market. What we've seen, you know, we're selling to the largest companies, the leaders of their industries, the brands that we all know, and what we've seen is a lot of resiliency in that population. Now, that's not to say that they're not under pressure like, you know, any other, you know, for-profit business.
But I think what you've seen is a real resiliency and a real commitment to digitally transform, and quite frankly, still very available budgets to be able to execute that, even given the fact that the industry spend rates have actually come down, you know, by a few hundred basis points over the last 2 or 3 years. So I think things have been holding up reasonably well, and I think our clients, we've been very engaged with our clients, you know, and Blueprint just gives us another really exciting engagement dimension. But I would say we're holding our own in terms of the software, you know, the technology space and spending. But once again, there are uncertainties. Interest rates are high. There is global conflict. We have an election coming up.
There's a lot of reasons to fear stability in the system, but the largest and best-known brands cannot wait because they're the ones everyone's targeting. So we feel like companies like, you know, J.P. Morgan, they're- you, you have no choice, right? You have to be a leader in technology, and we believe that's why we're so focused on our Target Org Model and focusing on those exact organizations, and why we, you know, why we continue to grow almost exclusively with those organizations. So, Alan, I don't know if you have any thoughts to add.
No, I think that's right on target. Yeah.
Okay, so let's talk about go-to-market for a second. In the last couple of years, you have made some changes, and I'm not sure if that's well understood in the investor community. So you know, you have moved into a targeted account model. You have optimized the sales force a little bit. Maybe talk about those changes, and any go-to-market initiatives that you have put in place going into this year.
Well, so, I think that we've made all the significant changes that we really feel we need to make to get the go-to-market into a better shape, and a lot of it now is execution and us getting more effective at following the Target Org Model. We went through a period in which we decided we were going to do a lot of marketing, generate awareness. You guys have probably heard all these words from CMOs, you know, generate awareness and build sort of demand pipelines and whatever. Let's just say that didn't work so great for us, and we found that it was much easier to spend money to budget than it was to hit sales to a budget.
So we decided that in our hearts, we really were an organization that wanted to do great things for our clients.
Part of that is having clients that we've mutually agreed with those customers that we're going to engage. We sometimes make that agreement first and then try to coax them into wanting to engage. But really picking targets, staying focused on target organizations, you know, still have the random things sort of come in over the transom, but really not trying to cultivate the classic demand funnel. I believe that has both led to better results. Certainly, it's led, from my point of view, to more satisfying outcomes in terms of working with these customers. I believe that's one of the major reasons you've seen some of the improvements in financial performance that have come along with that change.
And the second change we made to go to market, just to build on that, was the first iteration was really focused on the Target Org Model. And by the way, the Target Org Model does not mean no new logos. It means a targeted new logo approach, targeted companies that we know have the propensity to buy, that actually fit the scale that we have. The second change that we made was we realized once we kind of went through that first iteration, we realized that our client engagement teams were still dispersed across the organization, and so we actually centralized, really...
So we wanted to have, like, kind of the concept of, like, a, an account or an org captain that actually worked with all the parts of customer success, of the solution, of the partner organization, as opposed to having these orthogonal groups that might, you know, all well-intentioned, but were actually having different conversations with our clients in different parts of the selling process. That was the second one, and that ended up helping us consolidate some of our management layers. And so those were the two pieces we did. One first, which was the Target Org Model, the second, which was kind of refining that Target Org Model, just to help for people that aren't aware of that.
Yeah. Ken, I want to talk a little bit on the ACV growth that you have. Q1 was a little bit of a step down, and you kind of expected that it'll be, it'll be a slow start to the year, it'll be more of a back-end loaded year. Maybe it so when... You added something like $18 million in Q1. I think when I look at the guidance, it's almost $120 million that you have to add in Q2, Q3, Q4. Trying to understand, maybe help us understand the puts and takes, you know, what, what gets you there, what could be an upside? What could be a downside there?
So I just... You're absolutely right with the $18 million. Currency had a little bit of a headwind in Q1. If you adjusted for currency, it probably would have been more like $28 million or $29 million of ACV add. But we knew that the dimension-- If you look at the linearity of our ACV last year, it was a little bit of an anomaly, and when you think about ACV growth, you're comparing last year to this year. And so if you look across the Q4 , last year, we had a very big jump in ACV in Q1.
We actually had a slight decline in Q2. We had a flat Q3, and we had a big jump in Q4. If you look at the way that just rolled out and we look at this year, we didn't see that, we didn't see that trajectory.
We saw more of a level trajectory between the first Q3 . So if you just do the math only, what you would see is ACV would drop, and then it would come back up, and that's kind of what we had anticipated. And it's not a demand environment thing. It's just, you know, we work with large clients and typically deals that are not insignificant. And so sometimes you have lack of linearity in the booking flow. Actually, probably many times you have lack of linearity in the booking flow. So that's really what we saw happening. And so we saw that happening when we guided, and so we signaled that in our guidance, and it happened in Q1.
It looks a little suspect because you want to say, like, well, wow, you went from 11% to 9%, why is it not going to go to 8, and then 7, and then 6? But the reality is, in Q2 and Q3 last year, we've added very little, quite frankly, no ACV, and we know we will this year. So some of it's just the math.
Yeah, understood. And in terms of AI kind of contributing to ACV, what are we looking at? The time frame, is that more of next year?
We've said that our target for this year did not involve any incremental impact from AI. If that happens, you know, that's awesome. We believe it will have pipe, right? Yeah, yeah.
I think you mean AI SKUs.
Yes, sorry. Yes.
I believe-
AI SKUs, yeah
... already the success or lack of success at Pega is gonna be very, very tied to customers liking and adopting things like Blueprint and the other types of features we've included. But breaking out AI out as a set of distinct SKUs, we have a little of that, but it's not, it's not meaningful. AI is being driven right into the core of our business, which, I, I think is actually pretty exciting. When I talked about going through another kind of generational change here, where we've seen the technology change so much over the years, this is, a generational change to the, to the core of the business. Happily, we can do it, leveraging what we've already got, which is, which is a nice piece.
And actually, regardless of whether it impacts Q2, Q3, and Q4 in the numbers, it's critical to the go forward, right? So we need adoption of Blueprint. We need our sales teams and our partner teams and clients who use Blueprint because it's gonna build pipe, it's gonna advance, it's gonna give us an opportunity to target some very specific new logos as an opportunity. These are all reasons why it's critical for us to have momentum. Now, whether or not it generates bookings in Q2 or Q3 is not really the bigger... The bigger question is, does it actually help us accelerate our growth rate going forward?
Yep.
Yeah.
Let me see if anybody in the audience has any questions. No questions? Okay. So, so on that point, obviously, partners is a big channel for you, and you have some very large GSIs built, really big practices around Pega. How do you think of that in a GenAI world, right? Blueprint seems like might encroach into some of the consulting knowledge that, that they do. So talk, talk about that partnership, or the partnership go-to-market. How does that change in a GenAI model?
So we wanted our customers to be excited about what we were doing with GenAI and Blueprint. So a full month before we released it publicly, we actually went to all of the major partners, and we had a very deep dive, formal set of discussions with them, and we made it available to them. And, we've also said that we're going to allow our partners to be able to submit their own blueprint. So for example, they can create a blueprint for something if they believe they have the world's best approach to licensing in government or to collections.
They can create their own blueprint, which will be owned by, you know, an Accenture or Cognizant or whatever the firm is that does it, and that they'll be able to use with customers, but will not be available to everybody else, and candidly, will not be available to Pega. We're not gonna be able to see their blueprint so that they can feel that they're really building their own IP. And I'm thrilled to say that, I just saw the end of last week, our first partners submit multiple Blueprints. You know, we have to upload them into the SaaS environment so that they can be available for their use with their customers.
So I think we're gonna get a lot of partners actually using this as a way to show off what they know, which is exactly, I think, the way we make this positive. I mean, the reality is, AI will simplify a lot of the traditional, particularly lower-value tasks around building and implementing a system, but there's still a need for that thought leadership and design thinking and creativity. I just think they'll be able to do a lot more for customers for the same amount of money.
And just something to add to that, if you talk to some of the large SIs and you ask them, one of the biggest costs that they'd love to deflect, it's the upfront business development cost around showing, like a kind of an early version of a product. I mean, that's. Blueprint is really solving that problem for them because they're not able to typically charge for that, and that ends up being business development costs that they want, you know, companies like Pega or other companies like us to help support. But to be honest with you, if they could deflect that, you know, that's just a, that's an amazing accelerator to them being able to show clients a vision early.
Yeah. Ken, free cash flow in the model has been an outstanding few quarters now with huge free cash flow growth. Maybe talk about that, that trajectory. How much... You know, where do we see leverage from here on out? Seems like S&M is probably still a category that you might get some leverage from. But talk about more from here on out, where do we see leverage in the model? Where does that free cash flow margin go?
Sure. So we knew when we exited the cloud transition, what we wanted the business model to look like. And I think as we exited it, the business model looks pretty close to what we had hoped it would look like. If you account for some one-time items and things that are not, you know, kind of systemic in the business or more just episodic. I think when you think about where is the next frontier of opportunity, gross margin in Pega Cloud is certainly one, and that is operational improvements we can make, things like leveraging Kubernetes in the cloud and some operating scale. When you look at the leverage, we wanna get from things like selling R&D, G&A; we don't wanna get those leverage points by actual cost reductions.
We wanna get those by accelerating our growth rate with an efficient cost structure to do that. So I think that there's an opportunity to accelerate our growth rate with really, really nice accretion of margins with that growth rate. We could certainly, just like any other software company, I mean, you can cut your way to whatever cash flow number that you want. You're not. It's not gonna help you in terms of accelerating your growth rate. And we believe Target Org Model, large organizations, we need to be able to be in front of our clients.
We need to be able to support our clients. And so that requires a level of attention and investment that might be different than if you're selling a, you know, a cost solution that someone buys. So we have to be really thoughtful.
We will be very thoughtful about the target org model. We'll be very thoughtful about new orgs, but the reality is, we're gonna get our operating leverage by being able to expand our growth rate and not have to expand our cost at the same level. With the exception of one area, which is gross margin in Pega Cloud, which is in the high 70s% now, which will be 80%, and has the potential to go above 80%. Quite frankly, at our scale, if you think about a $600 million SaaS business with gross margins in the high 70s%, that's pretty respectable in terms of other...
You know, if we can get above 80% gross margins before $1 billion, which we will, right, which we are certainly on track to do that, I mean, I think we're, you know, we should, you know, be proud of what we've accomplished to get to that point.
Yeah, I think I heard expansion of growth or acceleration a few times, so I like that. With that, in that note, let's end it. Thank you so much for the time.
Thank you.
Thank you.