Thank you everyone. Peter, why don't you to kick off the agenda?
Sure. Thank you, Ken. Good afternoon, everyone. It's great to see the investment banks that cover us, our investors, and even members of our board here in the audience today. Rick Jones, great to see you here. Thanks for coming. We're very excited about today's session, being back at PegaWorld in person, we have what I think is a fantastic agenda for today. We actually started planning for this back in January. Talked to many of you about the topics that you thought would be really important for us to cover today, that's what we're going to focus on. Before I get into the agenda, just gonna cover a couple quick housekeeping items. First one is our safe harbor statement. Certain statements contained in this presentation may be construed as forward-looking statements as defined in the Private Securities Litigation Reform Act of 1995.
I would encourage you to take a moment to read the safe harbor statement. If you're listening to the audio today as well, you can actually go on to our investor relations website, take a look at our most recent filings to get up to speed on the safe harbor and the forward-looking statements. As Ken mentioned, we are gonna be taking questions live today. We're doing an audio broadcast of the session for today. Please wait for the microphone. If you're in the room today, speak into the microphone. Don't be shy. Please speak right into that nice and boldly, the AV guy asked me to say to you guys today. If you're online and you're listening to the audio today, we'll also take questions.
You can email your questions to Pega Investor Relations at pega.com, or you can email them to me at peter.welburn@pega.com. I'll be taking a look at the questions that come in. We can pass those along. In terms of the agenda for today, my expectation is we're gonna run about two hours, maybe two and a half, if we get a lot of questions, but we would like to have dialogue with you all. We're trying to shoot for sort of two to two and a half hours. Alan's gonna come up first. He's gonna talk a little bit about AI. Alan will take questions right up front, so if you've got questions for Alan, please be ready with those. After Alan wraps up, Jacqueline van Wees is gonna be speaking about sales.
Last year, we had Leon Trefler speak and John Higgins speak about sales from an executive perspective. We thought it would be great this year to give you guys a view from the field, so someone who's working in sales every single day, what's her experience? She's been here quite a long time, and won't steal our thunder, but very excited for Jacqueline to be here. After Jacqueline wraps up, John Higgins is gonna be speaking. I've known John quite some time. We acquired John's company. I've known John since the first cold call I made out to John, saying, "Hey, John, you might be interested in acquiring your company." Very excited that John has been with us since founding In The Chat and since Pega acquired his firm. He's gonna talk about Launchpad.
When I talked to many of you about planning for today, you said you were interested in Launchpad, wanted to learn more. We've got the chief of Pega Launchpad here, and I'm also very excited about my friend Steve Bixby, who's gonna be coming up here, VP of Product Engineering, along with Cara Manton as well. They're gonna be talking about generative AI and the power of the Pega Platform. What I love about their presentation is they're going to show it to you. When preparing for this session today, a lot of the feedback was: "Don't tell us about it, show us." They've prepared a demo. Looking forward to seeing that. Then Ken will wrap up with a financial discussion and also take your questions. At this point, again, very, very excited about today. Thank you all for coming.
I know many of you in this room are thinking to yourselves, "Was this worth the trip?" I'm hoping at the end of the session today, you say, "Wow, Peter, that really was worth coming out to Las Vegas again to hear about Pega, the innovation in the market." That's what I'm hoping happens today. Alan, why don't you come on up? Share a little bit about AI and answer a few questions.
How's that? Excellent. I would say as impressive as, you know, Peter and any presentations we have should be, that one of the real opportunities for people who are visiting here is to actually talk to clients. I think the client stories, I was, of course, thrilled to get two, you know, marquee names like Citibank and Aflac to stand up and say the things that they did to all of you and to their competitors, for that matter. Ultimately, I think that the success of our clients defines both our value system as a company and defines our value in the market to both them and to other organizations that we want to work with. Mingle, get to know, just ask customers what they're doing and what they think.
I will tell you from my experience, they can respond with enormous candor, especially the Dutch, but that's a whole different question. A whole different question. You heard my opening today about AI, and you got a chance to see, you know, Don, the Omelet King, and Karim and Cara really show off some of what we've been working on. I will tell you that we've been deeply engaged in this. The thing I tried to explain a little in the presentation is that this Gen AI stuff has really set a lot of imaginations on fire, but it is really a facet of the AI continuum, and it's one that we have lived in for a long time and find, I think, very natural to understand the power, but also the limitations of it, and I think are doing.
From my point of view, I think the team is doing a really, really exceptional job at understanding how to, at pace, harness the potential, make it be a well-organized, sustainable way to go to market and be in a position to do great things for our clients. I think that will come across to our customers over these next few days. We've made a number of announcements about it. I will tell you the level of hype and bullshit that's in the world around this is. I was gonna say it's unprecedented, but we see this sort of stuff all the time, you know, where there's some, you know, new miracle that's gonna change everything.
This is more superficially impressive than a lot of those miracles, but like many of these cases, it takes a lot to be able to operationalize business change and operationalize business improvement. That's, I think, where we come in terms of being able to provide the elements to deliver it. You know, the reality is we see companies that we're engaged with being full of workflows and AI-driven decisioning. We have an architecture that when I look at what others are doing out in the market, what other people are even saying, I believe we operate in a completely different level and a completely different scale of sophistication, which candidly might not be necessary for every client or every problem, but there are huge swaths of customers where this is exactly, we believe, what they need.
I was talking to a prospect client actually just yesterday, and I said, "Your job must be very hard because all software companies tell you they can do everything." They laughed, and they said, "Absolutely." Our mission is not to say that we can do everything. I think there are some things where we can do them extraordinarily well, and I think the folks who are, and have listened to the people who say they can do everything, you know, I think that they find that everything might be with a really, you know, lowercase E and, you know, may not even be fully spelled out at all. Really thrilled with what the layer cake does for us in terms of creating the fulcrum, the architecture for being able to drive AI.
Our center-out architecture, which is what really lets you have collections of engines that know how to bring brains and process and case management together. Those which anybody who's been, you know, to PegaWorld before will have heard me speak about. These are all fundamental elements that are accelerated by technologies like Gen AI. The change that I see Gen AI providing to the way people build and use these systems is gonna be pretty startling. You saw the demos of how Gen AI can really be an autopilot for the work you want to get done. It can guide, it can coach, it can avoid errors, and, you know, what you were seeing is a real set of capabilities coming to our customers this summer.
I think, you know, Steve and Kara will give you a little bit of a sense of where that is going, which candidly is up at another level of, I think, augmentation and helping our customers be successful. I think this is a enormously exciting time. I think we're very well positioned. It's wonderful having enterprise clients who, despite the travel bans and the patina of Las Vegas, let's just say, it's not the easiest place to get people to come to if, you know, they are in a tough economic setting. Many thousands are here.
I think if you look at those, I think if you look at our partners and talk to our partners and the enthusiasm they have for what we can do together, hopefully, you will, well, concur that a high level of excitement is necessary. We certainly think it is, and we think this is breakthrough stuff that's gonna make an enormous difference. I'm glad to answer some questions, though there are others who could answer as well. Yes?
Rishi, if you could just state your name and the name of your firm.
Of course. Rishi Jaluria , RBC. Alan, thanks so much for the time. Love the session this morning. Two questions on Gen AI for you.
First, you kind of hinted this at the beginning of your remarks right now, but a lot of companies out there, including a lot of your vocal competitors, are talking about their Gen AI strategy, whether it's real or not. Maybe from your perspective, because there is a lot of maybe marketing fluff out there, how should we, on the investment side, really recognize what is truly AI and what isn't? How do you think about differentiating that? Then maybe if you could just, you know, at Karim's keynote this morning, he was talking a lot about the local models and how that provides kind of another layer of protection for those that want to train data on their own models, but not train the central model.
Can you walk us through a little bit how that works and, you know, what kind of the feedback to, or early reception or early feedback has been? Thanks.
Sure. Everybody's trying to figure a lot of this stuff out. Just to maybe bring some clarity for you guys, because as I said, we've lived with AI a very long time, and as it has evolved, paid close attention to it, and I think understand this all very, very well. There's really three levels in here, though most people only talk about two. It's funny because the level they miss is actually, in many ways, the most important.
One level is a public model, where you basically have a model that's sitting out there that, you know, OpenAI has done, and they actually know an amazing amount of stuff. I mean, we can go ask it for the steps in the loan procedure and look what it comes back with. It's just, like, shocking. I mean, because that's in the public domain. You know, your kid could ask it to write a resume. It will write a resume better than, you know, 95% of the resumes that are out there because it's learning from the probably millions of resumes it's looked at in the web and LinkedIn posts and all these other things. There's this generative AI that is based on, you know, large language models, though there are various sizes of those and that are available.
There's something called a fine-tuned model, in which you take an LLM, and you shoot your own data, your own use cases at it, your own documents, and from that, the LLM learns. It learns how to change the weights in the model. It's very CPU consumptive, and it's a very powerful way to augment a model, but it's actually not the only way to do it. We think you got to do all three of what we're talking about. We think you've got to be able to go out to the public models for things that make sense, right? Go out to a private model that will have been trained with your data. Like, when we're putting our contract data up in our work here, we're not going to put our contract data up on the web.
It's got to be in a private area, but it gets trained. The other way of doing it, which is kind of fascinating, is something using what's called embeddings in a vector database. This, by the way, is one reason that vector database stocks are going to be hot, but there are 100 of them, so you got to be little careful here. What that is, you basically take your data. Instead of training and tuning the model, you basically dissect and digest your data. You turn it all into numbers, so you can have things that are kind of close to each other. You know, all the dogs live together, and they're close to the cats. They're far away from the tractors on this, what's called a vector, a numerical scale.
With one of those databases, of which we have dozens for different purposes, you can actually take a question, and you can go to that database, and you can say, "Find me my five documents that most closely relate to this question." It will pull it in, then you can shoot that as part of the prompt. It's not part of the trained model, it's part of the prompt in with your question to a more public model. Your data doesn't really go anywhere, but you can shoot that to any one of three or five models then. You're not having taken a bet on a single one, and you can see what you like for the quality of answers.
That intermediate, sometimes refers to as an embeddings model there, is actually what a lot of people are doing, but we think you got to do all three. You got to use the public, you got to be able to train private, and you can use the embeddings, and different problems will be supported by these three different styles, which we support natively and will support natively in the Pega system as well. Richard, that's probably more detail than you wanted, but now you've got a tip on looking at vector databases here. It's, it's a really interesting space, and the marketing is mostly nonsense. You see companies just repackaging things they already had and reannouncing it, you know, which, of course, that's, that's what happens. You see brilliant demos like, you know, as beautiful as Karim's little pussycat that he showed.
Actually making this stuff work is hard, and one of the things you have to do to make this work is you have to be able to tie the models into an operational framework that actually does something and records history and can work across, if you're in big organizations, a big organization. That's exactly what we do with our workflow engines and with our layer cake. That's why I think we're differentiated. I would go and ask other people what makes them really differentiated? If it's quantity of models or if it's... Yeah, none of that is sustainable and doesn't actually, I think, you know, solve the core problem. Was that helpful?
Good. Thanks. Yep.
Thanks, Sam. Alex from Monster. A lot of companies are talking about benefits of AI to them, specifically noting that a lot of the competitors in the marketplace, a lot of the other software providers are kind of screwed up. I think you alluded to that this morning, talking about a lot of the very low-code automation companies that will be in trouble. Just building a little bit on what you just talked about, why is Pega a true AI beneficiary? Why do you think that Pega is really going to benefit from GenAI? Why will GenAI not disrupt a lot of what Pega does and isn't going to be a Pega solution at the time? Thanks.
Sure. I think the majority, if not all, of what I would describe as the medium to low end of low-code, is just shot. Because writing code has gotten way easier, and building tiny systems that can stand alone has gotten way easier. That was never our market. Our market was always: How do you create the sort of enterprise we use, enterprise Layer Cake, enterprise workflows? We have this structure, you know, and we have patents on the Layer Cake, and that, to our point, is that even though we all have been using the words low-code, you know, candidly, by my standard, many of these guys were not low-code. They were just programmed. They certainly didn't have the sort of structures.
You know, we're advantaged because we grew up with the enterprise and with companies that did lots of M&A and that had to try to integrate lots of different operations. That's how we sort of came to the problem, and that's led to an architecture that is empirically, I don't think you have to look very deep, to see that the architecture is really different, that our architecture is layered. It's a metadata idea. You know, we used to call ourselves no code. When low code became popular, we said, "Low code, no code," you know? What we really are is this layer cake. That's really the heart of what we do. We don't generate code like they will. Anybody who's generating code, I think that candidly, their success rates and their ability to compete is gonna be pathetic.
We generate the model, and then we can go into a piece of model and regenerate again. Because we have that intermediate control structure, we have a completely different way of looking at the problem than folks who are going from prompts to code, right? Because if you go from prompts to code and you build something that has 100 modules, it's gonna be super hard to change that. I don't even know how you figure out what it does. There's no place to understand what it actually does. In our case, the model is the layer cake, and it can contain information from a variety of models and a variety of pieces, but it organizes them. That's why I think it's gonna be great for us, because what you're seeing in our demos is we're updating the layer cake.
We're not just generating a little burst of code. When you go see what Microsoft does or what others do, they get people co-piloting the writing of code. Well, that's gonna be as candidly, as effective or ineffective as just writing more code faster. Candidly, writing shoddy code faster isn't so great. Yeah, this will help some, and there's a place for that. For the sort of enterprise workflow systems that wanna bring analytical AI, generative AI, and workflows together in a controlled fashion and manage them, I don't see anybody else who's candidly, remotely close or has the gumption and the architecture and has invested like we have to really understand this problem through several generations of technical development.
Yeah, I think, I think it's gonna be highly advantaged, and if that's not visible to you, having spent, you know, today and the rest of the day with us, I'd like to personally hear that, and we'll keep working it until it becomes visible, because I do think it's true. Yes?
Great. Thank you. Steve Enders with Citi. I wanna ask about kind of the GenAI into the marketing use case and maybe a bit biased listening to the Citi keynote this morning around CDH, but how do you view the generative AI opportunity within that part of the market specifically, and, you know, maybe shifting into either more content-focused solutions there, just how you think about the marketing opportunity now? Thanks.
I think GenAI and the way we're able to show GenAI even now, and we'll be continuing to push on this, is just going to be fabulous at being able to take something that's created by a human, like a human creates an offer for a particular type of credit card with a particular two or three characteristics. With GenAI, you can literally, in five seconds, have 10 versions of that offer. One, redrafted for a millennial, another for somebody who's really focused on savings, another for somebody who wants to think about the environment. It can take the attributes of, say, the credit card, the attributes of the offer, and spin them together differently to be a creativity multiplier and a force multiplier.
Using the Pega technology, you can test all of these and decide which are the ones that are working, which are the ones that aren't, with our Next Best Action technology. That's an example of a lot of marketing copy that was arduously written by hand now becomes generated. Our style is to have these, say, 10 bits, looked at and say, "Yep, yep, yep," by a person who might edit a word or whatever, but they're awesome. They're really super, super impressive.
GenAI is also gonna be a key part of how you in marketing might communicate outbound to your clients, where you can take kind of a template that says, I wanna talk to this customer about these five things, take the information about the customer, use the language model to craft a note that would make brilliant sense for an individual customer, or of course, you could do the same for a segment. I think in this sort of what I would describe as copywriting area, and communication writing, It's not gonna take a year. You're gonna see massive change in this year, and I'm really excited about those types of things.
This would be kind of like managing on front of it and kind of go to the next level?
Yes, absolutely. Absolutely. Well, we do that. We send outbound communications, right? We choose between what are called treatments about how to present an ad. Those are things we do. These all now become superpowered and way more efficient, I would say. Back? We'll get you, Doug.
Digging a little deeper into your generative AI commentary, can you talk about if there's any possible new use cases that you guys have found since introducing this technology within Pega? By integrating it, have you been attracting a new kind of customer so far that really hasn't been doing or haven't been approaching you guys for business?
There are lots of new use. The question is about new use cases and new customer types. There are lots of new use cases, and we're finding them all the time. Some of them you saw, out of what Karim and Kara and Don showed this morning. I think you'll see some more with Steve. lots of new use cases and new use cases constantly coming up. There are ways to automate testing. The whole ability to generate test data, which is an enormous pain in the ass if you're building a system, is like, gone. That's why there's gonna be huge productivity gains in a developer point of view.
We're not really, at this point, focusing on seeing this as a way to open up new clients, because our strategy and our Infinity business is to really focus on the, well, the customers who you see here, and maybe a small number of customers like them. We're not looking there to open up new markets. That LaunchPad technology that I mentioned, you'll hear a little more about today, that's our strategy for how we bring ultimately the entire Pega technology stack to market with a different motion, a different cost structure and a different set of expectations. I'll leave it to John to go through how we think that's gonna work over the next couple of years.
Alan, we probably have time for one or two more questions.
Here we go. Go ahead.
Pinjalim Bora, JPMorgan. Hey, Alan. Great presentation. The demo on the home application loan was pretty powerful. I'm trying to kind of think between two things: How much of it is actually generative AI versus you able to kind of templatize a home app loan, right? It seems like the AI part of it is the natural language processing that you're doing at the top, that's probably doing some kind of a semantic search, trying to understand what the user is trying to do. The next step seems like you could have templatized it and not necessarily need an AI. Maybe the sample test data would be a part where generative AI is being used. Help us understand what part of it could be just templatized versus AI generated.
We have spent 40 years learning how to templatize business workflows and decisions. You're absolutely right. The thing that loan application went into, with what we call stages and steps and personas, all of and channels and the interfaces, all of that goes into, and I'll use your word, a template, which is the Pega. That's what the Pega Layer Cake is. It is that template. It's because we have that template, that we can actually do this and control it and evolve pieces of it and plug GenAI into it. If we didn't have that template, we'd be just generating reams of code, and there'd be no structure to it. That's what Pega brings. We bring this very comprehensive, you know, carefully designed and highly scalable structure.
Candidly, without that, you'll see lots of demos that I think at the end are just gonna give you know, spaghetti in terms of what you have to try to maintain if you actually the people running it. Yeah, if you want to think of the Situational Layer Cake as really being built around the idea of templating in a versioned way, templating in a way that supports a concept called inheritance, templating in a way that knows how to be omnichannel automatically and have that built in, I think that's fair. The rules in our system are, in effect, templates for how you define pieces of your business. Our Gen AI feeds those.
Yes, the natural language captures information, gets a bit processed from the web, and we've built the bindings to go from what a natural language returns to how do we snap that in to the same template that people used to type into, right? You know, I think if you see that and understand that, I think you'll understand the heart of why this is hugely different than what you'll see with somebody who doesn't have a, you know, a large template layer cake. You know, that's our, that's our secret sauce.
Maybe one clarification.
Turn them on. Go ahead.
Thank you. I point of clarification, the template... This is Steve Bixby. I'm responsible for the Pega.
Sound happier about it, Steve.
Hi, everyone. It's nice to see you. I just wanted to make a point. It's a great question. I was actually worried about the way it was presented at one point. I think we might have misinterpreted what we're showing here. I think we're asking, the templates don't have any of the business context that we're running this on a, you know, a banking industry application. Pega is just a raw platform, no knowledge of industry. Yes, it's using the model to parse the statements and things like that. For the large language-... in a plausible business process in the format we ask it for, it populates that template, right, for this case and such. We'll show it in the demo, I just wanted to make sure I caught that part of your question. Awesome question.
Thank you.
Thanks, Alan. I just want to clarify one question that generative AI is, like you said, Alan, not we're not intending on going into new markets with new customers and new logos as a result of a new use case, that generative. That said, the developer productivity improvements, the time-to-value speed in terms of getting implementations or getting clients live, being noticeably faster, leveraging Gen AI, will give us the opportunity to potentially win some new logos in the target logos in our...
Sure.
That we may not have otherwise won because of that speed, and certainly to grow with our existing clients with use cases that we might not have otherwise won. I do think that just didn't want that point to be lost, that there. It's not the intent is not to open up new markets, but there will be some clients that may be compelled to come to Pega that we may not have otherwise won because of that differentiation. Thanks.
Yep.
That's great. With this, I think it's my moment to introduce Jacqueline van Wees, here from the Netherlands. Welcome, Jacqueline.
Thank you. Thank you, Alan. Can you hear me okay? I think the mic is on. Yeah? Can I try again? Is it better? Is it good enough? Good afternoon, everyone. My name is Jacqueline van Wees. I'm sales director at Pegasystems. I've been at Pega for 13 years. I was part of Chordiant, the company that was mentioned this morning during the Citi presentation. Before that, I was part of KIQ. That was acquired by Chordiant. So I come from the heritage of the decisioning engine that we've been hearing about a lot.
As Peter mentioned, at the beginning, I've been asked to share a little bit about you, from the field, how we are going to market, and maybe also how that differentiates us from, some of the other, technology providers out there. It's a great time to be at Pegasystems. Especially when times are challenging and conditions are tough, people look for reliability and for results. We believe that we offer both of that with, on one hand, our market-leading, platform and technology, and we'll be talking a little bit about the business problems that we are solving. We do that with renewed rigor around our go-to-market approach, where we focus mostly on our existing clients. It's not that that is entirely new, right? Alan spoke about that as well.
Pega has always been very focused on working with marquee clients, building long-term strategic relationships. Over the past few years, we have experimented with going a little bit more down market, actually, investing quite heavily in that because we were hoping that that would really accelerate growth. Although we saw growth, it wasn't to the extent that really justified the level of investment going forward. Most of the growth that we were seeing was still coming from our existing clients. We've decided to focus more on the profitability of this go-to-market approach. As a salesperson, and let's be honest, we may not be the most beloved species on the planet, it's really wonderful to work for a company that has always relentlessly focused on product excellence and client success.
Clients really trust us as a critical part of their digital platform to help them work smarter, provide unified experiences, and adapt to change. There's great determination throughout all of Pega to make sure that these clients are successful. Certainly, in my time at Pega, Alan himself, but also other members of the executive team, have visited the Netherlands many times to speak to our clients. Alan mentioned that they're not particularly shy, right? Dutch clients will tell us what we can do better. As a result of that, we really bring in our own expertise and our own skills, combined with what clients are telling us, where they want to go.
As a result of that, Pega has been recognized by leading analysts like Forrester and Gartner for years in a row as a leader in this space. We really try to make impact for the world's largest and most demanding organizations, particularly across these five areas. We help them personalize engagement. We heard a little bit about that from Citi already this morning. We really use real-time decisioning and AI to make sure that every client interaction is relevant, it's personalized, and it's driving client lifetime value. We help to accelerate acquisition and onboarding to improve customer experience, partner experience, and achieve higher penetration rates.
We talk about automating customer service for higher Net Promoter Score and reducing costs, streamlining operations, which means really automating workflows for mission-critical business processes and resolving exceptions where things go wrong or we need further investigation. As you can see on this slide, the companies that we work with are not for the faint-hearted, right? These are very visible brands that demand enterprise scale and stability. When you get it right with brands like this, then the results can be huge. If we think about Vodafone, and I happened to be there for the early stages of that engagement with Vodafone, they are seeing GBP 100 million profit every year because they are able to increase their accept rate three times.
Unilever, who was able to overcome supply chain disruption by onboarding suppliers in hours, not days. The one that I'm always amazed with as a European, is the U.S. Census, which is this massive project, counting every U.S citizen, including 400,000 field staff, with zero disruption and two times the productivity compared to the previous census. It goes without saying that serving clients like this requires a very deliberate go-to-market strategy, and ours revolves around three main pillars. We focus on this target account model. We spoke about that. It's really focusing on our existing clients that we expect that will be driving 90% of our ACV growth. These are clients that have been with us for many years, and sometimes even decades.
We heard from one of them this morning, but also clients that I personally work with, like ING and Rabobank, that I've been working with for 13 years, but that have been Pega customers for much longer than that. It's our job to know these customers inside and out, so we have cross-functional field teams that really focus on understanding these clients' challenges and goals to make sure that we make them successful and that they will want to do more with us. That requires a deep understanding, deep discovery, deep relationship building of these teams. At the same time, we are able to scale, flex up and down as client demand requires. Finally, this requires an integrated management across all of the client life cycle, from planning, sales, delivery, adoption, and renewal.
We do that with integrated management and metrics so that we can make sure we focus on the right activities and particularly the outcomes. We believe this go-to-market strategy will drive profitable growth for us, right? It allows us to focus on an addressable market that is huge because our existing clients represent a multi-billion dollar spend. It makes perfect sense for us to focus, as we have been doing for decades, to focus on building these long-term strategic relationships with these clients, making them successful, helping them to see results that they will be keen to replicate in other parts of the organization. As a result of which, we will be seeing, we will be expecting growth from migrating these existing clients up the ACV pyramid, as we call it.
That means clients that are currently spending up to $5 million with us a year, making sure they see those results, they invest in us for additional domains or regions or products, as a result of which they will move up into the tier, up to $2 million, sorry, $10 million. Similar, those are currently in $10 million, and you get the point, right? At the same time, we will still be adding also new logos. For example, if you think about the Netherlands, at the moment, we look after and focus on actually almost a handful of clients, but they're big ones that we think there is still a lot of potential for us to move them up the ACV pyramid.
There are other large organizations that we're not calling on today, but that may become part of this focus in the next couple of years. This, of course, requires great people and great teams that really care for the clients, that want to invest in these clients, but it's really worth it. We've seen it work, and we see, for example, that sales cycles with existing clients are shorter. These clients are more sticky with us, right? The retention rates are higher, on one hand, because of the relationship that we have with them, but also because of the breadth of the landscape using Pega across the organization.
The win rates are higher. Also we see, and I've personally experienced that as well, we work with people. If you can make people successful in their careers, maybe boost their careers even, and they move from one company to the other company, they're much more likely to turn to Pega again in their new organization. Again, to focus on that main goal of making our clients so successful that they will want to do more with us, moving them up the ACV pyramid, we really need this focus across the entire client life cycle. That starts with building extensive three-year plans. We have our teams really study our clients, looking at their business goals, their priorities. It can also be regulatory pressures that are relevant or an IT landscape that needs renewal.
Making sure we understand all of that, but also understand the client's buying cycle rather than focusing only on our sales cycle. That is really what we do during those three-year plans. We have a great client base that is really willing to share the often spectacular results and, of course, some lessons learned typically as well, and we'll be hearing from many of those this PegaWorld. Like I said, personally, I've been working with organizations like Rabobank and ING, who accidentally, both of them, started in one part of their organization, looking to improve their payments investigations process. We worked with them, made them successful, and then they started looking at Pega for other domains or other regions or even other products.
Over the course of the years, we've really been able to extend that partnership significantly to the point that they also come to our headquarters very regularly, say, one a year, to listen to our roadmaps, to share with us what they think we can do better. They do even summits. For example, we have an ING summit. Rabobank even has a Proud of Pega Day that they do every year. It's really wonderful to create relationships like that. It does require real partnership, right? It does require us to care enough to invest in each other, but also maybe challenge each other every now and then, and have the tough conversations when the inevitable bumps in the road occur. Clearly, partnership and trust like that does not get built overnight.
It requires flawless delivery and showing real business results. We focus on that regardless of whether a project is Pega-led, is partner-led, whether it's client-led. As an example of that, we have introduced advisory services that allows us to embed very skilled and experienced resources, Pega resources, into a program that will help a client guide their implementation, their designs, bring in the right business practice, best practices, and governance. I'm really excited about that because we're already seeing a big change as a result of that. Ultimately, clients will only renew and buy more from us if they're adopting what they bought from us. If when users are loving the systems, when they're seeing the value.
That is really what we focus on, and us increasingly moving to a usage and consumption-based model will naturally incentivize the behavior of our teams across the entire life cycle. We are very convinced that this is our path to growth as an as-a-service company. In summary, we believe we have the right technology, where delivering results and reliability will be what sets technology providers apart. We think this is the right time, and clients need us more than ever to help them build for that autonomous enterprise that we heard about this morning, to creating that for the future, but getting results today. We believe we have the right field teams focused on executing across that entire life cycle.
With that, we believe we are solidly on the path to becoming a Rule of Forty company that I'm sure you've heard Ken talk about. Certainly, we have heard him talk about that very often. I'm very excited, and I hope you are, too. I'm here to answer any questions that you may have.
First, I want to say that this is the first presentation in an Investor Day done by the sales team, where they referenced Rule of 40. I did not put those words on the slide, so I am super proud. To me, that was the biggest win of the.
You're welcome.
Sorry.
Anybody have any questions for Jacqueline? If not. All right, Steve, I know you have a lot to say.
You know, you mentioned that you're expecting 90% of ACV growth coming from the existing customer base, and, you know, really kind of focusing on that part of the strategy. I guess, how are you thinking about, you know, the net new opportunity moving forward, and, like, what structures are in place to potentially capture that? Then, I guess, secondarily, you know, I know there's a pivot away from the mid-market approach. I mean, what would need to change to potentially, you know, go after that initiative longer term to or gain confidence around potentially targeting that house again?
Okay, thank you. I will address the first question, if that's okay. I think I will refer the second question to someone else, Ken or Alan. With respect to the first question, what we have been seeing is, like I said, we see clients that are big enterprise clients start in one area of their organization. What we've also seen is that because we have this approach where we really work with our clients, we also work with enterprise architecture, they typically look at how can we embed this as an enterprise platform, also potentially for other business problems to solve, potentially reaching into other regions. If you think, for example, about an ING or a Shell, right? These are very large organizations that have business demand everywhere.
We put in place, indeed, structures, sometimes enterprise architecture, target architectures that reference Pega, or also commercial models that will allow them to quickly and easily grow into these other areas. Of course, we also have our other products. It's all based on the same platform, but we have clients that start with case management, like Rabobank did, and then adopt the CDH, customization hub, at a later stage. What we see is that, and that's why we've taken this approach, there's still so much potential in these clients, and it's less effort to sell into those existing clients. That's why we've selected that.
Well, I was going to make two quick comments on that. The first comment was, hold the question on the mid-market until you Launchpad, because I actually think that's really relevant there. The second point I was going to make is, Jacqueline, we talked about this yesterday, and I know that she has examples of this. Because of the deep relationships that we have with clients, and it's in every country. People, executives and people at one client go to another company, and that company may look like a similar profile that may actually not be a client of Pega, or we may not actually be penetrated into the division to where that person goes. When you think about new logo acquisition, that's a big factor.
I cannot point to, like, just off the top of my head, five to 10 examples of executives at one company that used to actually work at another one that had Pega or have said to me, "Yes, I know Pega because I used Pega at my last company." That's just something to think about in terms of the opportunity to expand into new logos that may be profiles of companies that look very similar to the ones we already have.
No, I like, before I get to after John speaks, we have a very specific strategy to broaden the market footprint. That's the same strategy we're talking about here. There's a good chance we're exploring.
It'll be a little bit more clear on how we see the opportunity after I go.
Thanks. Rishi Jaluria, RBC. I wanted to ask about kind of the consumption element and how that drives adoption. You know, we heard Alan again talk about that on the earnings call, but since you're on the ground and actually seeing what customers look like, what does that structurally look like in a contract where there is that consumption element? Alongside that, you talked about incentives with the sales teams to actually help drive more consumption over time, which makes a lot of sense. What do those sort of incentives look like? Thanks.
I think what is one thing that sets us apart from other organizations is that we have never been big in selling a lot of stuff that ends up on the shelf, right? We're just not built like that. Alan really doesn't like it, us, when we do that. The whole culture is very focused on selling a client what they need for that particular moment in time, solving a business problem with volumes that we can see for now. Once those people are. Once that team is successful, they may have more volume within that specific business domain, or they may want to do other things, and we sell them again some stuff.
Of course, we've moved from a perpetual license to a term license model, subscription-based model, where that is even more incentivizing us as salespeople to make sure once we sell that piece, we really need to make sure they are successful. Because not only will they not extend, but there is a point in time after a term of the contract that they may say, "I'm not getting the value, so I'm leaving you as a client." That is naturally incentivizing, not only, by the way, the salespeople, all of the cross-functional teams are also very much focused and incentivized on making clients successful, getting the renewals in, and then hopefully getting the new revenue, either in that domain, like I said, when we scale with volume as a consumption-based model, or moving into other domains.
Why don't you tell them how you're compensating, how the sales teams are compensating in terms of the consumption model and ACV? That might be interesting for your view, and not so.
Sure, sure. Happy to. We all have targets, and all we need to do as salespeople is to make sure, well, the majority of our incentive is based on new ACV, so we need to make sure that we create new ACV for clients that have been with us, and there is a small part that is based on retention as well. If someone reduces their footprint, that will go into the mix. If we sell something new, we have to make sure that we get up that ladder again, so that it's not like, oh, we can say, "Oh, this client is not so successful here, but they buy something there, and we get compensated on that." It's really looked at more broadly to make sure we are all working.
Not that it's not already in our DNA, despite what everyone may think of salespeople, but it really naturally incentivizes us to do it that way as well.
Let's wrap up this question.
Hi, my name is William Jellison. I'm from D.A. Davidson. When you're getting together with customers, and you're trying to understand their needs and goals and how to pay and how to get there, how far into the future are you typically looking for those goals? I'm wondering, when a technology that's as disruptive as generative AI comes in, do you see that cause customers to hesitate to spend because they're worried the technology might change again? If that is the case, how do you help them feel comfortable moving forward and diving into the online?
Yeah. Yeah. With respect to the first question, we try to build three-year plans that we actually vet with our customers, ideally. We look at what are these clients focused on and how much they share with us, right? Sometimes it's 18 years, 18 months out, sometimes it's three years out. We really try to build that strategic account plan, we also do tactical discovery. We look into an operation, and we try to see how much of that can be automated, how much of that can be consolidated, and what does that mean in terms of value.
Because we are very focused on creating results fast, we look at building, first, we call it MLPs, or Minimum Lovable Products, for a client that really generates results inside three, four, six months, so that the client will see typically within a year, really get ROI. That is what we do per domain that we look at. Of course, we try to set people up, and that's why it's really important that we have these enterprise advisory services now, that we help an organization already think about what could be the next step, so that you create the layer cake that Alan spoke about in a way that allows for quick and easy reuse, and then additional benefits down the line. Does that answer the first question? The second question was what again?
Whether disruptive technology-
Right. Yes. Yeah. Thank you for reminding me. The other thing that we think sets Pega apart from other technology vendors is we have this model that Alan spoke about, and being able to generate the model, the model that creates that allows us to innovate in a way that no one else can. We've always seen that before as well. When new channels would emerge, or when there was new technology available, we would be able to do that on the Pega side. We would generate the model. Therefore, the clients would not be bothered as much by it. Of course, generative AI, I think it's early days.
This is, this is massively changing everything, but it's not that dissimilar from what I just described, because client knows that they will never run out of runway, that they are future-proof because we have this model, and they've seen us do it. Some of our clients that have been with us have been through four from scratch rewrites of our architecture because we thought there was a better way of doing it, and they have seen that we can actually do it that way.
I think the key thing is, when 23 ships this summer, the layer cake the customers already have continue working. I mean, one of the cool things about having that layer cake, and I'm referring to our model as the layer cake more and more because everybody's using the word model, and so it sort of has gotten a little bit polluted. You'll have large language models this. That layer cake collection of templates, to use a language that's, you know, been used here, is unchanged. It's just we're now using Gen AI to inject assets into it. That's why I think we've got such a huge advantage from a structural point of view.
Yeah. Thank you so much. It was a pleasure, being here with you today. I'm glad to introduce John Higgins.
You have an idea to impact, disrupt, or even revolutionize the world with a new application, but getting off the ground is hard. 90% of new software products fail. It's time to change that. Take your SaaS idea and attach an enterprise-grade booster to it. Introducing Pega Launchpad, an all-new low-code platform powering your B2B SaaS apps from concept to commercialization. For the pioneers already solving for what comes next, get ready for takeoff.
Check. Rolling. Light, circuit, instrumentation, circuit on.
This is something that a lot of people and companies are looking to do. Think in the order of magnitude of about one million independent software vendors over the next five years that are looking to take the deep knowledge or IP that they have in their brains or in their business processes, and they want to translate that into a SaaS application that they can sell to tens, hundreds, or even thousands of companies or subscribers. I have a little bit of experience myself with turning experience into product, so I'll tell you a little bit about my story in that section as well.
Second, we're gonna talk about the entrepreneurial grind, and this is really about the challenges that innovators, whether they're entrepreneurs starting a new business or innovators inside of an existing business, face when they're looking to develop and commercialize a SaaS application. Think of that as like the things that drive cost and time, and that, you know, ultimately result in the 90% of the software projects failing that we just heard about in that villy video. Number one in our countdown is talking about Pega Launchpad. Of course, I'm gonna talk about Pega Launchpad the whole time that I'm with you here today, I'm gonna break this actually into two subtopics. First, I'm gonna talk about how Pega Launchpad simplifies the development and operations of applications for our application providers.
Second, I think most of us in this room are probably pretty familiar with what Pega Launchpad can do or what Pega Infinity can do with the world's largest organizations, helping them build for change with applications that they build and deploy for use inside of their own operations. I'm gonna talk more then about how Pega Launchpad is the low-code platform for app development that empowers our software providers to build and commercialize their SaaS applications that they will take to market and sell to others. You can think about Pega Infinity in this regard as those internal enterprise use cases and Pega Launchpad as built for those software providers who want to commercialize apps themselves.
With that, I'll start off with topic number three, and the translation of experience into product, and I'll start with a little bit about my own story. I started my career at Rogers Communications, which is Canada's largest telecom. I was fresh out of school. I was wearing a headset on my head, and I was talking to their wireless customers. Eight years later, I was their vice president of client management. I ran their call center strategy budgets at a number of their operations with a team of about 6,000 people.
At the age of 30, I was the youngest VP in the history of company that did not have the last name Rogers, and I stayed at Rogers for four more years, sitting at their leadership table and learning everything that I could about contact centers. In 2010, I left Rogers to found a customer service-related software company of my own, with a product that we could sell to large-scale enterprises that had big contact centers like Rogers, leveraging the experience that I had built there. My company was InTheChat. We were an early-stage player in the social media customer service space, and then extended our platform to SMS, chat, email, and messaging, bringing all digital channels together on a common platform.
InTheChat really took all of the business processes or workflows that have been established for voice over the prior 40 years and turned it into an application that could power customer service over convenient digital channels or even chatbots, which were radically new at the time. InTheChat actually ended up powering digital customer service for companies that included JPMorgan Chase, TD Bank, Universal Studios. We were a beta partner with Facebook Messenger and with Apple Business Chat, and the company was so compelling that in 2019, Pega sought to acquire it. I was so excited about the potential to put our channel capabilities together with Pega's workflow, automation, and AI-powered decision capabilities. Like, how could I turn that down? I spent the next two years integrating our technologies and running the customer service business at Pega.
About a year ago, Alan and I had a conversation, and here we are on the Pega Launchpad. A couple things that I want you to take away from that story, it's not just about me. One is, just going back to a couple slides ago, when I was at Rogers, I was the target buyer persona for Pega Infinity. I was, you know, working at a large-scale enterprise, running a contact center that had a lot of workflow, automation needs, and all that type of great stuff. When I moved to InTheChat, I became the target buyer persona for Pega Launchpad. I was building a SaaS application. I was translating my experience into product and building a SaaS application that would get used by multiple companies to drive a business-critical use case.
That's why I get really excited about Pega LaunchPad, because Pega LaunchPad enables business leaders, like I was, to turn ideas, experiences, or their IP into software-powered businesses quickly and cost-effectively. That's really important for somebody like me, because the challenge for me, like it is for those one million ISVs that I mentioned a couple of minutes ago that want to get started into the software space, is that it's really hard to build and commercialize your app and to take that to market, and that's why we're going to talk about the grind next. I always find this to be a really interesting picture. If you don't know what that is, that is a space shuttle sitting on a transportation device that's taking it from its hangar to its launch pad.
Does anybody know how long it takes a rocket to get from its hangar to a launch pad? What's that? 17 hours. Close. It takes 10 hours. I know. I'm totally gonna get there. It takes 10 hours for a rocket to leave its hangar and get to a launch pad. That particular instance, the shuttle had left the hangar, went to a launch pad, had a technical issue, and then had to go to another launch pad, which took another seven hours for them to move. The parallel for me with InTheChat, you know, I talk about my experience there, is that everything took longer and cost more money than I ever could have forecasted.
I mean, the amount of time that I had to spend as a business owner, even just raising capital, was such a distraction, let alone the dilution and the giving up of control that had to happen, you know, for me in growing that business. I really needed that money because building SaaS applications takes a lot of people, time, and effort. They're writing code and testing and QA-ing the code. They're spinning up cloud instances and servers. They're writing APIs, and the list goes on. Then, beyond that, traditionally, there's a lot of technical stuff required to be able to build an application. InTheChat was built from the ground up using Java and Rails and working with AWS and MongoDB, Atlassian, and others.
You pay money to these vendors, and then you also have to pay money to have employees who work with them and build with them as well. That's really what the entrepreneurial grind is all about. It's the time, energy, and money spent, often before you've even signed a client or generated any meaningful revenue, to build many of the same things that every other software company has had to build. They had to build a SaaS architecture, the app that it would host, and the means by which clients would use it. For my team at InTheChat, the grind was all about coding that app from scratch, hosting it, and building the capabilities to handle our subscribers. Those activities all slowed my time to market, they drove up my costs, and they impacted my margins, but I had no other choice.
That's just the way that software development worked. Then I got to know Pega, and I could see what Pega was able to do with the large-scale enterprises or leading organizations that were developing applications for implementation inside of their own operations. I thought if we could take those same capabilities and extend them into the provider world like I was, we could change the world of commercial C, B2B SaaS application development. Then we could also take Pega into pretty exciting new markets that we never had access to before. By my analysis at the time, if I had had Pega when I was building InTheChat, I probably could have been in market about 60% faster. I would have had...
been able to cut my development cost by about 30%-50%. The result of that would have been that I had margins in the first three years of my business that would have been life-changing. You know, well, not only would they have been life-changing, but also then you think back to that, like, financial raise conversation, think about how much less I would have had to raise, think about how much less dilution there would have been, think about how much less control I would have had to given up.
You know, the other thing I was going to mention on here is just what was crazy was with Project Phoenix, that Alan actually announced on stage back in 2019, just a few doors down from here, three weeks after I joined the company, many of the capabilities that we would need for Launchpad, like the SaaS and microservices architecture and the React-based UX and multi-tenancy, those things were already being delivered. This brings us to countdown topic number one, and that's Pega Launchpad, the low-code application development platform that empowers our software providers to build and commercialize workflow-centric apps. With Pega Launchpad, we have taken the power of Pega Infinity, with its prebuilt and reusable components, UX, and easy connectivity, and we've added a scalable Pega Cloud-hosted SaaS architecture and everything that providers need for subscriber management and multi-tenancy.
We've packaged that all together in a new product that significantly reduces software provider development and operations time and cost. Pega Launchpad enables our providers to cut through the entrepreneurial grind. It enables reduced costs, reduced build times, reduced vendor counts, and it does a lot more. It brings the power of Pega to entirely new markets that would have never had access to Pega before. We're doing that through a new go-to-market model in which our application providers provide our growth engine. In that growth model, Pega Launchpad provides the app development and operations platform to our providers. Those providers build and commercialize their apps, selling them to our to their subscribers. The subscribers add users and drive usage, and that usage generates revenue. Pega will take a share of that revenue.
Our early adopter providers, who are already building on Pega Launchpad, include system integrators, professional services firms, enterprises, existing software developers, and startups. Those companies are selling to subscribers and taking us into new verticals and new subverticals. They're going to bring us companies of new sizes and scales, as was being asked about. They're going to bring us new use cases, and they're taking us into new business models. As examples, some of the apps that our early adopters are building are designed to be sold to subscribers that include small to mid-size auto parts players or global pharmaceutical companies that are in clinical trials, and large enterprises that are implementing ESG standards, and even local governments that are looking to drive significant efficiencies in their inspection services.
These are all areas that we, as Pega, would not have pursued on our own, and those subscribers, those buyers that are looking for pre-built apps, they wouldn't have had access to Pega. Our providers will bring us together beautifully. Hopefully you can capture some of my enthusiasm about Pega Launchpad and see, you know, that I'm really excited about Pega Launchpad for three core reasons. One, I'm excited to help our providers turn their ideas into products. I'm really excited to help them be successful faster by eliminating a good chunk of that entrepreneurial grind. I'm also excited to see our providers deliver valuable solutions to their subscribers as a new market of users who may have never had access to Pega before.
Finally, I'm also really excited about the fact that at Pega Launchpad, with Pega Launchpad, we have achieved all of our targeted milestones for the product so far over the past year, and that includes having a product that was ready for our early adopter providers to put their hands on keyboards with, having those early adopter providers build their first apps, and now getting ready for our first subscribers to start using those early adopter apps. There'll be lots for us to stay in touch about as we head towards 2024 and beyond, as we grow out Pega Launchpad and take it from the Launchpad to liftoff. First, I'll give you the opportunity to ask some questions in here. One coming in. Two coming in.
Who do you view as your competitors?
You know, it's an interesting thing. Like, there are others that have, you know, solutions that are more traditionally pre-built that, you know, struggle to be able to deliver these. Like, you may hear about, you know, companies that have been built on other platforms in the past. What we find is that the agility, the microservices, you know, native structure, and the multi-tenancy, you know, are big parts of what makes us different in this regard. This is a pretty nascent market, I think that you would find. You know, at the moment, we're not in a place where we're bumping up against folks. You know, like, you would hear all the same names that you would typically hear about in enterprise, you know, normal enterprise, SaaS, app development.
I don't think anybody would be surprised by any of those names. What we're finding as we go to market, and again, it's very, very early days, is that the capabilities that we have with Launchpad are very differentiating in terms of how easy it makes them to get started and the fact that it's got the subscriber management capabilities built in on top of the app development capabilities built in.
Hey, obviously, appreciate all the color. If we think about, you know, part of this is the value proposition.
... I guess in the past, where we've seen Pega go down that route before with Pega Express. What differentiates this from Pega Express? I mean, what learnings are there from that we can kind of say, "Well, here's where we can get better traction with Launchpad than we were with Pega Express in the past?" Thanks.
Yeah. I think the biggest, you know, thing is, you know, first of all, this is a multi-tenant, you know, solution, which we wouldn't have had at the time that we were doing, Pega Express. There's a very different play. Instead of having to spin up instances every time that you want to create a new app and incurring the cost and the difficulties with that, you know, we have something that is much easier and lighter in terms of the multi-tenancy. The entire subscriber management component is built. Like, this is truly not built for Pega to go to market after the mid-market. This is built for Pega to enable software providers to build their own business, going to market to sell the applications that they build. I think that that's what the big differentiator is.
You know, really, this is an engine to enable providers to go to market for us.
Great. Steve Enders from Citi. I do want to ask about the go-to-market aspect here. Like, how do you get this out there and attract partners and other vendors in the market to begin utilizing Pega Air?
Yeah.
I guess, maybe from a money today perspective, you know, I know you're talking about how much cheaper it is to get up and running here, but I guess, how does that translate to the marketing potential for Pega Air?
Yep.
Thank you.
All right, first I'm going to go to, you know, how are we drawing attention? To be honest, Steve?
Yeah.
So far this year, we've been very muted on Pega Launchpad. You know, one of the things that's really important to us as it relates to Pega Launchpad is not overcommitting and under-delivering. We're very focused on building the product, getting that product solid. We are now working with, you know, we have multiple, you know, still single digits, but high single digit numbers of companies that are building applications on the platform today. Our intent is to continue to meter that as we go along. We've built the product out. It is still in development, it will be in development for a bunch of time, but we are at a place where providers can build. We're very metered on how many we can bring through.
When we announced this, you know, last year in July, we put out a small press release. We didn't promote aggressively. We really just wanted to make sure that you folks knew that we were going to participate in this space, and we didn't really want to drive a whole lot of activity. We had literally hundreds of companies that submitted interests, in the first week of that announcement, that we did no profile. We issued a small press release, and we put up a little website and said, "Sign up here." The interest is very, very high, so we haven't gone and pushed aggressively on marketing at this point.
You know, Alan talking about it this morning on stage, I think will probably inspire a large number of the partners that exist inside of Pega to come and play with us or at least want to talk to us. It's really for us, at this point, two parts. One, we want to make sure the players that we're working with actually want to build a product company and actually want to take a product to market and want to become, you know, an application provider. We really do the quick check with everybody on that before they go, that it's not just somebody that want, you know, is thinking about this. They actually have a plan to take a product. They actually have a way to commercialize it. They have a way to sell it and move it through. That's the first step, and the.
The second step is where, you know, how does their application fit with what our capabilities are at Launchpad? We work with them to define that. The first eight that we have, that we've brought on, great partners that we are working with that are building those applications. I guess I gave the number, I told you it was high digit, high single digits. The first eight that are on with us are, you know, a real mix, as I mentioned, all the way from startup to large scale SI companies that are working with us to start building their applications and take those to market. Not a huge push on for marketing for us at the moment.
You asked about the monetization, and I gave you a really long answer on the first part. On the monetization side, it's a revenue share model for us, and we have, you know, protective provisions as we build the contracts with these customers or with these providers that says, "You know, we're going to do a revenue share with you, depending on..." You know, we work with them to commit to a certain amount that they will sign up for. That determines, you know, that influences the revenue share.
There are protective provisions for us that says, "You know, we will not make less than X." I think, you know, Ken being here, we at this point aren't, you know, we aren't forecasting, I think it's fair to say, Ken, like, Pega Launchpad isn't in any of the forecasts at the moment that you will see, because we are still very much in the early learning stage on that.
I even think the economic model is probably something we'll learn a little bit, too, right? Just understand the levers that you store. I think John touched on it. It's a revenue share model with certain floors and protective measures around how they price, so that you don't actually get buckled in when something else is selling on it. There's protection there, but I think there'll be some things we learn from that.
We should probably move on, John, to our next speaker.
Okay. With that, hey, let's hear about some more, great Pega innovation coming, from two of our product leaders, Steve Bixby and Kara Manton. Thank you very much.
Thank you, John.
Awesome.
All right.
All right. Well, thank you to John. That was awesome. I think we can all agree we're really excited about Pega Launchpad. We have the privilege of being here to talk to you about Pega Infinity and what we've been doing in the product organization, since the last time we were all together. My name is Kara Manton. I'm the head of Product Operations.
I'm Steve Bixby.
Um-
I'm responsible for product engineering for the Pega platform.
Yeah. On behalf of all of us, you know, just thank you for being here to ask all these insightful questions and learn about our product. Pega Infinity delivers operational efficiency through our intelligent automation platform. It delivers seamless customer service, and it personalizes one-to-one customer engagement. We do that all across dozens of industries with many specific use case applications, but today, we really want to talk about our low-code platform. We are the low-code platform for AI-powered decisioning and workflow automation. This term, AI-powered decisioning, is definitely taking on a bit of a new definition today. I think we can all agree. We're talking a lot about generative AI, and I'm excited to show you it today. I'm excited to talk more about it today.
Pega GenAI is like taking a consultant with five years of experience and having them make a result on day one, right? We really believe that Pega GenAI is going to allow our clients to build more workflows, to automate more processes, to make more decisions with less skills and in less time. Of course, that's important, as we've been talking about today, because more decisions, more cases, getting work done, that's how we grow. We're going to show you why Pega and GenAI is a force multiplier, and we're going to do that with a live demo as well, which will be hopefully pretty fun.
Which will be totally nerve-wracking.
Yeah.
Exciting for everyone.
That, too.
Yeah, all right. I want to use a little metaphor, and I think the setup has been great. Hearing your questions, I think, has sort of given me some juice here to try to get you on board. Hopefully, people are familiar with Lego, right? Everyone's got Legos. I know, we know there's loads of them in my basement. I think a lot of families have that situation. There are actually 200,000 unique Lego bricks, right? Building blocks that help you build, you know, a cool Land Rover, say, right? Taking a pile of bricks and trying to get to that result is hard. Which is why every Lego set comes with step-by-step instructions, right? It's going to tell you exactly how to get to that end result.
In fact, they're going to give you only the pieces that you need to get to that end result. You know, it could be 300 pages worth, in the case of this Land Rover, but nonetheless, it's a step-by-step set of instructions. When building an enterprise business application, trying to get to that result can be, you know, far more daunting than, you know, building a truck. Pega, as we've kind of been the theme here this morning, this idea of templates and the model and the layer cake, we have about 200 plus unique building blocks in Pega. Realistically, it's only about two dozen that you work with every day to build an application, but we have built out these concepts, these templates, that then get populated and ultimately can deliver this. There is no instruction manual, right?
What you do, you have subject matter experts, you have consultants, and per the question earlier, these are the people that populate those templates with the business information. Like, what is an auto loan? What is a mortgage? What are the steps that need to go through? What is the data model that we need to populate? What are the systems we need to connect to? Generative AI is like putting a fast-forward button on that process, right? Making every person involved that much more powerful, allowing developers to do things they couldn't do previously, allowing end users to have more power to deliver more capability on the front end. What I'm going to show you today, in just a few minutes, is going from, you know, an idea to a working application.
We'll talk about what was a little bit different in how Kara did this earlier today.
While Steve gets ready to show you this live demo, just to kind of recap, if you were here this morning, you saw me on the keynote stage. I had the privilege of building an app as quickly as I possibly could. I did it, I think, in, you know, two minutes or so. What I did is I used Pega Gen AI. I said, "Can you give me the stages and steps of this workflow?" I said, "Yes." It gave me a data model. It gave me sample records. It gave me a UX built right out of the box. It translated it to Turkish, which is a little nerve-wracking, but it checked in with me. We kind of call that the human in the loop of Gen AI.
You know, we send out a prompt, it sends a return, and we choose whether to accept it or not. I could have changed it. I could have rejected what it sent. What we're going to show you here today is a little bit different. We're going to let Pega GenAI just run the show. We're going to say, "We want to build an app," and we're going to see, let it go and have it build the entire thing and return it in the end and see what the result is. It really is live. Let's.
I'm going to be honest with you and tell you that I tried this during the keynote this morning, it did not work. The request timed out. I did it again. It worked. We're going to experience this together. Here we go.
Of course.
I'm going to log in. This is Pega Infinity 23. This is an application that's sort of empty, but I did create, like, a fake company called Skyline Insurance. So far, the only context I've provided is that this is an insurance company, say. Right? I'm going to skip the traditional way, and I'm going to actually open up the new Pega developer assistant and switch on the Pega Gen AI suggestions. As Kara was describing, the way she built that application is she let the Gen AI just keep suggesting things to her as she navigated around. We're not going to navigate anywhere. We're just going to look at this thing and say, "Can you just build an app for us?" It's already suggested a couple things. Like, you can ask this any question.
Earlier, Alan talked about this technology called embeddings. We've already taken every piece of public documentation about the Pega Infinity platform and fed it into a vector database so that you can ask any question at any time, and it's going to give you a, you know, professionally authored response, but using our content. Right? This has already become hugely valuable to us. You could say, "What's a case type? How do I build a workflow?" Et cetera, et cetera. What we're going to do is just build a workflow. I could either type it in the bottom here or just click it, and it fills it in for me. I'm saying I want to build a workflow. All right, first click worked. That's awesome. Let me help you build. Here's a few workflows you may want. It's actually suggesting a few.
I mentioned it knows that we're an insurance company. It's like, how about a quote request, a policy application, a claim submission, or a renewal? Those are all great suggestions. In fact, the quote was kind of what I want to do. I was just going to type in auto insurance quote. That's the workflow I want to build. I could literally type anything in. When we were playing with this, we even did things like llama rentals. Unbelievable what it would suggest as these the processes. The data model was like, "Is this a mountain llama, a spotted llama?" Like, really cool. Maybe not as relevant for our business as auto quote, but-
We'll try that out.
Let's go. All right, this is the moment of truth. I have a backup video if this doesn't come back. What it's doing right now is it's, you know, again, to use that metaphor, it's populating the template. It's defining the stages and steps for the workflow. It's defining the personas that need to participate in the process. It's building out the data model. We may see it put some decisions in place. We're gonna see it for the first time together, based on what it responds from that prompt. I'll just keep talking until it hopefully comes back.
Maybe you can pull down the layer.
Yeah, while we're waiting, there is a layer cake picture there in the background, that sort of shows... Oh, we came back. I'm very excited.
Just in the nick of time.
All right. I hope it built a good one. Okay, successfully created auto insurance quote. Click here to open the workflow, preview it, or generate a mobile app. All right, let's come back to that one, let's just open it. I'm gonna open up that first piece of the model, which is the workflow itself, right? Behind this sits a bunch of process flows and all that. We call this the Case Designer or the case life cycle. This actually looks pretty interesting. It's not the most robust one I've seen it generate, we've got customer information, quote calculation, policy creation, and then an email confirm. There's a couple of approval steps, which is cool. There's a data model we can click into and see all that, let's just run it. Like, we typed in three words.
Let's run it. I'm gonna say, "Good enough, save and run." I could have also clicked that preview button at the top or clicked the thing in the panel on the right. Here we go. All right, we've got our first insurance quote, case, A-2025. Here's the data model it provided us: contact information, make, model, coverage. What's this? Coverage types. All right, it populated some stuff, a bunch of different data types on here, a couple true-false questions. All right, I'm gonna use another generative AI feature here. We talked about this during the keynote. This is the fill it with sample data. This is just kind of like a little accelerator, you don't have to, like, fill out the whole form every time as you're testing. It's also the same thing that generates all that test data.
This just populated it with appropriate information. John Doe, 2021 Toyota Camry.
Nice.
-for $20,000. That seems like a pretty good deal. Anyway, so it's doing this, so I can click through this and move along. I want to show you a couple of other things. This one's maybe not as exciting as building an entire application, but one of the things customers often do is change the theme of the application to match their brand. One of the things that I like about this is A couple things. I could ask it, "How do I do it?" It could tell me, or I could click over to the Settings tab, click Theme, and do all the steps that I know how to do because I'm sort of an expert in this product.
Then it would ask me this question: "What brand do you want to use?" I'm gonna say, "Brand my app." "Let me help you build. Type the name of a brand." Instead of just saying, "Oh, I want to make it green or yellow," or whatever, it's actually gonna give me a. I'm gonna give it a brand, and because it has the world's knowledge at its fingertips, it will actually figure out, you know, what's the color codes that best match this brand. This is an insurance app, so I'm gonna use a well-known insurance company. Let's do State Farm. Let's see. Okay, yeah, red. That's the red. "You want to do it?" Yes. Try it. This should take that, the State Farm red-
Nice.
Put it in. There it is. If I open up that quote and sort of see, yeah, like that. I'd buy that as State Farm. Do one more. Brand my app. Liberty Mutual is a recognizable color set. Yes, I, you know, definitely cheated and looked at what are the ones that are the most recognizable. My favorite one was actually T-Mobile, which was hot pink, but it wasn't an insurance use case. Here we go. Nice. All right.
Bright.
It is bright. It looks, you know, looks like a, you know, this could be Liberty Mutual. We're gonna get to the point where we can just ask it, you know, grab the logo, pop it in, all of that. That's all coming, right? We're not quite there yet. I mentioned earlier that when this thing was built, at the way at the top, it suggested generating a mobile app, which, just to make this even more risky, I'm gonna try. This should do is yep, return a QR code, and if I pull up my phone-
Ooh!
-here.
Yeah, this is generating a native mobile app that theoretically anyone could scan that QR code.
Hi, everybody. Okay, here we go. Scanning the QR code. Install. Download. I'm not on the Wi-Fi because I found it was a little flaky. Come on, baby.
Well, thousands of people joining the Wi-Fi at the same time.
Yeah.
Sometimes do that.
It's working. All right. There we go. This is cool. The mobile app also got the branding, like the theming, the color scheme, and all of that. Let's get started. Got to sign in. Yes, push notifications, and let's load this thing up for the first time. All right, cool. You guys can see that? Here, I can make it full screen if you like. Now you're seeing the mobile app that should be able to do an auto insurance quote. There we go. Auto insurance quote. Pop it up. Here's all that information. I'm gonna show you one more cool generative AI feature. Just to the left of the next button there is a little scanner thing. This allows me to use the camera to scan documents. Let's see, I have a registration certificate here.
This is actually from my Jeep. If any of you want to send me a gift, hopefully, it will have my.
Steve's address.
... address up on the screen in a minute. What this did is it took all the information from that form, extracted it, and then also sent that with the fields on the screen, and the generative AI then matched them up and populated it for me. This form is now in this phone.
We're never typing in a field again.
Exactly. Yes, there it is, the Jeep Wrangler. Same spelling on here, by the way. That's not a typo.
That's the RMV getting that right.
Yeah, there it is. Did this data model create a VIN number? It did not. There, I've done this in the past where it pulls the VIN, and it's impressive to be like, "All right, that's awesome, 'cause I never want to have to type a VIN number." Anyway, so we're using that to capture this. Let's click Next. Quote details. I want to get to one of those approval steps. Okay, there we are. If this is all working correctly, I should get a push notification for that approval. While we're waiting for that, let's do one more thing here.
Because this mobile app uses the best practices and all the, these templated building blocks that Pega provides, it has things built in, like the audit trail, that's gonna capture everything that happens, you know, behind the scenes with this workflow. It has utilities, like the ability to do attachments and followers and tag things. It has all the information that's being collected, captured in the app. It also has a activity feed that I could post to. If I wanted to post, like, a picture of that Jeep...
Maybe this one?
That's a good one.
Yeah.
Thanks, Kara.
Yeah, you're welcome.
All right. Oh, it's gonna be blurry. Sorry. I could even, you know, post a note and say, "Check out my Jeep," and then post that. Let me get out of the full screen. It's working. Like, we've created an app that, if nothing else, is the equivalent of a ticketing app with workflow. Like, it's capturing information, it's creating a case ID, it's routing the thing around, it's adding approval steps, it's letting me do very practical things like this, like to communicate with it, so I could bring other users on board. Finally, let's go back and look at the desktop. This is A-2026. If I go back to the home here, I can probably click on that one, 2026, and then click over to Pulse.
Hopefully, there it is, and there's our picture.
Where's the quote?
Where's the what?
Quote.
Where with a quote? The question was, where's the quote? Quote is. Oh, we didn't go through and get the actual quote. That's the, that was the question. No. In fact, I would not trust this to, like, build the rules, the rating rules, and all that stuff. In practice, that's typically done in another system that we would make an integration to anyway. As we're gonna continue to discuss here, the beauty of the generative AI is not that it builds an end-to-end application, and we all just kick back and drink margaritas. It's that it gives us an awesome place to start, and you're never lost, right?
I work with clients all the time that are saying, "You know, the barrier to entry with Pega platform is trying to get familiar with all of those building blocks, so you know how to put them together," right? With generative AI, with the ability to just query the system and just give you the documentation, with the ability for us to guide you at every step of the way, and with the ability to create an app that gives you an unbelievable starting point, I think this is incredibly powerful technology. Like, I think this is freaking awesome. I've built, like, I don't know, hundreds of workflows in the last week, partially just for fun, like I do with image generation, right? You just generate some images for me. It's fun. I've generated llama rentals, donkey rentals, whatever it is.
Then I look at it, I go, if I don't like it, I just go, "Nope, delete it," and it just wipes everything out. If I like it, I go, "Yeah, that's good. Let's start working with this." Then I can start to ask it to refine it even further. Hopefully, this gives you a sense of why those building blocks in the Pega model give us the leg up. We're not generating code. We are coercing the prompts to give us information that's, like, interesting. Give us information about the industry, about the business process, about the data that would get collected here, and then we match it up to our models. So that you're not stuck with a, you know, pile of code that you can't understand. It's the same stuff.
Like, you know, this visual here is something that anyone, I think, can understand. Oh, God. Okay, there we go. Anyone can understand what this business process is doing, right? This is what you lose in a code-based system. In fact, there are no other low-code vendors that have a picture like this, right? Again, I think Pega plus generative AI is, you know, big time. We'll get into it a little bit more.
Yeah, let's talk more about it.
Thank you. Thank you for rooting for me through that.
He needed it. He needed that moral help.
All right.
Can we switch back to the slides?
Darren, you can go back to the other one.
Great. Yeah, I mean, as Steve just said, it's an excellent starting point, right? We're not naive. We understand that we'll have to bring in professional developers to do that integration into the existing systems. It really does put app generation on autopilot, right? It did all these things. It created the workflow, the stages, and steps. The document OCR is very cool. It goes by so fast, you almost think, is this, you know, 'Couldn't we already do this?' No, like, that is brand new, supported by generative AI. The great thing about Pega is that once you build that app, and you wanna change it in a couple of weeks, in a couple of months, our model-driven approach allows you to do that quickly. You know, we are built for change.
Awesome.
Good.
All right, Alan talked about this in the opening about, you know, AI heritage at Pega. It's something we've been focused on for decades, right? Our acquisition strategy has been very focused on AI over the years, including InTheChat with John. Sorry, I didn't put your picture up here, John. We have a team of experts that have been focused on bringing AI to practical use cases at the world's largest organizations and dealing with some of the real considerations when you bring AI into the world's most highly regulated industries. Right? Back in 2017, Rob Walker introduced this concept of the T-Switch or the transparency switch. Some of you might have been here for that.
Where if you are operating in a world where you have to be able to audit exactly why a decision is made and how it got to that conclusion, if you set that switch, we will restrict the AI to only do things that it can explain. Right? In 2019, he introduced the Ethical Bias Check or empathetic AI. That makes sure that you're not introducing bias into these decisions so that we're not overweighting or a different constituency or a group. Right? We've been thinking about these problems for years, this idea of governance and security and managing risks of AI. They're not that much different with generative AI, right? We need to be focused on that, and we'll touch on that in a second. Everything that we've done up to this point has delivered tremendous value.
Yeah. I mean, our existing AI success is vast. You heard from some people this morning. I love these three stories. You know, Wells Fargo, with their 10 times increase in engagement across their 70 million users, all using Customer Decision Hub and doing Next Best Actions. I can't say that three times fast. Aflac, we saw on the main stage, automating millions with their customer service product, using email bots, chatbots. The savings that Sheila talked about were pretty remarkable. The Navy Federal Credit Union, they were actually uncovered 300,000 person-hours by automating processes within their organization. That's a huge savings for them. They're a great client who is constantly finding new workflows to automate on our platform.
All of those benefits that Cara just described there for these clients, the clients you heard from this morning, they're able to derive these benefits using the existing AI capabilities of the Pega platform. A classification of AI that sometimes I like to call left brain AI, right? The left side of the human brain is very analytical. It's responsible for, you know, making decisions. It's very data-centric. That's where your mathematics happens, all of that. What generative AI is, think of it as the right brain AI. It's where intuition and creativity and the ability to generate and things like art, that's where it comes from, which is why we saw that crazy painting that Alan put up this morning.
What we're really seeing here is that we're just opening the door to these new possibilities, this new technology of AI, which isn't all that new, to be honest. It just started getting really good, right? With ChatGPT and these availability of APIs that we can call, it's now possible for us to use this for very practical things. It's going to still require that we focus on governance and security and managing risks and things like that, but we are super well suited for that. I'm crazy excited about this. I hope you're feeling it. My whole team is so energized by this. It's been an unbelievable journey since November, the team just getting so excited about generative AI, and then once we had access to these APIs, then we started building our own local models. Just really exciting stuff.
I mean, we had almost the entire product engineering organization do a hackathon where we got hundreds of ideas, many of which we're gonna be seeing in Pega Infinity '23. We'll be releasing Pega Infinity '23 later this summer. Pega GenAI is, of course, one of the critical components we're very excited about, but there's a lot in the platform. If I can just take a couple more minutes, I would love to talk about a few of them, one being our Constellation user experience. Our user experience is one of those building blocks that Steve was talking about. Constellation gives a new, modern user experience across any form factor. If you're on your desktop, if you're on your mobile device, if you're in the front office, if you're in the back office.
any industry, across any language, Turkish included, but, you know, even languages that read right to left, right? It gives a consistent experience and, of course, any brand. All that, 100% consistency, with the ability to build apps up to 40% faster. Companies like LeasePlan, who actually are here and are speaking tomorrow, if you're still around, they're a fleet management company. They used Constellation to build an app starting in 2021. Since then, they've built 11. They're currently in 29 countries worldwide. They're looking to grow to 60 in just the next couple years. What they say, anecdotally, I was chatting with them, and what they said about the Constellation technology is that they can deliver two to three countries every couple weeks.
That allows them to take a single app, deploy it to many countries, different languages, but all have that consistent feel. When they need to change that app, using the layer cake, they can do that quickly, right? They can make a change, it can go to every single channel, which really empowers them to be successful.
Yep. What enables that, what you're seeing in the little animation here, is instead of trying to manipulate the UI for that Fleet Ops desktop app, they simply are saying: "You know what? These are the primary fields, these are the secondary fields." Primary fields get a different treatment. That treatment is based on the template that they select, right? This is what empowers them to build so quickly and to deliver this across the channels.
Yeah. Another great capabilities are new accessibility capabilities. It's really going to allow us to expand our footprint into public sector clients. Accessibility is all about inclusion, diversity, and it's not really an option anymore, right? Especially in the public sector, you have to have accessibility. We have, you know, many clients who are excited about that. Of course, process mining. Karim announced it on the main stage today. Process mining is all about uncovering how processes actually work, not how you think they work. What is actually happening? What are those divergences? Fixing those problems that we find through process mining lets us have new workflows, new decisions, and if I can say it one more time, you know, that's the engine for growth. I think we'll end.
Thank you.
Yeah.
All right, just to wrap up, Pega Infinity '23 is all about improving this experience of building applications so that people can build faster, improving end user experiences, and then, you know, getting AI injected everywhere. With this explosion of generative AI and our ability to integrate this so quickly into the platform, so that we're generating these models on your behalf, giving you information that's relevant to help you do your work. Giving tools to end users to make them more efficient, like the scanner that I showed, the ability to, you know, fill data. We are as excited as we've ever been. I'm really, you know, honored that Ken and the team asked me to speak to you guys. My goal was to help you understand kind of how it's working a little bit better.
Hopefully, we were able to achieve that. I think we're going to be taking questions in just a second.
Yeah.
I think the first question, though, that I want to ask of Ken, is maybe you could help us explain the monetization strategy here.
Sure. I mean, this was a question that came in from a few of you. I think that certainly this is our early view of monetization, but just start with the bottom of the slide, because that's the punch. That's the, that's really the end point, which is more volume of usage in a company that is predominantly getting value from usage. We want clients to drive more applications, more use cases, faster time to production, lower cost to deploy, lower cost to manage. That's what drives the monetization for us, not selling a package of something necessarily, but the value that it gives to the clients to be able to further adopt Pega faster and get to higher volumes.
Two specific areas that you've heard today and you actually heard on the main stage, was really the ability to get a quick start, right. To be able to get started with the application, to make sure that you're leveraging the Pega resources that you have, the skilled resources that you have, on the right things, and being able to get that quick start so that you can... Steve talked about it. The application that they built here is not one that you would use, but it gets you a long way down the path, right. It takes a lot of time out of that build so that you can actually put your resources on finishing the app and really making... That's kind of that first theme.
The second one is really kind of that differentiation of how you scale and sustain, right? That's another theme. Alan talked about using the layer cake and how the layer cake becomes, these are my words, kind of a governor, so to speak, or a structure to allow you to ingest information in a structured way and not let the AI get out of control and actually you end up with something that you can't manage. That's the risk with...
You saw the example that in, I think it was Krim's section, with the prompt to code, and you're just prompting to code, and you're like: "Oh, wait, there's a bug in there." How in the world would AI be able to go back and actually manage all the, all the kind of spaghetti of, like, the different prompts and the code and the bug? That's, I think, a really important thinking about how Layer Cake, our Layer Cake, our model, is going to create this structure that allows us to leverage AI. The main key point is the bottom. Faster consumption means higher consumption, means more apps, means faster time to production. Clients see value, and they see value, we see value. That's how, that's the way we think about the value of AI.
We are running-
I was thinking.
We are running. I said we'd stay here as long as you guys need us to. I just wanted to just pause for one second. If there's any questions on AI for these guys, and I know Alan probably has to run 'cause he's. I wanna give it just a second. If there's anything else that wasn't asked earlier around AI, Brown?
Sorry. No, no.
Yeah. Hey, Brown McCullough from Ranger. Just a quick one on what you're actually seeing. I know it's super early, but in terms of client experimentation with regards to this app development. How much of an issue, if at all, is the fact that sometimes the AI is wrong or has errors?
Yep, it's a good question. That's the question, I think. We have this executive briefing center in Cambridge. We've had clients coming in. Those are really the only clients that have been able to see this. I've been showing this off since the moment it was available to show, because I think it just blows people away, and clients want to see, prospects want to see: what is our strategy in regard to GenAI? The feedback has been incredible. I don't know of any clients that have been hands-on other than in the briefing center, we don't have any real feedback on that. That will come as soon as Infinity '23 is out the door. The second question was about, you know, what if it's wrong?
I mentioned as I was even preparing just to do this demo, I'm like, "I'll do an auto insurance quote, that's pretty straightforward." It's pretty different each time, right? Sometimes it's really good, sometimes it's really bad. Sometimes it doesn't even have vehicle information in it, and I'm like, "Eh." You know, we can improve the prompts to say, "Oh, always include this or that." I think it's about understanding the tooling, understanding what it's doing, and, you know, taking measures to control those hallucinations, to control the things that are inaccurate. I talked about the knowledge library of Pega documentation. Because we use that embeddings approach, we've pretty much mitigated that completely. If we took the...
You know, you can just go to ChatGPT and ask about Pega, you're gonna get some really wrong answers because it's not gonna have the right context, right? It's gonna pull all of our documents together, pick keywords, and give you weird stuff. That's something we're very focused on, but we're gonna make sure that we're not providing tools to people that are going to put them in a tough position, where it's doing things wrong. Keeping the human in the loop has always been sort of our way to do that.
I would add one additional thought, which is the clients that I've talked to specifically around this are very concerned about the library, so to speak, or the LLM that they use, and they don't want their information going outside of their organization, and they're worried about using information from the public domain because it could be really tarnished. I think, in fact, I talked to a client this morning right before the keynote, and basically said: "Listen, we're not even comfortable with the quality of our own data, so I'm not even sure I want to use all of my own data, actually, to drive AI yet." Like, I do think that is something that clients are wrestling with, right?
Can they create certain data lakes, so to speak, or maybe ponds, that actually they're more comfortable with the data that they might actually use? Even their own internal data, they're like: "I don't even know what it would return if it allowed me to just go..." They certainly are not ready to go outside. I do think that's. Maybe that's something that you guys all know already, but that is something that we've heard.
Absolutely.
Awesome. Thank you, guys.
All right. Thank you.
Okay, so Safe Harbor, you guys have seen this. I'm gonna talk through a few concepts. The first is just kind of reaffirming the concept of what is the opportunity. I think that's been a question that's come up about whether Pega needs to or should, to drive growth, go after new logos in a much more aggressive way. I wanna. We're gonna talk very, very, very directly on that. I wanna connect the transition ending and just kind of maybe put a pin in that one in terms of the movement to subscription. The model that we're looking at in terms of the predictability of our model and how that plays into Rule forty and value creation. Let me just go...
Let me start hitting some of these slides, and naturally, we'll stay as long as you guys need for questions. Actually, I have a three o'clock, but I'll stay anytime between now and three. I think the market opportunity, and if you look at this data, you'll see some of our competitors talk about a much bigger market opportunity than this number. We've actually subset this market to the verticals and the regions and the segments of kind of the customer pyramid, so to speak, in terms of the size, to be able to be representative on what we think is the addressable market. As you can imagine, even a $78 billion for a company that's doing, you know, a $1 billion plus at ACV, that's a massive market.
Can we, you know, can we win all $78 billion? Of course not. I mean, let's be practical here. What do we need to do to be able to drive a significantly larger amount of ACV as a company? I think is well within our reach, and I'll give on a future slide a specific view of that. We have completed the subscription transition. I always fear saying "completed" because it's never done, but I would say we don't sell perpetual license anymore. We're five years in. We don't have a lot of the headwind that we had.
This is different, though, than having complete predictability and consistently with how revenue is recognized within a quarter. I want to clarify that because I think that sometimes when you look at this slide, you say, "Oh, great, then that means revenue will match ACV exactly." No, it will not. ASC 606 has prevented that for any company that sells anything other than SaaS, because as all of you probably know, depending on the duration of the deal, depending on the size of the deal, depending on if it's a renewal, depending on lots of other factors, the revenue is inconsistent in terms of when it is recognized. This is more of the view of the billing consistency of when we were originally billing up front as a perpetual, when we moved into an annual billing, and the cash flow inconsistency.
Revenue will be much more connected to ACV, it will still never be perfect within a year and certainly not within a quarter. I don't wanna mislead you by suggesting this slide would go there. As we move to an increasing amount of Pega Cloud, I would be as bold as to say, I would like for all of our business to be Pega Cloud, I think over time, more and more of it will be, our business will actually become more consistent because as a SaaS business, naturally, it's a much more structured model in terms of when the revenue comes in. The majority of our bookings are Pega Cloud. 50% or so of our ACV is Pega Cloud.
In the coming years, as the majority of our bookings being Pega Cloud, we should be moving in the right direction in terms of building more, a larger part of our business that is consistent. Real quick snapshot, we were about a 50/50 business when we started this, and when you look at 2022, the 19% is largely professional services, right? We're really, in terms of non-professional services, we're pretty much a 100% subscription business. When we started this process in 2016, if I showed you 16 on this slide, 2016, we actually had more one-time revenue, so to speak, than we actually had subscription revenue. We went from a business that was, call it 50/50, to a business that's a 100% subscription.
In this model, like I said, the blue of the 250 is largely professional services. What are the key metrics that we will continue to focus on as we finish the transition? They will be the growth in ACV. Why do we care about ACV? Because ACV connects to billings. Whether the deal is a term deal or whether it is a term license or a SaaS license, the billing parameters are unchanged. You bill in advance a year, you bill in advance a quarter, whatever the term is, whatever that specific customer negotiates, most of our contracts are billed a year in advance, with some of them being quarterly in advance. We don't bill, to my knowledge, any or substantially a low amount, like maybe a 1% or 2% of our clients may bill in arrears.
It's very rare that that would happen. Some of our usage measures, for example, if a client has a contractual element that they can increase usage, but the usage may not get measured until some point in the future, naturally, that's when we would recognize ACV. That's when it would tie to the billings. We're not recognizing ACV for anything that isn't contractually committed from the client and would associate to a billing, to an actual invoice being sent to the client. I'm not sure where that came from, but, the next piece of this is free cash flow, which I'll talk about in a second. Annual contract value we're not measuring that because we think it connects to revenue. We're measuring that because it connects to billings.
We believe billings, ACV slash billings and free cash flow is the fundamental important part of our business, and revenue is something that naturally, as we become more of a SaaS business, will become more neutralized in terms of the variability that you might have within a quarter or a year. We have a very strong subscription model in terms of the retention level of our clients. We have almost 100% gross retention. Our net retention has been above 110% forever, I mean, approaching 113%, 114%, and even close to 115% in previous years. This is...
Our business is really built around keeping that gross retention at as close to 100% as you can, and really being able to sell more to many of our clients, not all of our clients, but many of our clients, 'cause not every client is gonna increase their ACV every year. In fact, a minority of our clients increase ACV every year, and that's how we get that, you know, that kind of that historical 13%-15% growth that we're looking at. We're about $1.2 billion right now, and this kind of shows the, you know, the stair stacking from Q1 of 2022.
If you go back over time, you know, this number of ACV, when I started, I think, was like $350 million, just to kind of give you a frame of reference of where we were in 2016 to where we are now. A question that has come up and is one that I think about is, well, what are the ACV growth drivers? How much of the ACV growth driver is really innovation, and how much of it is more focused on productivity and execution? Execution also being who we target, like the markets that we target, how we staff, how we support, the go-to-market model.
I just threw this slide up, just as kind of a frame of reference of how I'm thinking about things like GenAI, process mining, Process AI, Pega Launchpad in the future, Voice AI, being really innovation levers, things that would help us drive increased value to our clients, that we would receive some of that value. Then there's things like moving, you know, our target org model and trying to drive deeper, dense focus with the existing logos that we actually do business with, pre- predominantly, at 90% plus, and moving to a consumption-based contract model. I wanna clarify what that means. Consumption-based does not mean the client is committed to zero, and it's pay by the drink. That is not what consumption is in our vernacular.
What consumption is, where a client commits to a minimum spend over a period of time, but has the ability to surge up and leverage the flexibility in the contract to drive more usage. There is a commitment that client is making over a fixed period of time. It is analogous to the way Google Cloud, AWS, Azure price with their clients. The 2 ways that you get discount with AWS is you commit to a longer duration and a higher average amount per year. Our model is not dissimilar to that. The higher amount that a client will commit in a year, helps the economics for them in terms of pricing.
The longer the duration is a lesser factor for us because of our retention rates and because the applications that our clients are typically deploying have a high level of predictability and consistency. They're not like, you know, leverage one billion transactions one year, and then the next year it's zero, right? So if we had that kind of situation, we might care more about trying to get long-duration contracts. Long-duration contracts for us can actually end up being a negative. And you might say, "Look, that's crazy, Ken. How does a longer contract commitment become a negative?" Two things. One, longer contracts customers expect better discounts. If with our retention rates, better discounts, why do we wanna give up value?
Why do we wanna give up discounts if there really isn't any value to us as the provider of that value to the client or helping them realize that value? That is just a reality. That's one. The second one is longer contracts sometimes make it harder for clients to adopt new technology, to adopt new apps, to increase volume, because a contractual event becomes a compelling event many times. Think about the clients that we're dealing with, right? The JPMorgans, the Bank of America, the Aetnas, the Highmarks, the very large companies, they actually have a very structured way that their procurement thinks about buying and engaging. Sometimes when you have long-duration contracts, it doesn't work as well to midterm or mid-cycle increases. That's a. I'm generalizing there, but that is a factor to consider.
That's the reason why, you know, we really think this consumption-based move of really focusing on usage. Usage for us is measured in an actual technical usage in our product called a case, right. A case is a unit of measure, and the cases can be measured. They can be measured in the product. The clients can see their usage, we can see their usage, we can actually mon-. Where there are types of cases, we can measure, we can measure the amount of cases that are related, we can measure parent-children cases. Like, we can come up with different structures to be able to create the measure, so to speak, of how we're going to share value with our clients.
The second piece of this is the free cash flow, which has not been something we've really talked about at all from 2017 until 2021, really, because we were in the midst of the transition, and free cash flow was just not something that was anything we could even talk about because you'd say, "Well, when's that gonna happen?" We'd be looking out so far in terms of when the free cash flow. When we started to hit 2022, we knew that we were coming to the end of it, and we actually started to be very specific in how we connected Rule 40 to being free cash flow.
You know, we're a year later from even that discussion, we actually have even more visibility to free cash flow, which I'll talk about in a second. This has been our free cash flow trajectory as we moved through the transition. By the way, this free cash flow measure is consistent year-to-year, right? Meaning I haven't changed the way that we define free cash flow in any of the years. We did $146 million in free cash flow in 2017. That was the last year of really being a perpetual business. Even that, you might say, "Well, where'd that 146 come from?" Even the perpetual deals that we booked in the fourth quarter of 2016, actually, many of those collections happened into 2017.
We were at 146. We went all the way down to -53, and we've basically been coming out the other side, and you're seeing the first significant jump, which is in 2023, as we begin to fully exit the subscription transition. For those of you that may have not heard me, just explain what I think is the simple way to think about this. In the first year that you move from perpetual to subscription, you lose essentially four or five years of billing, because that's essentially the equivalent of what that perpetual bill is, and you replace it with one-fourth or one-fifth of that, which is the actual recurring amount that you get. The second year you've lost it again, but you've only replaced maybe half or two-fifths.
The next year, you replaced maybe three-fourths or four-fifths, and then you get to that fifth year, you've kind of got back to where you started, which is not ironic that we're getting close to back to where we started in 2023, which is the year that we exit the transition. What I wanted to share with all of you today is, we've decided, based on what we've seen through the first couple quarters, that we are now targeting $180 million or more of free cash flow in 2023. We do not historically update guidance. We have chose to update the free cash flow guidance because we feel like we have really good visibility to that number, and we just felt like it was something that we wanted to share with you.
We're looking at a $180 million number, up from $150, up from $40 in the prior year. We think that much of this has to do with the exit of subscription transition, but it also has to do with our really big push for Rule of 40. I mean, you know, all joking aside, you've seen, you know, the Rule of 40 on Jacqueline's slides earlier, which I actually did not even see those slides before she did that. That is a true statement. I think what you'll also hear is, if you go out and walk the halls and you find someone with a Pega badge, and you say, "What's the thing that Ken talks about most? What's the thing that Alan talks about most?
What's the theme of Pega in 2023? I mean, maybe you might hear AI, just because. I think if you ask them of a financial metric, they would say ACV and Rule of 40. I don't think everyone completely understands the concept of free cash flow. I mean, not everybody is at the sophistication level of the people in this room, but I think they understand in their mind, we have to be more profitable. We need ACV to grow, we need to be more profitable, and that's actually gonna help us on the Rule of 40 measure. I think we've made tremendous amount of progress with messaging and alignment inside Pega around at the importance of us moving down this path. We're not anywhere near Rule of 40 in 2023, but we're making progress.
We're committed to it, and I'm happy to communicate that that number is now 180. I'll touch on just one last point on this slide, which is the criticality of driving free cash flow is not being messaged because it's free cash flow. It's being messaged in really important differentiation. It's not just, "Hey, let's drive free cash flow because that'll drive the stock price up, and everyone will be happy." That is not a message that resonates with customers, with partners, with our team members. What does resonate is if we drive a company to free cash flow, and we're running this business as a Rule of 40, we could be proud about the business that we're running.
We're gonna have really good free cash flow to invest in the business, to maybe do strategic acquisitions, to position the company in a much stronger way, and that really resonates with all of our constituents and all of our stakeholders. Just to be clear on that's the approach that we've taken in the communication. Our long-term financial model that we've been thinking about is, if you go out three to four years, if you just three to five years, excuse me. If you just take out our ACV growth at low double digits, we're gonna be a business generating $500 million of free cash flow within a few years. I'm not sure that that's, you know, you could just use a very simple number, and this is...
I'm saying these numbers as an example, not as in any way of guidance, right. If you just say, what would you be at 25% free cash flow and $2 billion, that's $500 million, right. It doesn't, you know, it's not, it's not, it's not hard math to see how you could get there. If you took where we are now, and you cascaded out 13%-14%, three to five years, you probably end up confirming that math. That's I don't believe that's the endpoint. 25% free cash flow is respectable, but companies of, you know, that are of our size and the retention of our clients, I think that there's certainly no reason why that needs to be the only thing that we could achieve.
It's a big change from where we've historically been. I think that I don't wanna get ahead of myself there, but this is the kind of business that we wanna run, right? This is what we're committed to: generating hundreds of millions of dollars while not losing a step on innovation and not actually missing any of our engagement with our clients, but focusing on driving efficiency in many parts of the business. What are those parts of the business? There's three fundamental pieces to it. Gross margin, excuse me, sales and marketing, and R&D. Gross margin, 74% is where we are now. I don't like to refer to things that running a business is easy, but scaling Pega gross margin is relatively predictable thing that we can see happening.
For those of you that, you know, when you looked at our gross margin five years ago, it was, you know, 42% gross margin on Pega Cloud, and I said, "We're gonna try to get to 70." Many of you said, "Geez, well, how are you gonna do that?" Then we said, "Not 70, 75," and now we're getting pretty darn close to 75, even in 2023. We think that number is 80, and we see a path to get there. It's not a path that requires a structural change in the business. It just requires, you know, an increased amount of discipline and some engineering time on some of the innovation to be able to help us drive that.
Sales and marketing is a big lift if we did not change the way that our go-to-market model was. If we continue On spreading sales resources in markets that we had very low productivity, it would have been incredibly difficult to get our sales and marketing down to something that would be more representative of what we should be for a company our size, which is kind of more in that 30% range. Right now, 43% is what we were in 2022. You could probably engineer the number in 2023 based on our free cash flow and know that we'll make pretty noticeable progress down from that number. We won't be at 30, but, you know, we could, you know, potentially be, you know, getting, cutting that number in half or close to it.
That's a really big one. That's a lot of dollars. You know, we spent $500+ million in sales and marketing. That's a big number. That's a very important part of us increasing our free cash flow. R&D, that 20 to 17, that's really operating leverage. It's largely operating leverage, and I'll talk a little bit in a second about some of the other specific things. I think the big... Gross margin, I, like I said, is probably the easiest, quite frankly, for us to commit to.
I think R&D is the second, I think sales and marketing is where we need to continue to spend a lot of focus and effort to make sure that we're getting the outcomes that we intend, and also that we're actually spending the right level of sales and marketing and making sure we think about that growth comparison to the spend. Gross margin expansion, three things: scale Pega Cloud, increase the automation around Pega Cloud, both in our control, and implement Kubernetes multitenancy. Kubernetes is completely within our control. Multitenancy is how much can we leverage things that can be run multitenant, even on Pega Infinity. This isn't about Launchpad. This is on Pega Infinity.
When you're, when we're running microservices, and we have each of the compartments of the container of the product that's actually isolated, you can actually leverage some of those as multitenancy. Others, you can't. What are the things that we can do in the architecture of our product to be able to drive multitenancy? I mean, we've made tremendous progress here. Very excited about where we need to get. The second one, and just hitting on a couple of these points, this is our scaling of Pega Cloud. You can see from an operating leverage standpoint, we got a business that's, you know, gonna be approaching $500 million. It was, you know, I think when I started, it was, like, 25... I think we had $25 million or $28 million in ACV for Pega Cloud.
The margin was, like, 30%, right? The gross margin on Pega Cloud was, like, 30%. This has really been a fundamental change in the business. This shows you how Pega Cloud gross margin is converging with our overall gross margin. When I say gross margin, We can put almost interchange Pega Cloud and gross margin because they really are largely the same in terms of the relationship between them. This slide was one that Jacqueline showed. I'm gonna basically just make 2 comments here.
One is, that is to scale of the amount of growth that we would need to get from our existing clients, that's actually over the next three to four years in terms of the yellow, over our current ACV. The gray is the total opportunity set in terms of the market opportunity. We do not need to go out and win. We don't have to cannibalize spend in our area. We don't have to win, displace other vendors. We have to really just win our fair share and expand or cannibalize a little bit in to the overall addressable market to be able to grow the numbers that I had shown. If you look at the one on the right, we have about 200 organizations that are above $1 million in ACV.
We have about 700 clients or so. About 200 of those are over $1 million in ACV. The majority of those over $1 million in ACV are not over $5 million in ACV. You think about the number that we have between $1 million and $5 million, every single one of those clients, I've looked at the list, every single one of those clients could be spending $10 million, $20 million, $30 million, $50 million. They are marquee names, but maybe we're earlier in the journey with them, right? Maybe we just haven't covered the organizations as the way that we wanted to or should have in the past.
This is a tremendous opportunity, and it doesn't require us to build any new product, go into any new markets, and quite frankly, not even target many new logos to be able to achieve our growth. R&D efficiency is really highly focused on Pega Cloud. Why is that? Because clients that are on Pega Cloud are going to be on to, I would say, overwhelmingly consistent versions of the product, meaning they're gonna be on a current version of the product.
When clients are on a current version of the product, the engineering time that you spend on other versions of the product, like, so an example, if you had product version one, two and three and all of your clients were on version three, you don't have to spend any engineering time on version one and two because there are no clients, there are no bugs, there are no security vulnerabilities. Nobody's on version one or two. If you have clients, if one third of your clients are on one third of your clients are on two, and one third of your clients are on three, you have to develop three products. When we were selling not Pega Cloud and clients weren't staying current on their products, meaning upgrading, we had, I don't know, Steve, how many versions?
30 versions of the product in some cases that we're supporting, 50 versions of the product. Can you imagine out of $250 million of R&D, how much R&D you're spending for one customer that might be on one version? Right? What if there's a security vulnerability? Well, we have to go back and figure out. That's all custom work that needs to be done for that one. You have to do it, right? Because we owe the client that, you know, and we also have liability if we don't. I mean, there's a tremendous distraction that happens in R&D when clients are on old versions.
By the way, it's why every single enterprise company has an end-of-life policy and only supports typically two versions back or two years back or whatever their policy is, because they know they'll be spending tons of engineering time. Pega Cloud helps us tremendously on that because nobody's on old versions on Pega Cloud. Literally, nobody is on. We can control the version within reason. I mean, naturally, we might have to tell them we're doing it on one weekend versus another, but, I mean, we have a lot of control over that. The second one is generative AI. I know this is one that's like I've had people in this room ask, like, "Well, how much efficiency will it drive with R&D?" We don't really know. What we know is it will drive efficiency. We don't know.
Some people say 20%, some people say 50%, other companies have said, like, 20%-25%. Sure, pick whatever % you want. It's not going to hurt our efficiency. It's going to help our efficiency. I think. Remember, this is not on the platform and configuring the platform. This is actually engineering about how we actually build and evolve our products. I mean, even having GenAI, I was thinking about this use case this morning, was even having GenAI go out and say, "Look at all of our old versions and tell me where X and Y happens in the code base, that might actually be a security vulnerability or something that might be a high risk." Even that alone, like, that would be tremendously valuable versus the way that companies have to do it now.
Once again, this is going from 20 to 17. This isn't a cut in R&D. R&D will still be higher, right? It's just that from the standpoint of the incremental operating leverage, we see value here. What does this all mean, right? This has been the slide. This is basically a slide that I showed when we first started talking about this, six years ago. If we want to sustain growth and expand margins, which margins mean cash flow, we have to have a business that does not require us to start from 0 every single year, and that's what you are as a perpetual business. That's one component.
Second piece is you have to have a sales machine that is reasonably productive, or the amount of time, the amount of money that you're gonna spend on sales and marketing is just gonna constantly challenge you to understand whether that growth is really worth it because of that cost of sales and marketing. We have to stay focused on consistent and predictable growth. How fast can we grow and still be efficient? We cannot accept a model that trades off, and I actually think Jacqueline said it when she was up here, which is that, once again, it's pretty amazing that she's that people are listening to this and not, and not, and we didn't put words in their mouth, so to speak.
She actually said, "We tried to go after higher growth, we saw it, in many cases, it wasn't worth it." This is, now, this is someone that runs a major region in Europe that saw the disadvantage of actually going after organizations where win rates are low, sales cycles are low, the competitive dynamic might be different, the product-market fit might not be the best. I mean, all these things. That's where the risk is of just trying to put sales and marketing focus on areas that we may not be effective. The counter to that is you go, you thin out. I don't think we're near this risk, you thin out your sales and marketing organization such that you put your exact clients at risk, meaning retention is at risk. You're not covering those organizations.
I do not believe at our in any investment level that I've shown or implied in this slide deck or anything that we've said, we would be anywhere near that in terms of getting to the Rule of 40. I think we can do both. I think we can actually drive an incredibly efficient business that generates, that spits off a tremendous amount of cash flow and still do our clients exactly the justice that we think they should have, covering them and focusing on how we can grow with them. In terms of capital allocation, which this was probably one of those questions that when someone asked, I said, "We're not generating free cash flow, so I'm not really sure I need to worry about capital allocation strategy.
It's, now we're at a point where we're generating free cash flow, and that number's gonna become big over the next couple of years. How are we thinking about it? We viewed the repurchase of bonds. As many of you know, we bought $98 million of face value of the convertible notes. There was a slight arbitrage we played there. Our overnight rates, we get, or call it 4%, and the yield to maturity on those were, like, 6.25%, 6.5%. We decided that that was worth it from a couple standpoint. One, we didn't have a near-term use for that excess cash flow that we felt we needed to preserve it for.
Second, we really felt very strong, the very strong message that we felt we were sending to the bondholders, to the market, about our confidence and our commitment to generating continued free cash flow, and that opportunistically, we would pay down some of the bonds. We did not go out to bondholders and solicit to them. Not that we couldn't if we wanted to, but we didn't. We actually had people come inbound, that had asked if we were interested and had put offers in front of us, and, you know, we felt like it was, you know, it was compelling enough to consider. This is something that we decided to do. I've talked to a few of you about that.
I just think in general, naturally delevering in a situation where you know that the debt is coming due and you may want more flexibility, is something that, you know, we think is valuable at the level that we did it. I think this all kind of ties to just strengthening our overall financial position, which gives us options. I went through that relatively quick because I wanted to make sure. By the way, this slide deck is posted as an 8-K. For those of you that don't already know that, if you wanted to look at the deck, my section is posted as an 8-K, so you can grab that whenever you want. I think it was filed at 4:00 today, 4:00 Eastern Time. I'm going to stop there. Questions. Rishi?
Michael, Rishi from William, obviously. Two questions for you, Ken.
Yeah.
First, if you think about the kind of three to five year model and do the math on that 500, I know it's five million plus, so it's higher than that. But we do the add-on on kind of the ACV growth, and, you know, assume that even at that point, you're at 10% professional services. One good way of saying that is you're still a little shy of the Rule of 40, and then that three to five-year time frame. Are we saying that it still might be beyond that time frame, or are there kind of other things that need to happen to hit that Rule of 40? The second question, I'll ask, which may be a little bit related to that, but so you talked about GenAI and the leverage that it can bring on the R&D line.
Is there anything that can bring on the gross margin line from reducing the services attached and kind of getting stuff more out of the box and get your services mix to go down over time as well?
Okay, great question. The first question, let me clarify. The rule right now, the only calculation we can do on Rule of 40 is with revenue as the denominator and free cash flow as the numerator. Because revenue is not always consistent in terms of how it gets accounted for, I'm basically trying to say it directionally, it'll be in a range. Revenue could be $50 million up or $50 million down, and that could change the free cash flow % by a few points. I was trying to just get to a directional free cash flow number, not trying to walk back Rule of 40, or it's more just the. Because, like, for example, if we have low revenue in a year, we actually might get a point or two of help on Rule of 40.
If we have higher revenue because of the term, we might actually lose a point or two. At the end of the day, it's really about the actual free cash flow, because the billings are not affected by that. It's just accounting. That's the one. I've actually thought about maybe thinking about the free cash flow as a percentage of the ACV number or something else, I just didn't want to start to confuse everybody by a new metric. That was the. The second question was, it's a really good one, which is: What about all the other areas that you. Two things. I'm going to touch on one you didn't ask about, I'm going to hit on the.
Think about how much generative AI could help in fielding customer support calls, triaging problem solving and support, trying to figure out commonality of bugs that could exist or vulnerabilities that could exist across, like, a tremendous amount. Right now, what our global customer support team has to do is when they see something and they're like: "Geez, I don't know what to do," they have to escalate to an engineer. That's another person that gets pulled out of Steve's team that actually has to go work. That would, that's an example of an area. By the way, sales and marketing. I may have told some of you, we have this thing that we're doing, we're testing. It's called Sales Buddy. It's basically a chat.
Essentially, it's a Webex chat channel that you can go in, it's tied to generative AI. You can go in, and you can ask it a question. You can say, "Write me a compelling note to a data scientist to come to PegaWorld from Canada that's worried about budget." It'll basically produce like, here's all the. Imagine that in the selling motion of, like, trying to help a salesperson or someone or a CSM. The question that you specifically asked, our strategy for professional services is slightly different. We want to move as much of our professional services to recurring professional services as possible. We want the work of the actual implementations to be done as much as possible from our partners. Our services resources are, I would view, as the best in the space.
We want them to be focused more on the enterprise architect type role, where they're actually supporting clients and helping advise them on their roadmap, almost as a bundle of when you buy a license, you buy a third of an enterprise architect, a half of a. We want more of our professional services to move there, become recurring. A second part of our professional services is the, is a little bit more of the technical account manager professional services. Someone that really wants someone to be like a, almost a triaging support person on all of the Pega applications. Our strategy there is a little different. Generative AI would definitely help to speed up implementations, and maybe it'll definitely reduce the average implementation, but what it might mean is just more applications can be done.
I think the shift there is a little bit more on the recurring side, so that we can be advisors to our clients and helping our partners, and let our partners really get more of the, of the... Now, that will take some time, but we've already started to move in that direction.
I know Yeah. The 30% number would actually be on the more efficient side of the measure that we showed in the previous investor deck. We just, it's We're focusing more on the % because we're trying to get to the free cash flow number, as opposed to a metric that may not directly tie to the free cash flow. If you do the math on that, if you go back and look, the percentages were slightly more as a sales of the % on the base case. I think it's a little... This one would be in that same range, maybe a little bit better. Yep.
You talked about. The one of the biggest pieces of feedback that we had from our clients consistently was that we were too difficult to contract with. The second one was, you're too expensive. But what, by the way, what vendor wouldn't, what customer wouldn't say that a vendor is expensive, right? But the reality is, why was that? It was because we created this complicated structure of how we would work with our clients. Not the, not the relationship side, but the way we structure the deals. We would say, "Okay, you can have 370 cases, and you can use them on Tuesdays and Thursdays, but not Friday.
If you use them on Sunday, then the case gets divided by..." Like, we had all these ki- and I think that we were trying to get as much value out of that as we could, which is why the pricing comment. When clients would say, "Well, geez, if I go over my number, I don't know what's gonna happen. I just want to make sure I don't go over my number." Their feedback to us was: You're not incenting us to actually adopt more with the product because it's too damn hard to do it, right? That was probably about four years ago that we really started to hear that in a big way.
By the way, a lot of other companies were kind of moving in the direction of moving away from user models and moving more to consumption models. That's when we started this in earnest probably about four years ago. Now if you look, like, most of our new agreements have some level of. They're case-based, and they would, as much as possible, have some level of like a flexibility on consumption or at least trying to consider that. We're trying to go back, and we call it modernize the contracts. We're trying to go back to existing contracts. Clients are very supportive of it.
They're not like, "What are you trying to do here?" They actually like the flexibility because the other flexibility they get is if they end up coming down by 10%, they pay 10% less, but they also lose some discounts as they go down. There's like a... Our view is we know the use cases. They're not going to actually go down in most cases, so we actually feel like that's hedged. Yep.
I was thinking more, just I was thinking with the GenAI idea, could that actually might be a solution for the adoption, or are you changing the ideas of, you know, how you were looking at it between on-premise and cloud? Could that be a thing for you also?
I'll start with one question, you might wanna follow up. The product is the same, with the exception of the managed service and the specific technology and engineering around the control plane for the cloud. The product of a term license or Pega Cloud, there isn't a difference in the product. The reason why GenAI is available on Pega Cloud first is because we can control and manage the customer experience and value more than just saying, "Here you go, you can do whatever you want." Actually, it will become a motivation for clients to move to Pega Cloud. Look, client cloud choice is not something that we would choose to do from a business standpoint.
It's more because clients want us to do it, right? Like, they're just like, they're saying like, "Hey, you know, I don't want to buy." We have to like, we have to be really careful with that because we wanna make sure we're giving them the right way to get value. We don't have to do things that would actually further encourage them to actually not actually leverage Pega Cloud. Because we think we know that when our clients are on Pega Cloud, they're on a current version, they get way better experience, they're getting more value, they have NPS scores, lower GCS tickets, lower drain on engineering, and they actually like the application more. I think we know that. We just wanna make sure that we're not...
We're getting them there with a carrot, not a stick, so to speak. That's the sensitivity there. Now, I don't know if there's a follow-up that you wanted to ask.
Yeah, I got one.
Yeah.
The component goes on this, where you're talking about multitenancy, where are we in that? Is there a long way to go on the multitenancy side of things?
We're in the early innings. Yeah, we've done things like we can multitenant like things like search. There are things that we can multitenant within an organization, so we're a little bit further on within a client, but because remember, clients have multiple applications, so even multi-tenancy within a client is very helpful. We're looking at the next level, which is, can you multitenant certain parts of the product that you could share within, say, like the United States versus Europe? That we're very early stages on that. Launchpad is actually one of the things that's driving a lot of our insight into that, because Launchpad is fully native multi-tenancy. I think, Steve. Sorry.
Hi, there. Just, what are the current on GenAI and the cost side? On GenAI, back in the growth targets. As you offer this to customers increasing this year, 2023 rolls out, what's the potential impact to you on the cost side of those deliveries? The second is.
Yeah.
-on Launchpad, how much do you think of how big is the effort on the sales and marketing side of this, is this, you know, the same percentage of normal cost to that?
The variable cost for Pega Cloud really is driven on processing. Storage is a very small component of it. GenAI does not have a tremendous amount of processing. It has a call, it pulls back data, it might store data, it's not a tremendous drag on the AWS cost to actually run GenAI because the actual, the calculation model is not heavy. If I look at the total cost of an AWS for a client, it's 92% processing or something related, and 8% storage. It'll be some small amount, but it'll push more volume, I think it'll be probably net accretive to margins. Your second question was?
Remind me? Sorry.
Launchpad.
Launchpad. Well, sales and marketing. We didn't answer this question, but I think somebody asked it, maybe Steve asked it earlier, and I meant to interject. Launchpad, think of it as a channel model. It's not a channel model, but think of it as a channel model. You do not have a tremendous amount of direct selling. You're trying to get the channel partner enabled, and they will sell. The sales and marketing right now is, I mean, less than seven people. It's tiny. I don't think that's gonna... Like, if Launchpad was, like, massively bigger, I don't think it would maybe it might be 50. It's just not gonna be a noticeable. Marketing, we haven't done much marketing.
I mean, I should say, we've not done any marketing on Launchpad. We've actually consciously tried not to market Launchpad. I think we probably need to spend a little bit of marketing dollars once we get some momentum there. No, it's very little on the BD side. Steve?
Hi, Steven Lynch with Citi again. I'm gonna ask on the growth assumptions that you're assuming here going forward for the 13% and 14% growth. I think as you mentioned earlier, you've seen 100% retention down to, you know, 115. I mean, how should we think about where that hits timing going forward in the assumed year, maybe also in the case of GenAI and potentially accelerating adoption of consumption? How do we kind of move all those pieces together?
Launchpad's not in any of my assumptions. Pega Launchpad is on top of anything that we show here. You know, realistically, it could become a mitigator if we have any, but it's not, it's not a, it's not something that we've modeled here. I suspect that Pega Launchpad will require a more in-depth thought about how we actually go to market and what our game and how fast do we wanna scale it. I think right now we're really more focused on gauging adoption and understanding the subscriber model and understanding the economics. And let me tell you, the early read is, like, amazing, but I think we need to not jump in before we actually know a little bit more about the economic model. That doesn't mean I'm not extremely bullish.
It's just, you know. In terms of the net retention rate, in our model, we've taken down our net retention rate slightly from historically what it was because we don't have a lot of growth from new logos. That's more of a function of our view of the economic environment over the next couple of years versus what it was over the last few years.
Just macro?
Yes. It's just us looking at the macro environment right now and saying, "What do we think is the..." just assuming that the macro is not gonna get noticeably worse, but is not gonna get noticeably better either. It's gonna kinda, you know, maybe anchor a little bit, maybe be a little bit more consistent over the next couple of years. We don't assume, like, some big bounce back. We also don't assume that there's, you know, a two-year recession either.
Yeah, this is Will Jefferson from EA80. You gave us the piece I needed to get to just as operating income. In the last few down years, guys, do you expect there to be any structural benefit from Pega Cloud transition to working capital intensity, and whether that's expressing receivables, payables, as we go around?
Yeah, I mean, that's a really, that's actually a more advanced question, I think. Yeah, I think there should be some. If Pega Cloud becomes our entire business, you should get a little bit of operating cash flow accretion because of the deferred revenue being higher and actually starting to be outsized over the AR number, versus right now, where we have a 50/50 model. If we were a, say, a 75/25 model, you'd have 20, you know, that 25% shift. That's not something. We've modeled it, but it's not something that's like, it's not a massive number. It is slightly helpful. Yeah. Any other questions? Listen, I know we're well over our time.
It's quarter till 3:00, but, hopefully, this was helpful, and, I think, the turnout was great, and your questions, clearly, you guys are very engaged, so thank you so much. It's the innovate, the innovation center is open until, I think, 6:00, in terms of the, in terms of the. I think, oh, sorry. Peter's giving me the. The innovation hub is open until 6:00 today, and it's still 6:30 tomorrow. Use the Pega app. It's actually pretty good. Talks about all the scheduled events.
I would definitely, I would highly suggest anybody that's here tomorrow, come to the morning session and listen to the AI discussion, because we have some of the, I would say, some of the most advanced AI thinkers in the industry, and I think it'd be. Specifically, you know, Rob Walker has been with us for a while, and he's been through many generations of AI. You know, when AI was like, you know, everybody laughed at it, and then it was real, but not real, and now what it is now. I think it'd be really interesting for you since there's so much interest in it. Thank you, everyone. Appreciate it, and I'll be around for the next couple of days, so.