Hello, everyone, and welcome to our Investor Day at Interactions 2025. My name is Marty Cohen. I'm Head of Investor Relations at NiCE, and I know most of you. I hope you enjoyed the general session. What I want to do is just take you briefly through the agenda for the rest of the day, and then we can begin. For this session, we're going to hear presentations from different members of our senior management team, as well as one of our customers. Then we'll end with a senior management panel where you'll have the chance to ask questions. We'll hear from Scott Russell, our CEO; Barry Cooper, who's President of our CX; Elisha Wright, he's Global Director, Learning Design and Delivery for Hyatt Hotels; and Beth Gaspich, our CFO. Scott's going to provide a general overview on our strategy.
Barry's going to discuss in more detail our CX and AI strategy. Elisha will talk about how he's using NiCE AI to improve the customer experience, and Beth will discuss our financials. Please hold all your questions until the Q&A session. For those of you that are listening on the web, there is a space to ask questions. Again, we'll take those questions during the Q&A session. After the presentation, we'll take a quick break. We're going to grab lunch. The lunch will be in the back of the room. We'll come back, we'll sit down, and then we'll begin our Q&A session with Scott, Beth, and Barry. After lunch, we've arranged for you a private tour of the Innovation Hall, and you'll see several demos of our new AI solutions. I think you'll enjoy that. That'll basically end our day.
For those of you who are staying tonight, we do invite you to our customer party, what we call our customer appreciation party. That begins at 6:30 P.M. It is at the Virgin Hotels, and the band OneRepublic will be playing. Buses leave this hotel at 6:15. Let me put up the famous Safe Harbor slide here. Finally, all numbers in the presentation are non-GAAP. I will now invite Scott to the front of the room.
I thought you were going to talk through the Safe Harbor slide as well.
No.
Thank you. Good morning, everybody. I appreciate that you all joined us here today. I know many of you are not able to join for the full two days, but the opportunity to share directly with each of you a little bit more detail, obviously, about where we're at, where we're going as a company, and also an open dialogue in the Q&A session so we can answer your questions and make sure that we provide as much information as we can to help you not only in your analysis, but also, hopefully, a positive view about our company. I want to just highlight two things before I go into a bit of detail. The first is, and you will have noticed, today we're really drilling into CX, and it's in line with the event. It's obviously the major part of our business. Okay. Okay?
Can't hear me. Oh, you didn't mic you up. Oh, you didn't mic me up. Yes. I rushed here, so I'm non-mic'd. I can do this. I can juggle and chew gum at the same time. We obviously focused on we're focusing on CX. The presentations today from Barry, myself, Beth's obviously more at a company level, but even then it's zeroed in on the CX side. I wanted to highlight that. The second, and it's probably even more important, is we realize, and one of the things that I'd spoken about in earnings and other forums is the need to provide more disclosure, more transparency to help you.
Whilst we're really excited about what we're going to present today, as you can probably appreciate, being six months in where we've got a lot that we've done, but with a lot that we want to do in sort of building out those midterm and those longer-range targets, we're going to come back to you with a capital market day, another investor day in October. I can hear a bit of can I hear myself?
Yeah.
Woo. That was awesome. We'll come back to you with more details. Beth and Marty and the team will come back to you on a capital markets day in October, early mid-October. We'll come back to you with the exact dates. The intention at that point is to give you even more detail than what we're sharing today, and we're going to give you a fair bit, but the opportunity to be able to combine the innovation, the roadmap, the expectations of not only what you'll see this year, but for the midterm as well. Today I'm going to focus on three things. I'll be covering three pillars of our strategy. You obviously heard me on stage this morning talking more at a high level about what we're doing as a company and creating a NiCE world.
By the way, I have said this to my own team, but I will say it here as well. That is not a marketing slogan only. Clearly, we're using our brand and our name, but the opportunity to create incredible human experiences in the world of CX, in the world of self-service, in the world of customer experiences, it is a differentiator. I believe that if we deliver seamlessly and consistently and purposely against that vision, it will be another reason why NiCE is the winner. I'll be covering our market, what we're seeing, and why we have confidence because of the market dynamics. I'll then share at a high level our innovation, but clearly, I'm going to wait for Barry to come on, and he'll talk in much more depth about the innovation, the capabilities that we have.
I'll also spend a bit of time on our go-to-market, and I'll conclude with a wrap-up of that as well. Let me go to what's happening in the market and what do we see. This is really important because the changes in the market are trends that give us confidence on not just our short-term, but our mid and long-term growth potential. The first, and we've been talking about this as a company, and the industry has talked about this for some time, the move from CCaaS, and we've been talking about on-prem to cloud. What I'm here to tell you is companies, and you heard a few of them today, are moving from CCaaS, and they're moving to AI-powered platforms to drive customer experience. They're moving from human interactions to AI-powered self-service, and I'll talk in some more detail.
They're going from scripted bots that are largely deterministic, so very tight boundaries of what they can do in retrieval of information, to true AI agents. Moving from orchestrating interactions, and we've been great at this over generations, orchestrating an interaction between a brand and the consumer, but moving that to automating workflows. By virtue of that, automating workflows that go beyond the interaction, go into the mid-office, into the back office, and you heard me say that a few times this morning. We are definitely moving from an agent-centered world to an interaction and engagement-centered world. Moving from the 15 million agents that everybody talks about in this industry and going to billions and billions of interactions, and you will see it is growing dramatically.
Last but not least is moving from a world where there's been a high emphasis around labor spend and about how to manage the contact center into technology spend and why companies like ours are going to be a key pivot point of the way that customer experience is being delivered. Let me drill in a little bit more detail to give you some context of why we see these market changes and what's happening. As I mentioned before, we are moving from a CCaaS on-prem to cloud. Remember, I do say this knowing, of course, that 35%-40% of on-prem has moved to cloud. There is still the move. We are still seeing an abundance of RFPs, an abundance of customers that are sitting on those old legacy systems. You heard Carnival UK today. Walmart was the same. Disney was the same.
They moved from the on-prem, and they moved to the cloud. What we're now seeing is they're making the jump. They're not going on-prem cloud, cloud AI. They're going from wherever they're starting. They could be at an on-prem. They could be a cloud-enabled platform that we have, but they're going to an AI platform that drives their interactions and that drives their experiences. Why this is important is we have an opportunity, as a company, for example, to monetize the agent and the work that they do. From an interaction standpoint, as you're growing, every time they use a Copilot or an Autopilot or for a supervisor for agent to be able to do an AI agent, we're able to do it in a cost-effective way. I'll come to this later.
Not only is the opportunity for us to be able to grow and earn, but for our customers, they're able to leverage the AI-powered platform to be able to serve their customers more effectively, not limited by a siloed tech infrastructure. If you look at this, this is a depiction of our AI traffic. You can see it. In the last 12 months, in fact, I would argue in the last three or four months, the growth has, it's going way more than what it was a year ago. That will continue. That trend line continues to explode because companies start with a small interaction. You heard the experiences today. They start with one, and then they move to another, and then they move to another, and they're able to then get the benefit of that AI-powered platform.
The beauty is it's all on the same platform, the same data, the same models. Whether they do it for a Copilot or an Autopilot or they're able to do it for an AI agent, it's the same platform that they're leveraging from, which means it's consistent. The second is that we are definitely moving from human interactions that are agent-based to rapidly expanding AI-powered self-service. Self-service is clearly here. I guess there's a couple of points that I would like to make to this. While at the moment there is a lot of self-service and a lot of companies, and you will hear about those companies, there's a lot of startups out there that talk about their self-service capability, their AI bots, and they're really good. The reality is they're only delivering 14%. 14% of the service issues can be delivered by self-service.
Even though there is a self-service capability, the potential, so nearly all customers are using some sort of self-service, but full fulfillment, full resolution, full task is only done at 14%. We call it the self-service resolution gap. What we see is the opportunity with AI to be able to bridge that gap, more and more complex use cases. I really loved Anderson's presentation today. You think about it for a second. They've got an AI agent that used to handle a delivery issue for pharmaceutical, and then they had another agent that handled delivery of groceries, and they had another agent around calling for support for tech issues. The role of a human agent now can handle it all, but it's done with the power of AI, delivering self-serving a lot of that, which means the human agent then does only the complex tasks.
The opportunity for us and the opportunity in the market is to be able to help companies increase the amount of self-service through our platform, but still coexist in a seamless conversation with our customers. The third is we are definitely seeing the move. We launched Mpower Agents today. Barry will go into it in a level of detail, but moving from scripted bots that are very deterministic, they've got tight guardrails, to true AI agents. I'll put it in very simple terms. A bot will have certain things that it can perform. It will be pre-programmed, determined. Yes, you have that interaction. By the way, my experience with the finance, that's real. I literally logged in, and I was visualizing the experience, and it was a NiCE customer.
It was a great experience with a bot, but there were certain tasks that it was not scripted to do. The reality now is with AI agents, not only can you handle the contact and the interaction, it performs tasks. That is the most important part. Those tasks are not limited to the interaction. It is fulfilling tasks that a human agent might have done, but it also is fulfilling tasks that the mid and the back office could be done. You might log in and say you want a new credit card that needs to be available within 24 hours. Historically, that would have been a bot that flipped to a human agent that would have had to check with credit whether they can issue that card and whether they're allowed to go to the limit, completely automated with an AI agent.
We can give you countless examples of that. The fourth is orchestrating interactions, and this is obviously a really important one. It is tied to what I just said, orchestrating interactions into automating workflows. It is no surprise that you will see when I announce the strategic partnerships, they have a clear purpose. This is not just partnerships about go-to-market. These are partnerships that help us achieve end-to-end fulfillment via our platform. Whether it is ServiceNow, because they are building AI agents themselves that very much cover the mid and the back office, what humans do in different tasks, automating workflows, or whether it is AWS and leveraging their underlying platform and their data and their Bedrock and Q Business, or whether it is Snowflake and they are able to federate data real-time and data sharing, the ability for us to be able to perform tasks and automate workflows that goes beyond the interaction.
This is really important because I believe we as an industry have been limited to what happens between consumer and agent, the interaction point. Once it goes beyond, we did not really participate. AI gives us the ability to participate and engage and deliver value across the organization. Not exclusively. It will be through partnerships as well as our own, but our workflow orchestration definitely has the ability to do that. You can see that we are going to blur the lines. The tasks, and I guess this is even more important, is not only are we doing what the agent does, but you would be shocked at how much, and you will hear the example later on with our customer presentation with Hyatt, is one interaction with a human agent often has one, two, five, ten.
There's a lot of people that are working on behalf of that intent, that interaction. Our platform, the intent is to be able to blur those boundaries and be able to go from intent to fulfillment using CXone Mpower. Second last one is around the agents versus the interactions. You can see here in simple terms that our growth of digital interactions is exploding. I need to highlight voice calls are not reducing. There's no material change in voice. In fact, I'm amazed. My kids tell me this, but Gen Zers are just as likely to make a voice call as what people of my generation would be. Voice is still there. The digital interactions to be able to interact with their brand of choice is growing exponentially. It's not just the chat.
The chat is the first point, but then it's the chat on any platform, and then the interoperability between those. Think of the proactive side versus the reactive side. You can't do this on separate platforms. Siloed solutions and being able to handle different interactions. Again, you heard the examples. You'll hear more too if you've got one experience that handles it, does a bot or an interaction digitally one way, and then a different experience on another, and then a different on another, and then you've got voice on NiCE. Honestly, there is no way you can interoperate seamlessly and have a great consumer experience or make it easy for a human agent when that occurs. A unified platform truly matters.
Last but not least, and I know this is probably important to you when you think about our addressable market and our modeling, the way we view this is very clear. The market will definitely shift in the contact center, in customer experience, where labor spend ultimately will come down. We have not seen any dramatic reduction or any material reduction on the human agents at this point. A lot of businesses are still using human agents. They have become more productive, but they are doing more and more, and they are doing other revenue-generating things. Whether they decide to reduce it or get more efficiency, the reality is the technology spend is definitely increasing. It gets bigger.
It gets wider, which means beyond 2025, our addressable market, as businesses move from where they are to where they're going to go through AI, the increase of technology spend will absolutely increase, and it's a massive TAM opportunity for us. The market is a dynamic one. Yes, there are competitors there, but it is a market where we are well-positioned to win. I just want to touch on in a little bit more detail and context the three pillars that I presented this morning when I talked about reimagining customer experience, being the platform. It is so critical, and it was work that was done before I got here. The re-architecting of CXone into CXone Mpower is a critical foundation because you're able to handle automation of workflows. You're able to handle human and AI agents and any sort of bot.
You are able to consolidate and then leverage and learn from the knowledge all in one place, no matter what deployment that you choose. First of all, on workflows. Workflows, let's just remind ourselves what they are. They're a series of tasks, underlying process, a series of tasks in order to complete from an intent to customer calls or texts or chats or whatever, and they have an intent. Often, it's not just one. It's many. Many things, per the example that Walmart gave. There might be four or five different intents in the same interaction. The workflows is the ability to be able to orchestrate that of one interaction, but the billions of interactions, and they're able to do it seamlessly, that cuts across the interaction into the mid and back office. For example, in the—sorry, let me go back one. Where's the workflow slide?
Okay. I'll just talk to it. Oh, there it is. Okay. Apologies for that. This example is a good one. It's a financial services provider. They had three million appointments that they received each year, three million appointments. 84% of those appointments were scheduled via a call, scheduled via a call. A high number of the tasks, scheduling appointment with their financial services analyst or provider, were handled by their contact center. What they did with us is they replaced their workflow with our self-service. They were able to connect not only from our self-service that handled the interaction, but we connected automatically to the appointment system. We took the human out of the loop, and we're able to then fulfill that with a containment rate of nearly 70%.
You think about it, the volume of those voice calls, we're able to get the fulfillment of a very simple but very important task for that organization. It's a simple example, but you multiply that across every industry, it's an enormous opportunity for NiCE to take a broader step in what actions get taken in the back office that used to be the domain or the world of either the CRMs or the ERPs or the hyperscalers or other platforms. Let me be clear. We're an unavoidable contact point. They will come to us first. Please. We're not the human. The contact will come to us first. By coming to us first, if we've got the technology platform to fulfill, we have the right to solve it without it ever needing to go to any other technology platform out there.
When the customer comes to their brand and they leverage CXone Mpower, we're an unavoidable first point of contact, the single pane of glass, and we get the ability to expand our TAM seamlessly, and it's easier for companies because they can do it all at that point of interaction, resolving it real-time. Why go handing off to other enterprise systems? The second is agents, and clearly you heard today the launch of Mpower Agent. It's exciting. I hope when you go to the innovation hall, you'll see this and you'll see it live. It's really important because it does move from the scripted bots to AI agents, and it truly is as simple as I described it. Barry will show it tomorrow on stage as well, but we can show it.
Literally, you'll do an English prompt or a command prompt, I guess, in any language in the future, Barry, but you're able, and it will create the code, create the AI agent automatically, ready to be deployed. The thing is, if you think about it, AI agents are not going to be only in the domain of when it's going to be an auto when it's going to be a self-service option. Think of a human agent that's in contact, and they want to get a task done. They want to update that appointment. They trigger the AI agent that updates the schedule, updates the appointment, and tells the customer real-time. Yes, it can be in a full self-service scenario, but quite often, it will interoperate with a human agent and an AI agent together. Why is that important?
Because if you're not on the same platform, you can't do it. Everybody asks me this question, "Oh, other platforms have also got AI agents." Yeah, they do. But they don't have the context of the customer engagement, and they can't do it without knowing that. You have to have that interoperation, the guardrails, the experience, the AI models that are explicit for CX. Barry will share that in more detail. Last but not least, and I'll keep moving, is knowledge. I was interested in what John Wells said this morning at Carnival UK when he said that knowledge was the, I think it was the last bullet point that he presented, but it is the foundation. It connects the dots.
The single platform, including all data, which right now for most businesses is spread across a series of different bases, but we unify it onto CXone, onto the knowledge base, including our AI models, our CX-specific ones, as well as the foundational models. You can choose your LLM, and you're able to transform that knowledge into actions, into outcomes. You can transform it. The opportunity, obviously, is to be able to go with the data and knowledge and the AI workflows is to be able to give more contextual insights for agents, to be able to self-serve as models, but the guardrails do matter. You'll hear that a lot with customers. The guardrails that they put in place about what can be delivered is really important. They're not going to give a premium account upgrade to every customer.
There are barriers and there are limits, and we've already got that built in. Okay. Let me move forward to our go-to-market. There is a number of levers, and I won't go through all of them, but needless to say, we do have the industry's largest customer base. We do have a vast ecosystem. We have strategic partnerships that we've started, and we're not done. There is more to come. We have an enormous opportunity of growth through international expansion. Honestly, we've grown really well here in North America, but the growth opportunity internationally in Europe and in Asia is a tremendous opportunity. Obviously, I've got a lot of experience in that, and we've got a well-diversified business across all verticals. I do want to highlight that our partnerships are not only the ones that I presented on stage.
We have an enormous, we added 110 new partners last year. This year, we're on track to do something similar. And 75%, 75% of the CXone Mpower new logos are delivered through partners. Yes, we've got an amazing ecosystem, and yes, we've got a direct interaction with our customers, but we are very focused on leveraging the power that the ecosystem is to be able to extend our breadth and our depth and our reach, especially in international markets, but not exclusively. The other part is that all of our large enterprise deals or two-thirds of our large enterprise deals are also partner-led. We have a great customer market. Our customers are an enormous opportunity for us in financial terms to upsell and cross-sell.
Many of those customers, they're at varying stages of their journey from their on-premise siloed systems to an orchestrated single platform that they're able to deliver and leverage the AI benefits that we talk about. No matter where they are, the point is every one of them are on an AI journey, and the opportunity for NiCE is to drive that AI journey using our platform and using it consistently, not only to automate what they currently do, but to transform the experiences with their consumers. I guess this is the best way I can describe how it manifests itself. If you want to have a depiction of how a journey of an existing customer goes on the journey with NiCE, this is a good one. This is a global entertainment company.
In 2022, they started with base platform of ours, non-AI, but they started with CXone Mpower, agent experience, OCR, recording, and they implemented the CXone platform. In the last three years, they've gone from an ARR of $3 million to $10 million. 40% of the ARR is now AI and self-service. You know that we represented and we're sharing that stat. Why is it important? Because not only new customers that go straight to AI, but existing customers who are leveraging our platform are increasing disproportionately the amount of services of AI services rather than historical platform services. That gives us growth. The thing I love about this is economically for them, it was a no-brainer. It was a no-brainer because they were able to automate what was human tasks. They were able to get more efficient and reduce redundancy in the workflows of their business.
Obviously, they were able to deliver delightful experiences for their customers while containing cost. Our view is very simple. We have an amazing market of which we're operating in. We've got a great innovation platform that gives us the potential and the opportunity to grow and succeed. We lastly have got a go-to-market and a customer opportunity that gives us the right to win. We have got a proven track record, but we are impatient. We are impatient. If we sit and wait, this market will be taken up by someone. There is no doubt about it. I actually thrive on the thought of it's a competitive environment. Yes, there are others coming in here because they see what we see: a huge market opportunity, an increasing total addressable market, but what they do not have is the domain expertise and knowledge in customer experience.
To share with you why that is different and why that is important, I'm going to hand over to Barry. He's going to talk about our innovation.
I got this. Going to pick.
Cool.
Hey, guys. So I asked for three hours to share our innovation with you. I got 20 minutes, probably now 10.
Sorry about that.
That's all right. Those of you, seriously, those of you around tomorrow, we're doing a main stage as always. It's like 50 minutes going really deep in a lot of the new innovation. If you can stay for tomorrow, guys, we're going to show some incredible, incredible stuff. Anyway, a couple of foundational slides, and I'll just go a little bit deeper into some of the things that Scott mentioned. Look, I'm going to not repeat everything here. It's clear that AI is real. It's not hype. I think everyone agrees with that. It's not only allowing us to automate customer service and augment customer service. We're actually redefining it. We literally are tearing up the rulebook and starting all over again. That point that Scott just made about the compression of the middle and front office together and back office, this is real.
I'll talk through some examples of that that we have. Those words we show here, they are chosen deliberately. Yeah? Hopefully, you know that the orchestrating workflows, agents, and knowledge by workflows, not just interactions, not just calls and chats, end-to-end intent fulfillment by agents, not just human agents in the front office, human agents and AI agents, and increasingly middle and back office humans and AI agents as well. Finally, knowledge, John said it perfectly. It's actually our terminology, but it's right. Knowledge management was something that was the depth not really managed by organizations. It wasn't a priority. Along came GenAI and basically put some magnifying glass on that knowledge. Suddenly, it's very important to make sure your knowledge is in order and structured and correct because suddenly, it's available to everyone. Knowledge management is AI management.
One last thing, that last line there as well. Please do not underestimate that. There may be many competitors at the low end of the market. There are very, very few at the high end of the market because it is so complex. One thing, again, I think is really important. All three customers that spoke on stage, including Hyatt's, who will speak to us here as well, they are all multi-brand organizations. Yeah? If it is Disney, and they are talking about Disney, ESPN, Hulu. If it is Walmart, it is Walmart, the online Walmart, the stores, it is Spark, you name it. If it is Carnival Cruises, it is P&O, it is Cunard, it is all of those different brands. All of those customers are our largest customers that actually put all of their brands onto one instance of CXone. That is huge. Only very few providers can do that. I have included this slide here.
This is from our sales deck. I think it's really important because it really communicates our two value propositions. This is the traditional one everyone's familiar with here. Scott mentioned it already, but we are the single pane of glass between consumers and organizations. What does that do? It solves problems for three stakeholders. For consumers, we've all been consumers. I had an experience on an IVR or whatever, and then that experience is not carried over to a bot and vice versa. Your consumers, the silos are broken down. Everything's in one place. For an employee of an organization, they're not alt-tabbing between five different systems trying to find the history of what you did. It's all in one place. Really important for the organizations, they're no longer SIs. We've pre-integrated everything for them so they don't have to integrate it themselves as well.
This is our traditional value proposition. It's a nice little picture. Really, our value proposition now goes further than that in that we are hiding the complexity of what it takes to deliver fulfillment through workflows that reach into the middle and back office. There are two really important things about that. In almost any organization that goes beyond a basic level of sophistication, there are multiple systems of record and multiple systems of workflow. When we hear about Salesforce or ServiceNow with their bot that goes to the front end, they're only seeing their part of the business. We see everything because we're the single pane of glass. That orchestration is managing multiple steps of fulfillment into multiple systems of record, systems of workflow.
The other thing, and this has been something we're doing for three or four years now, and I hope you guys are aware of this. This hub strategy is, in my opinion, genius. What hub does basically is we allow our customers to use their own technology or use ours. It's up to them. Because it may be they have a strategic need to keep knowledge in a certain place and not migrate it to CXone, or it may be a large customer can't migrate all in one go and needs to do it over three or four years. The hubs allow our customers to keep those technologies while adopting the value of the platform. Knowledge can be in CXone, or it can be in Salesforce, SharePoint, Agent Assist. You can use our Copilot or use Agentforce or 10 other kind of Agent Assist services.
AI services, you can use our LLMs, our ASR, or bring your own LLMs and own ASRs. Our virtual agents, same thing applies. This hub strategy is absolutely key to making us very viable for the high end. What really changed over the last six months, and this is really important since I last spoke to you guys and some of you I met at Enterprise Connect, these are the three things that really changed. When we first came to market with our AI solutions, our Copilot, our Autopilot, our Actions, Autosummary, Exo, each of those, our focus was getting them to market as fast as possible. They came with their own AI services. They came with almost duplication of functionality amongst those things.
What we did in release 24.4 at the end of last year is we removed all of those AI services from those individual applications, and we put them inside the platform. The platform now manages LLM selection. It manages the prompt editor. It manages LLM selection. It manages our ASR selection. It manages our knowledge management. It manages our RAG indexing all in a single place. That was really powerful because it means all of our applications now are getting their AI services from the same place in a consistent way. That, by the way, was the prompt, pardon the pun, to change CXone to CXone Mpower. CXone is the CCaaS brand. Empower is the AI brand. Together, that's CXone Mpower. First kind of major change that happened over the last six months.
In the middle, the second major change, and this was driven by a lot of our customers, including the likes of Hyatt, who will speak here later, increasingly our customers are asking us to move and work with a different kind of user, a different kind of function as we do what Scott's talking about around fulfillment. At H&R Block, for example, we're working with tax pros. At Walmart, we're dealing with fraud specialists. At CSAA, we're dealing with adjusters as part of the insurance process. At UHG, we're dealing with pharmacists. At some of our financial customers, we're dealing with mortgage specialists. These are functions and people that aren't traditional front office people. Our customers are saying, "We want you to support this function and automate this function because it's part of our fulfillment of medium to complex intents." This is huge.
You will see new capabilities like CXone Desk and other things we are doing around fulfillment. It is obviously a big part of what we are doing with the Mpower Agents, supporting that new kind of user, that new kind of function. Because when you align those middle and back office functions to customer service, you achieve the holy grail of amazing customer service and reduced costs. The third thing along there related to that is you talked to us about two or three years ago, a lot of our automation, we were automating the dissemination of knowledge. We were automating basic routing, getting to an agent. Increasingly, our customers are asking us to automate really complex intents and multiple intents that reach into those systems of workflow and systems of record. You are familiar, hopefully, with this. This is how we picture the CXone platform.
Really, three things to take away from this. It's a real platform. It's eight years, millions of lines of code, incorporating data visualization, data store, UX, as I mentioned before, kind of the AI services, cloud security, user security, all those things in a single place. Every time we build a new application, it inherits all the capabilities down there on that platform. We're basically done with about 40% of the application before we write the first line of code. Here is our core value propositions. We augment the workforce, making them incredible at their jobs. We orchestrate workflows, not just interactions, and we automate service, not just knowledge dissemination. They're our core value propositions. I'll talk about this in a second. CXone Mpower is a platform that allows organizations not just to operate, and that's really important. Operations is really hard.
It is why the likes of Salesforce and ServiceNow have got a surprise when they come into this market because operations operating a mission-critical system is very different to operating a database with a front end. We allow organizations to design, build, and operate through the capabilities that we offer. As I mentioned, a core part of the platform is our AI. I think you're constantly asking people at NiCE, "So what's different about NiCE's approach to AI compared to everyone else?" It is extremely different. I can't tell you how different we are to the likes of Genesys, Five9, and even Salesforce and others because what we're doing, we're combining two technologies together in that platform. We are using large language models and large action models.
By the way, we're swapping and changing models all the time based on the usual criteria of the accuracy, the cost, and the speed, depending on the use case. We're combining that with our Enlighten models, our CX-specific models that we developed starting back in 2019. Every use case, almost every use case we do around AI, those two things come together. I'll give you a very, very simple example. Okay? AutoSummary. Everyone's got an AutoSummary. Most organizations, they just take a transcript of a call or a chat. They send it to an LLM, and they write a prompt to say, "Summarize what went on in this." That's great. You can do that. You all know by now the way LLMs work, they're not consistent. Everyone's going to be a little bit different.
It's just information theory condensing down a whole bunch of words into a summary. What we do, our approach is not that. Our approach is we first run our AI models, our CX-specific models on that transcript. We say, "What were the intents? What were the actions that took place? What were the outcomes that happened during this interaction?" Those are then the metadata we send to the LLM. Again, it's an example of creating the guardrails. What we get at the end of the day is an autosummary that's referencing those intents, actions, and outcomes in a consistent way using the industry terms. That's one example, a very simple example. I can give you 20 other examples of how we combine those two things together. The result, far greater accuracy, first of all. It's CX-specific outcomes, not generic outcomes.
Something you guys may find quite interesting, it's a lot less cost for us because we're actually not sending a full transcript to the LLM. We're just sending the metadata, which is a lot less. I guess the key point of all of this is when you have AI embedded in the architecture, embedded in the platform, unlike point solutions, you're using AI consistently. It appears in the same way in all of the applications. It means the customer goes to one place for their AI governance, not 10 different places. It means that all of the information captured by these capabilities is feeding the AI data so it gets better. It's all fed together. It's all connected. Bringing me to the announcement we made this morning, our Mpower Agents.
You're going to go down to the showcase, and you're going to go see these in action. It's incredible technology. It's release one of the Mpower Agents. I'll tell you that now, but it's a release. It's out there today. Mpower Agents, the way to think about these is if what we have with Autopilot and Copilot, that's conversational AI. It's focused on the conversations with humans, employees, or customers. Mpower Agents are process AI. They're actually doing the fulfillment that all of our customers are asking us to do, working with adjusters, working with mortgage specialists, working with tax pros, those kinds of things. They sit behind our Autopilot and Copilot and actually execute those tasks. You'll see the way these are created, and they are truly created in an agentic way. I'm choosing my words very carefully here.
I'm going to say they are created in an agentic way. You don't build a flow like you traditionally build an AI bot. What you do is you give it a job description. You give it a prompt. You say, "Your job is to go solve, go and find," as I said today, as Scott said today, "go find customers in the database who are eligible for an upgrade based on this criteria. If they're available, go and then reach out to them in their preferred method of outbound communication and then offer them an upgrade. If approved by them, then do the upgrade." You literally write that as text. Then the Mpower Agent builder goes away and builds that flow. It's going to use all of the capabilities available on the platform to do that.
It's going to use the APIs that are set up on the platform into the system to record system of workflow. It's going to use all of the channels that are available on the platform. It's going to use the preference data about the consumers that's on the platform. It's going to use the integration, going to use the knowledge and the knowledge bases on the platform. Those resources, they form the guardrails to make sure that it stays on track even though you're building it in an agentic way. These Mpower Agents become a big part of the design, build, and operate. With Empower actions and some of the capabilities on there, and I'll show you just in a second, we allow customers to uncover opportunities to design a new, highly efficient CX.
With the Mpower Agent builder, we allow customers to build capabilities on the platform in an agentic, descriptive way. With our Copilot and Autopilot, we allow those agentic bots to operate the business and go deliver those results and take out huge costs and improve customer service. Just a screenshot on each of these things. I'll be very quick because I'm limited to my 20 minutes. This is going to be our main stage tomorrow in a lot more detail. Within actions, remember, actions is our AI capability for the CX leader of an organization. Because all of the customer's data is on CXone, we can see everything that goes on. Again, we're applying the CX-specific Enlighten models, first of all, to uncover opportunities, things that aren't working well.
Sometimes benchmarking against other industries or other companies in the same industry to find what works well, what does not work well. Then, again, we are using LLMs and foundational models to actually visualize and verbalize what that is. You can uncover opportunities to improve or reduce cost to improve the customer experience. I am not showing it here. Again, once we have used our CX-specific AI to uncover those issues, we are using generative AI to create the prompts. I will say it again. We are using generative AI to create the prompts that describe what the Mpower Agent needs to do to solve that problem. The kind of problems we are able to find with this, and this is absolutely transformative for our customers. We are doing this right now for about 50 of our customers.
I'm yet to come across a customer who hasn't been blown away by the opportunities we're finding. We're finding generally opportunities to reduce cost, improve customer experience, increase revenue, or make sure they stay compliant in a highly compliant industry. The way we do that is either through automation, take people out, and that could be people in the front office or middle office, augment people to give people tools to do more in a more consistent way, or increasingly, and that's incredible, proactive outreach. It's very interesting that frequently the best resolution is actually to go and not wait for an incoming interaction, but reach out proactively to an organization or, sorry, to an individual to do that. That then results in very, very, very practical things that go out.
A new IVA, an Mpower Agent that gets built, changes to a website, for example, updates of knowledge, for example, changing the workforce scheduling of people, updating the quality plan. These are all things that are very easy and relatively incremental changes to make on top of the CXone platform that then delivers that ROI to our customer. Here's the builder for the CXone Mpower Agent. Again, this doesn't do it justice. If you go downstairs, you'll get to play with this. You can play with it yourself. This is where you write your prompts. This is where you can copy and paste a prompt that came out of that first design thing to build a bot. It then goes away and creates this. You can then click on these nodes, and you can make change behind each of these codes. Nodes is code.
That code can then be changed to customize those things. We expect our customers to create hundreds of these Mpower Agents to do all the different kind of business tasks that they need to do. They'll sit behind our customers, Copilots and Autopilots because, as I mentioned before, as they operate, here's an example of Copilot. An employee, because our Copilot, again, is listening in on a conversation, it will trigger an Mpower Agent or give the option for an Mpower Agent to run and do a bit of fulfillment for that agent, making it extremely efficient. Of course, on the front end, when we have Autopilot running on our website, make that Mpower Agent available to the end consumer as well, taking out the human being altogether because you can actually automate that process from the front end.
It was rushed, but Marty will be happy with me. I've got 75 seconds to go, so I'm going to take my time on this last slide. If I want to summarize what differentiates CXone Mpower, it's a unique AI platform. We combine foundational LLMs and CX-specific models. No one else does that. We do it this way because we have those CX-specific models that we've been doing since 2019 with Enlighten. The fact that we're using AI consistently as part of our platform, we're not tacking it on or a point solution. It means that AI is used consistently across all the different applications. Those applications feed data back into AI so the AI gets better irrespective of the application or the touchpoint. Our automated insights, again, leveraging those CX-specific models that appear within actions, it highlights those proactive decisions, the actions, the outcomes at scale.
It's how our customers can really transform their organizations and deliver the ROI, save them money, build them revenue, basically funding what they then pay us in license fees. Finally, our new Mpower Agents, truly agentic in terms of how you build them, but actually focusing on the resolution of complex intents around fulfillment and delivering on those workflows. Marty, I was one second over. I apologize. Thank you. Now, Elisha from Hyatt. This is an amazing customer. I'm going to say a few words now. I'm going to take a bit more extra time. Elisha, great to have you here. This customer, Hyatt, and I'm sure you're going to speak to, they've been a partner of NiCE for a long time. What your team has just done in record time is so impressive. I get daily updates from Andy and others about how the project's going.
I am so proud of what you guys have done. It's amazing. Congrats.
Awesome. Thank you. The pressure is on, I guess. Today, I want to just talk to you guys a little bit about our journey with AI adoption with Hyatt, particularly in the knowledge management space. I'll just go through why did we approach this in the first place? Why did we even look at knowledge management? Why were we even partnering with CXone for this? I also want to touch on something that we call the digital rocket ship as well. On top of knowledge base, we also looked at Copilot as well, a tool that we use to make knowledge access even easier and faster for our colleagues. I'll walk through how we evaluated Copilot and how we partnered with our teams that we have across our operations to make that come to life.
I want to talk to you guys a little bit about some feedback that we're getting from our colleagues and from a performance perspective. What exactly are we seeing now that we're more than 90 days past our rollout for Copilot? First, Hyatt Hotels. If anybody's not familiar with Hyatt, we were founded in 1957. There are 28 unique hotel brands that we have. If you guys can think about that, 28 brands, that means there's a lot of information that our colleagues have to memorize in order to better serve our guests. For the contact centers, when the guests are calling us, they think that they're talking to the hotel. We want our colleagues to be as knowledgeable as possible so that way they can help them fast and make it as easy as possible for them.
You guys can probably tell that's a little bit difficult when you have to memorize 28 brands and then all the booking rules that go within those 28 brands. You have all the different segments like luxury. You have all your select service, everything there, a lot for our colleagues to remember. What we're going to focus on, I know we see 160,000 global colleagues. That's Hyatt overall. We are going to focus in on the contact center space. If you hear me say GCC, that stands for Global Care Centers. That is the name of our contact centers. For the contact centers, we have a mixture of property services and guest services. On one hand, you have the colleagues that take care of the guest experience.
On the other hand, you have the colleagues that take care of everything behind the scenes to make sure that our hotels are operating. It was like a perfect blend for us to really try AI in this space because we have two different tracks of work. You have two different levels of complexities that they have to go through in order to get their job done. For our guests, they have to move fast. They have to make sure they're getting them answers as fast as possible, making their reservations as fast as possible. I think, like you said, people are still calling us, but they don't want to be on the phones forever either, right? They want to talk to somebody, but they don't want to be on the phones forever. That's exactly where our colleagues come in.
On the operational side, they have to support our colleagues. They need somewhere to go that way they can get information fast to resolve things as fast as possible. That way our colleagues have the space to take care of our customer. Just a little bit about me. My name is Elisha Wright. I have been with hospitality for about 14 years now. Two and a half years of that has been at Hyatt. Fun fact, when I came to Hyatt, I did not expect to have knowledge management under me at all. I was strictly learning and development. Think about being in a classroom, training people, falling to the floor, and then you do it over again. Knowledge management was not in my purview. I really had to think about how do I associate this now with my role.
That is when I started thinking, wow, knowledge management to me is way more important because that is what happens after training. That is how we embed knowledge in the flow of work, which is what they really need. We just went through four weeks of training. Do we really expect them to remember all of this? We need to give them a tool where they can find that information fast. Now, it is my favorite part of my job, to be honest with you. I was scared of it at first. There is a lot of content that has to go into it. Once you start realizing, wow, you have to write it like this for the colleagues to be able to digest it easy. Next, you know they are helping their customer.
When you see the excitement on their face saying, wow, I learned all of this and I do not have to memorize it because Elisha or Tia back there, who is on my team, by the way, showed us where we can find that information. Something that I try to tell people is true market intelligence is not everything you know. It is how you find that answer. Why do anything in the first place? When I came to Hyatt, we had a knowledge-based system. It worked. It worked, but it did not work well. You had to be very, very specific on what you were wanting to find. At Hyatt, we are very known for our acronyms internally. If you cannot think about, okay, what is that acronym, and you get one letter wrong, you were not getting results.
Sometimes you would put the article number that goes specifically to the article that you're looking for because you use that one every single day, and you still can't find that information. Also, the system was a little bit slower. Obviously, if it's slow, that means you're slow responding to your customer. That means they're getting more upset because it's taking a long time for you to resolve either issue, make their reservation. That way they can get on their way. It also was outdated UI. I think a lot of feedback that we got, I'll quote a colleague specifically, "This is not really pretty. This is ugly. I wish we had something a little bit more modern." We knew that we needed something more modern.
We knew that we needed something with natural language search and not just a keyword search because there's going to be times where you say, "What is something?" or you're thinking like something that's close to it, but I don't know the exact word. We needed something that our colleagues do not have to think much. They can just type it in and find exactly what they need. That is what Expert gave us. The rollout for us was absolutely amazing. Honestly, it was like night and day for the colleagues. Something I was super nervous about, it was my first big rollout that I have ever done, even at my previous job. You're expecting to get that feedback where you're saying, "I don't like it.
We're so used to the other way. It was ugly before, but we were so used to that ugly that we want to keep it and don't call my baby ugly. When they got this new baby, it was almost like immediate positivity, especially when they started realizing that this article that I searched for before, it was at the bottom of the list. Now it's at the top because it's realizing that I'm looking at this every single day. Although we have fears of AI because our colleagues hear AI, they start freaking out because they think it's going to take over their job. We were able to do that subtle AI for them. They don't even realize that they're using AI.
CXone Expert, I'm going to touch on Copilot a little bit more as we talk here because Expert for us was just a launching pad. With Andy's help over here, we came up with something that was called the digital rocket ship. What that means is we have Expert as that launching pad. It's our foundation. We start to layer tools on top of that so that way we can continue having knowledge access, not even just to our colleagues, but also to our customer. That's when you start to hear things called Autopilot, for instance. Although I wasn't part of Autopilot, it was part of that digital rocket ship that helped bring knowledge to our customers as well and also still supported our colleagues. The main next step for us, though, with the digital rocket ship was going into Copilot.
What that was going to do was instead of now me having to go into knowledge base and find this answer, Copilot is analyzing my interaction and is sending that knowledge to me based off the intent of what the customer is looking for. That has been a game changer for us as well. I am going to go a little bit into Copilot. You probably could have taken a few minutes too, by the way, because I am looking. I still have some time here, but I probably could have given you a few more minutes. Evaluating Copilot. Again, we started with CXone Expert as our single source of truth. We had to first partner with our digital team and our AI council.
The cool thing that I love about Hyatt is that once they realized that we were looking to adopt AI, they did not want to slow us down. They did not want to stop us. What they did was they formed a council. That way they were able to put governance in place and also help us make sure that we're assessing the risk and we're doing all the right things as we start to bring AI into our environment and making sure that we had a soft landing for our colleagues. We partnered with the digital team and the AI council to go through what were all the pitfalls that we may expect, what do we anticipate, what do we need to partner with NiCE to make sure that we have done before we even bring this in front of any of our managers even.
That was a big help for us. We also partnered with the NiCE team because they helped us with making sure that we made all those changes. They made sure that they got all the answers as fast as possible for us as well. It really helped us move fast. That way we were able to adopt Copilot as fast as possible. Just like you said, it was record time. That was the partnership that we had internally and with our partners at NiCE, which I already talked about here for the partnership with NiCE. We also partnered with our operation leaders as well. Operation leader was somebody that I noticed that they were being ignored in our rollouts that we had. Even I had to say for CXone Expert, I could have done a better job with bringing our operations leader in.
That was a lesson learned. If we wanted to move fast, we had to make sure that we had their buy-ins because ultimately, they are the ones that are going to be leading the colleagues. They are going to be the ones listening to all the complaints if there were any. They are also going to be the ones to be the cheerleader for us as well. We made sure that we partnered with the operations team so that way they can not only support the colleagues, but also support us as we were building the training around Copilot. We made sure that we were talking about all the right things and ensured that we understood what are the fears of AI that they are hearing out there. Let's make sure that we're addressing that. A strong partnership with the operations team.
We also have our partners at ALG. It's one of our brands within Hyatt. You guys are going to be hearing Nick on the main stage tomorrow speaking a little bit more about AI. They were also going through Copilot. We ended up partnering. As we were testing, we were sharing our lessons learned. As they were testing, they were sharing their lessons learned. Because of them, we were able to really find that ticket to say, we know exactly what we need to do in order to get Copilot working. I want to stress the partnerships that we had to really make this happen because it really helped us with that real-world impact that we want to see. That's the last slide here. Copilot, adopting Copilot for our colleagues, we focused more on chat and not voice first.
It was a little bit low risk for us because we know that if they have to read it twice, three times, they have a little bit more grace from our customers. Where if you're on voice and you just don't understand something, they might say something a little bit out loud and the customer's like, "What are you talking about?" We decided to test it within chat and specifically with North America. We had our voice colleagues still using Expert. They were still doing the manual search. The one thing that our team had to do, and I have to give a big shout-out again, Tia, she's our Senior Manager of Knowledge Base.
One thing that she did as she was doing all the assessing with our ALG team is we recognized that the way that we had our articles written, they were great for our colleagues. It was amazing. They were able to find information faster, but our AI was not able to find that information. Our colleagues get four weeks of training. Copilot gets zero weeks of training. We had to go and reformat almost all of our articles so that we can make them a little bit more concise. We can make them shorter. We can make them more action-oriented. That way Copilot can understand, "Okay, this is exactly what the colleague is trying to do," and be able to give them that correct answer.
When we first started testing it, we started noticing that we were getting some responses that like, "Man, why are we getting this response?" As soon as you're ready to give up, you're like, "Wait a minute." We're realizing that these two articles look the same. We also saw opportunities to even combine some articles and realize that we're really duplicating information where we don't need to. I talk about our voice colleagues still using Expert manually. It also helped them out. Even though they weren't using Copilot, now they have a knowledge base that they can go back to, and they're not finding redundant information anymore. We also had to make sure that we addressed the emotional side. Again, there was fear of AI. People thought they would be losing their jobs if we incorporated AI.
It was important for me to have people understand that this tool is here to support them, not to replace them by any means. How we did that was choosing a few colleagues to sit with us, letting them use it, give us their feedback. Let's make sure we're getting back to NiCE. We go back to that partnership that we had, going back to NiCE to make sure that we addressed any of those concerns and getting back in front of them and getting them more comfortable. We created cheerleaders within our colleagues from our pilot. That way they went to the masses, able to sing the praises. I have to say that we had an amazing rollout. It's probably the best rollout that I've ever had, even with Expert. We have a heavy population of outsourcing colleagues now.
If you think about it, those outsourcing colleagues, they're from different countries. They're trying their best to understand the terminology. They're trying their best to understand hospitality overall. Now they have this knowledge base. They have Copilot to help give them the information as they're trying to get comfortable with, "Hey, there's this new company I was with." If you guys don't know, outsourcing colleagues, they typically are with another company first before they come to us. Now they're thinking about all the different companies they were with before and having to get out of those habits. Copilot and knowledge bases would help them do that. Now I'll show you guys just a little bit of a snippet of how Copilot works in action, I think. Wait, let me go back. Oh, that paused. Okay. I didn't press anything. Oh, okay.
They have their CXone open. They see a new app for Copilot. You guys see they're talking here with the guest through chat. We can see if the customer sentiment is positive. We can see that if they're getting angry with us, their confusion. You have a summary as well. It summarizes the interaction. If they are looking for something, they can always go back to and see, "Wait, why am I here in the first place?" This is what we call the KB answer. This is where the answer from the knowledge base comes to our colleagues, and they can read it quickly and then be able to respond to the customer. If they wanted a little bit more information, this web link, it takes them straight to the article.
It is still cutting down on that search time because they're not having to go look for it manually. The article link is coming to them directly as well. This is an example of the web link. It took them directly to the article. Another thing I love is that colleagues can give us feedback as well. We take that feedback to continue making Copilot better, continue making knowledge base better. Like I said, as the conversation progresses, Copilot provides updated responses, additional resources, and then continues to update that sentiment as the conversation goes on. They might start happy. They might get mad in the middle based on the answer that we gave them. Our goal is to make sure it's positive by the end. That is Copilot.
Again, I want to talk to you guys a little bit about some performance highlights that we've been seeing. Almost immediately, we saw a decrease in our average handling time. That means that the colleagues are getting the answers fast, and they're responding to the guests faster as well. Now we are saving the customer time. We are saving the colleague time and saving the organization money as well. Another thing that I did not include in here, though, is this customer satisfaction. The one thing that you can see is that things can decrease. What if they were giving wrong answers? Because I could send you something and say, "Hey, you're good, but do we know if it's right?" We did not see any decreases in customer satisfaction. That means that they are getting the right answers.
They're not having to call back and say, "Hey, I have this complaint because this person told me this, and I was given this instead." Customer satisfaction is still exactly where it needs to be. This is just some quotes that we have from our colleagues just talking about Copilot. They see the value in it. They look forward to seeing all that it has to offer because we continue working with our friends at NiCE to make enhancements to Copilot. We have a few more features that we want to release. They're happy with what they have now, and they're excited more that we're going to give them more features to make this easier for them.
My favorite thing is always when people say that they would recommend Copilot to others because you can come back to me and you can say, "Hey, I like it." It is something about recommending it to somebody else that, to me, is the true result. The true satisfaction I'm looking for is that people are telling other people, "You need to use this," especially because we have colleagues who still have that mindset of, "I don't want to use it because if I do, that means I'm contributing to it taking my job." When they start talking to their friends, they start talking to other colleagues, they're saying, "No, I use it for X, Y, Z," and now they're getting best practices as well. Looking to the future, we are looking to expand Copilot to different languages. Specifically, we're looking at German right now.
We've actually done demos for our contact centers globally. That's within China, Japan, Korea, Germany, and India. Every single one of them saw it and said, "They want this, and they want it now." So much so that now I have weekly meetings with them. Even though we're not ready to go there yet, they just want to know more about Copilot and what we can do to get it to them. German is where we're going to start. I'm excited because we are a global company. We want to make sure that we're rolling out a tool that everybody would be able to use. All the products that we've had with NiCE so far, they had that global reach. If we go back to Expert, in China, they have a different level of expectation when it comes to service. People want you to move fast.
So they needed a knowledge base that was going to provide that for them. Expert did that. I believe that Copilot is going to do that as well. I think that this is just the beginning. Again, this is something that I was not expecting to come into whatsoever, except with knowledge base. Now it is my most favorite part of my job. I look forward to just continuing building with the NiCE team as we implement more features and probably, hopefully, more products from NiCE so that way we have an ecosystem of NiCE products that take care of our customer.
Okay. Thank you, Elisha. There is nothing better than hearing directly from our customers how CXone Mpower is really driving tangible results and really enhancing customer experiences, your consumers, and your employees as well. Thank you for presenting today.
I want to start off today by really talking about a high-level overview of the last five years of financial results at NiCE and demonstrate the strength we've consistently delivered. We're really well-positioned for a strong financial foundation to launch our next era of profitable growth. At NiCE, we have consistently delivered expanding revenue and growing at a compounded growth rate of 13% over the last five years. This total revenue growth, not surprising to anyone, is really being driven by our cloud and AI platforms, and in particular, of course, our CX Mpower solution. At NiCE, we do have the largest cloud revenue base in the industry. Our cloud and AI platforms are made to scale to the organizations you've heard from today: Hyatt, Walmart, Disney. Very highlighted earlier that really there's only a few companies that are able to scale at this level.
You can see that our growth is demonstrating our success. Our cloud platforms, with that underlying ability to scale, means that we consistently deliver high-margin cloud platforms. You will see that is playing out in terms of how we actually manage our business at NiCE. We have always maintained a disciplined approach in cost management. You can see this ongoing continuous margin expansion where last year we exceeded more than 30%, delivering a 31.1% operating margin in the business. We have a proven ability to optimize our resources, scale effectively, and continuously enhance productivity. This is something we have always maintained as a balanced approach at NiCE, which is driving the top-line growth while always having a focus on also our profitability. This produces our long-term financial results and this profitable growth model. All of this profitability is conveyed and translated into our best-in-class free cash flow generation.
Our free cash flow generation was more than $700 million last year, with a free cash flow margin that exceeded, almost reached, 27% last year. Our free cash flow is really unparalleled when you look at our industry. It provides enormous flexibility and optionality for us at NiCE to make ongoing strategic investments, execute on value-added M&A, and also continue to return capital to shareholders. This is extremely well-positioning us as we look at the next phase of our profitable growth with this strong foundation that we have. Next, I want to spend a little time on our five key growth catalysts at NiCE. These key growth catalysts quickly are both catalysts that have been recognized and represented in our financial statements over the last five years but are also key to driving this ongoing growth looking ahead.
The five key growth catalysts are first, the secular tailwind of organizations that continue to migrate from legacy on-premise providers into the cloud. The second is our global profile that we have at NiCE and the international momentum that we have been delivering. The third is our proven ability to continue to cross-sell and upsell into our existing install base. The fourth, Scott talked quite a bit earlier today, which is that we already have a strong ecosystem of partners that is continuing to be expanding. Finally, the rapid adoption of our AI. These five key growth catalysts are really a common thread across all parts of our business at NiCE. Given that our customer engagement business is such a large portion of our overall revenue, it represented 85% of our total revenue last quarter.
When I go from here and dive into each one a little bit, I'm actually going to focus mostly on the customer engagement business and provide some more specific CX data. The first growth catalyst is what we talked about, this ongoing secular tailwind of organizations looking to shift to the cloud. You can see that we have benefited from this strong tailwind over the number of years with ongoing increases in our cloud revenue. The secular tailwind for NiCE over the earlier years was really in the small and mid-sized markets. We've increasingly taken a larger profile in the large enterprise, which I'll talk about a little bit more momentarily. Importantly, as I highlighted earlier, we are the largest cloud revenue base in our industry. In fact, in the most recent quarter of Q1, we exceeded $2.1 billion of ARR in the cloud.
It is really important to highlight. Here you can see this migration and the tailwind that has happened that over a five-year period, we have gone from less than half of our revenue at NiCE to being now at 73% of our total revenue. This was at the end of 2024. Actually, in the first quarter of this year, that has further increased. Now we are at a 75% growth rate. When you think about our overall total revenue growth and the expectation there, more and more of our total revenue is being driven by this higher cloud revenue growth. That is what we expect to continue to see in the future as a cloud-centric and AI-centric company.
Before I leave the slide, the last thing that I would highlight that's really critical is that every CCaaS deal that we sign at NiCE is incremental to our revenue. Our business was really built on the CCaaS offering, which was something that we introduced several years ago and continues to provide that incremental revenue to our top line. I'm standing here with the key growth catalyst of the secular tailwind of cloud migration. I talked about how at NiCE, our customers have really shifted over the years to starting more in the SMB. Today, we continue to work with the SMB, but as well, we have expanded into having more and more large enterprise customers. In fact, we now have more than 401 cloud customers. You can see as well the great impact that it's had on our revenue.
Now if you look at our cloud revenue, more than one half of our cloud revenue is coming from those $1 million plus large ARR customers. Our entryway into the large enterprise is really just beginning. Scott mentioned earlier that the market is really estimated to be about 35%-45% penetrated at this point. That means there's still a significant opportunity looking ahead for us in the large enterprise where we have always played extremely well and have years of experience at the high end of the market. Further, the penetration of AI in these organizations is even less, which means that we just have an enormous opportunity as we expect to see these large enterprises continue to accelerate their expansion and shift to the cloud and to our CCaaS platform over the next several years.
The second key growth catalyst is the international momentum that I talked about. Our international business now represents approximately 10% of our overall cloud revenue. That's grown at an impressive 52% compounded growth rate over this five-year period. We've recently announced two significant wins in our international business. One was in the APAC region about a year ago and more recently, one in our AMEA region. Both of these impressive international deals were eight-digit ACV deals, and each had a TCV exceeding $100 million each. Really great momentum that we're seeing. This is an area we've been focusing investment in the last couple of years. Given the strong growth rate that we're seeing, it's continuing to be expected to deliver more and more of our revenue from our international business.
The expectation is that as this concentration grows, you would see a positive impact and a creative impact on the overall cloud and total revenue growth rate. The third growth catalyst is our proven ability to cross-sell and upsell into our existing install base. Here you'll see the validation of that. Starting with the first quarter of 2022, you can see how our customers are repeat customers. We have a very sticky and loyal base of customers that are continuing to come back and buy more and more with us. This is increasingly important if you think about what you heard from Barry and others talking about today, the complexity at the high end of the market, the needs that they have. We see these customers continuing to come back over and over. These customers really, as I said, are very loyal customers.
Once they're with NiCE, we see that they're customers for life. The fourth growth catalyst I won't spend much time on. Scott talked about it quite a bit earlier. It's the strategic partners that we have. I think Scott mentioned already that we had now 70% of our business was partner-led during the course of 2024. This is a very big transition from where we started at NiCE. Typically, many years ago, we were really a direct go-to-market. Now more and more of our business is being led by the partner ecosystem. In fact, 70% of our new large enterprise CXone ACV deals were driven by our channel partners.
In terms of some of the newer partnerships that Scott highlighted earlier today, we would expect them to start coming into our financial results really at the very tail end of 2025, but much more gradually to see those results during the course of 2026. The fifth and final actual growth catalyst that I'm sharing is our AI adoption. You've heard a lot this morning on the main stage. You heard from both Scott, Barry, and Elisha from Hyatt about the strength of this business. We wanted to share that in the first quarter of this year, 25% of our new cloud bookings were our CX AI bookings. When you look at the success we're seeing, you can see it's going to continue to have a great impact looking ahead for our growth at NiCE.
One year ago, we had growth of about 29% year-over-year. That has already accelerated to a 39% year-over-year growth. What's equally important that I want to mention is if you were here last year, I talked a lot about our consumption-based pricing model. When you look at this AI and self-service revenue, more than half of this revenue, the majority, is coming from a consumption-based pricing model. Our customers are really in the very early stages of AI adoption. You've seen some of the data that shows the great increases we're seeing in transactions. Those transactions will continue to come into and drive this growth ongoing. That's an important facet. Similar to what I shared for international expansion, the growth here is expected to be accretive.
Based on the amount of AI that we're selling, we'll continue to have a positive impact looking forward in our growth rates. Another slide that I'd like to share with you here is how much of our base has actually been penetrated by AI so far. Two years ago, about a quarter of our customers had adopted at least one of our AI and self-service solutions. Today, that stands at about a third of our customers. We are seeing this greater adoption. What's extremely important to highlight is that on average, our AI and self-service customers have an ARR in the cloud, which is typically about seven times higher. This is really important also when I show the opportunity ahead in the large enterprise. The large enterprise typically has a significant amount of complexity. They have greater needs for the AI.
I'm going to share with you now a few success stories of customers that demonstrate a few customers that have already deployed our AI and how it's playing out in terms of the monetization opportunity for us at NiCE. I'm going to go through three customer success stories. The first is a long-standing customer of NiCE. They are a global hotel chain. This customer, several years ago, adopted our CXone platform back in these early years where you see the OCR and workforce engagement. This customer decided that they wanted to go through a self-service transformation. They went through a competitive bidding process. They included us in that bid as well as some of our competitors, but ultimately selected NiCE.
First, due to the strength of our self-service application of Autopilot, which they adopted, but also because they had a really great deployment with our CXone Mpower platform earlier on. That experience they had and the stickiness that I've talked about as a customer are what keeps those customers coming back along with our best-in-class technology around AI. Now our AI, as a percentage of this customer's ARR, is about a third of their recurring revenue. This is the existing customer example. Now I want to go into the next customer success story, which is a new logo win. This is a large, well-known financial services company. This company decided to go through a DIY or do-it-yourself initiative in their organization. They decided that they would go on their own and try to deploy self-service capabilities.
Ultimately, they found out that it was much harder than they expected, and they failed. Ultimately, they also went through a competitive bidding process. They selected NiCE due to the strength and best-in-class technology we have in AI. They have now adopted three of our different AI applications. They're using Autopilot. They're using Copilot. They're also using some of our Enlighten models, which were specific to their vertical as well. We have a significant amount of data that we've accumulated over the years that's really focused on intent. That also was a significant factor in this customer ultimately selecting Empower and our platform with the AI capabilities. You can see that for this customer, in less than one year, we've increased the recurring revenue to $8 million, which is a multiple of four times increase year-over-year.
The third customer success story and the final customer success story is really interesting. We have gotten a lot of questions over the years around customers and what will happen to our business as agents decline. This is a terrific example of a customer that does not have human agents with NiCE. This customer is using only our proactive AI agents. They are a 100% self-service customer. This customer has had a great increase in their ARR as well with almost a four-times multiple. We have a very satisfied customer. Ultimately, they were looking for containment in their organization. Because they have had such strong results, they have also been able to now shift more of their focus and are sharing with us that they are driving a lot of top-line growth as a result.
These are three examples that share the success we've had with some of our customers. These are real examples of well-known marquee brands. I want to just transition to going from how these customer success stories are playing out in terms of our financial results, these customers, as well as many others. This is a Sankey diagram. This is demonstrating our 2024 results. I wanted to share this with you because it really demonstrates the strength that we have at NiCE and the resilience and continuous focus on how we drive our business. You can see our impressive 71% gross margin, the operating income that we delivered last year of more than a 30% operating margin.
Of course, all coming with the robust and durable free cash flow generation due to the strength of the top-line growth, as well as keeping that focus on profitable growth. It underscores the strength and durability we have of our strong operating model. It also demonstrates that at NiCE, it is one of the muscles we have always maintained, which is a focus on delivering operating leverage. We continue to do that through our best-in-class cloud and AI platforms. When we drill in a little bit further, you can see just the strength that we've delivered from this profitable growth model and how that played out during the course of 2024. We delivered more than $700 million of free cash flow last year. This really makes us unmatched in our industry.
It provides us with optionality to drive business quickly, to make strategic investments, and have a very clear focus and much flexibility in terms of what we do around capital allocation. This is a view of the three prongs we have in our capital allocation strategy. We continue to always have a robust, and we have an ironclad balance sheet at NiCE. We want to maintain that strong balance sheet. We want to maintain that capability to be able to put our liquidity at work. We had about $1.6 billion of cash available to us, which we continue to put to work through the next two prongs, the first being the strategic and disciplined investment. Regardless of whether we are investing internally or whether we're looking to execute on acquisitions, we are always doing this in a very targeted and focused way.
Earlier this year and today as well, Scott talked a little bit about where our focus is. First and foremost, we are continuing to drive innovation around our R&D spend and specifically focusing and continuing to hire in those areas. In addition, one of the things you'll see shortly is where we've invested in our cloud business, meaning specifically in our international regions. Over the last 12-plus months, we have been very focused on making investments there. It's really paying off from what you've seen. The third and final program and the final prong of our capital allocation strategy is around our share buyback. I talked about the strength of our free cash flow that we generated last year.
As well, in the first quarter of this year, we delivered a record cash flow generated from our operations that allowed us to complete a record share buyback of more than $250 million. In parallel, we announced a new buyback program of $500 million at the same time. All of our capital allocation strategy really comes together. We are always looking to balance investment with value creation. Of course, it is all underscored by our ongoing commitment to return capital to our shareholders. I have talked a lot about the discipline that we have at NiCE and the way we intentionally focus where we spend. Here you can see the strength of the cloud margins and the overall gross margins we have at NiCE.
One of the things that I would highlight that I mentioned as we came into the year of 2025 is that we do expect to see a flattish gross margin in the cloud this year and the overall gross margin. This is as a result of those intentional investments that I talked about. They are already paying off. We have created foundational infrastructure outside of the Americas to continue to grow that 10% overall concentration from our international business even faster. That is one of the key areas that we have focused this year. These investments are continuing to pay off. You can see that we continue to maintain that focus on this really constant delivery and a great operating model driving this profitable growth.
Next, I want to flip back and actually reiterate all of the guidance that we shared last quarter. Last quarter, we reiterated the top-line expectation of total revenue at 7% and the cloud revenue expectation of this year of 12% growth. Our operating margin, we expect to see about a 50 basis point expansion this year in the operating margin. Despite accelerating investments in key areas, we have other efficiencies we will utilize to continue to deliver on this expansion. Finally, the EPS growth of 11%. I can say that with Scott at our helm, we are super excited about our next phase of growth at NiCE. We started the day by mentioning that we do plan to have a Capital Markets Day during the month of October.
Currently, we expect it will likely be in New York City, just so you can mark your calendars. The expectation is that we would unveil some of our financial expectations beyond 2025 at the Capital Markets Day. Please stay tuned for that. Finally, in summary on my end, just want to remind everyone about the fantastic market that we operate in and the competitive advantages we have. We operate in this fast-growing TAM with these continued tailwinds that we talked about of organizations and the market, which is still highly underpenetrated, especially at the high end and even more so with the AI opportunity. We have an increasing global presence with the international momentum we are delivering. We are very excited about that business, which is really thriving.
We have an expanding customer base with a proven validation that they come back and are continuously looking to purchase more from NiCE. We have a very broad and deep, rich base of our platform, CXone Mpower, that allows them to continue to come back and meet more of their needs and their organizations. Finally, we have decades of CX domain experience. Of course, all of this is underscored by the financial strength that I talked about, which really is unparalleled in our industry. The amount of liquidity available for us to invest, the amount of liquidity available to return capital to shareholders, all of that is very unique. The runway ahead for us is significant at NiCE. We are extremely well-positioned to continue to capitalize on that opportunity. As I mentioned, very excited about the growth ahead for NiCE.
I think I'm going to hand it back to Marty before we take a short break. Thank you.
Why don't we take a 15-minute break? They're going to roll lunch into the room. We can grab lunch, and we'll come back in 15 minutes. For those on the web, please stand by. You'll hear music for the next 15-20 minutes. Then again, we'll be back with Q&A. Thank you. Why don't we begin with the Q&A session? We have Barry, Scott, and Beth, and they're here to—they're happy to answer any questions you guys have. If you have a question, just raise your hand, and Coron will bring over the microphone and just identify your name and your firm.
Right. Siti Panigrahi from Mizuho. Great presentation. Keynote was a lot of energy. Scott, I think your message, if I understand correctly, like you're saying 15 million agents that we use to track the market opportunity. Now look at the billions of interaction, massive interaction. That's where the opportunity is. Help us understand how should we think about the market opportunity there and your monetization strategy. How are you going to monetize that whole billions of interaction?
Yeah, sure. I'll start, and Barry, I might hand to you if I miss anything. You're right. If you think about the history of Contact Center and the world that we've been in, we've been centered around the agent and the platform to interact. That interaction, whether it be the ACD and being able to handle voice or any type of channel in the interaction, or with the agent and the experience, the performance, the accuracy of the agents all around that initial contact. If you think about that world, it started and stopped at the interaction.
Now, we are the best in the market, and we're really good at that. If you think about our addressable market going forward, the interaction will only increase. Whether it's via voice or whether it's via chat or whether it's email or any other mechanism that happens going forward, that interaction volume continues to increase. What you do with the interaction, how you solve the intent, that is a huge opportunity for us. That is why when we talk from intent to fulfillment, we used to talk about our role on intent, and we would be the intent to interaction. That is only a very small step in the journey. Intent to fulfillment means that you're handling workflows, you're performing tasks, you're delivering services, you're automating outcomes. You're doing it by humans and by, through, obviously, AI agents and bots.
That gives us two means. One is it means that we can increase value for our customers to serve the end customer, which means we've got a premium that we can drive there, which right now, if you just talk about bot conversation, it's a pretty commoditized market because it was based on such limited information and knowledge that it could use. The more value we drive, the more savings, the more out there, the more we can drive. That's the first part. Secondly, if we're performing tasks that were otherwise done by backend systems, people, processes that were fractured and redundant, and we're able to optimize that, then clearly there's an opportunity for us to be able to monetize that.
Ultimately, the way I view it is, right now, if an interaction occurs, we have one shot of monetization, and that's usually through the agent. One interaction, if you take the Walmart example this morning, that one interaction might have five or ten different intents within that one interaction. Every one of those intents will trigger an Mpower Agent. Every Mpower Agent might use Copilot, and then you might then use auto summary. The AI monetization for one interaction could explode. That's part of the reason why we're best presented. It's not a one-for-one, and that's the opportunity in front of us. Barry, do you want to add anything?
Yeah, I mean, you said it. I mean, we've talked about it a bit earlier as well, but ignoring agent-based pricing for the time being and looking at the interaction-based pricing, the market's, as Scott said, very commoditized. A conversational AI bot is going to cost you about 20-25 cents if you're a customer to use that. And Salesforce came along saying, "Oh, it's going to cost five bucks." Everyone laughed because it's not going to be the case. You can't go there in a massive market and change that overnight. We realize it's very hard to differentiate what you charge for that conversational AI experience. It's going to stay 20-25 cents, I believe. What we can do, as Scott just said, is monetize the backend and the fulfillment. Our Mpower Agents are charged per database that's updated, information that's retrieved.
That is tied directly to value because that work is replacing middle and back office work by expensive people, and it is differentiated from the Copilot cost or Autopilot cost as well. That is the plan. Initially, we will bundle Mpower Agents with Copilot and Autopilot so it is easy to get adoption, but then it will be turned on and we will start monetizing fulfillment.
One more question on CCaaS migration. I think you talked about maybe 35%-45% right now since a massive 55%-65% are still on legacy on-prem. First, what do you think is going to catalyze that kind of migration, that massive legacy base? Second, why are you not going with your AI solutions, like going to this legacy? Because we keep hearing from customers they are using some other standalone AI.
Why are you not going there like a beachhead strategy, trying to capture them? When they're ready, they'll move to your platform.
The answer is we are and we will, and that is a really critical point. We have competitors who are very clear about trying to monetize either their own install base and moving from on-prem to cloud. That has a limited runway, and we will win more than our fair share of that. If you only get them to a cloud, but you really get them to an AI-powered platform, and then you find a way to lead with that rather than an add-on, you are truly changing the game because you are giving them a leapfrog capability, and it gives us a differentiated way of being able to—that's not just about our CCaaS platform versus somebody else's.
We believe very strongly that our AI platform is leagues ahead. We're obviously making it even more powerful through partnerships, but you're absolutely right. There's no surprise that the volume and the usage rates and the expansion that you've seen in the last six, even in the last six months, is a direct correlation to that.
Just anybody on the web, remember if you have a question, just please type it into the web interface, and we can take it here. Tyler?
Yeah, hi. Tyler Radke from Citi. Thanks for doing this. Great to see you and see the energy from the keynote. First question, just kind of bigger picture. NiCE as a company had grown its cloud revenue north of 20% for many years.
I think as you look at some of your bigger competitors in the space, they're talking about well north of a 20% cloud growth rate. Scott, just as you've come in, I'm sure your aspirations are much higher than 11% or 12% that you've guided to this year. How do we get back to that 20%? Is it going to kind of require these new use cases around knowledge management and orchestration, or is it simply just kind of a timing thing and macro-related?
Yeah, I think there's a few things. First of all, I think I might have mentioned once before, I've got a pretty good track record of taking businesses that have got moderate and good growth in the cloud and making them exponentially. It does cover a number of pillars, but I'll just reiterate.
One is using the market forces to our advantage. In the market force of AI, we have not seen the potential in our—you are seeing the glimpses of it, but it is not yet at a mainstream. We have just launched Mpower Agent. You fast forward 12 months, 18 months, and our ability to be able to monetize that addressable market, that is a growth catalyst. More broadly, back to the earlier question, starting, finishing, and all being in the AI platform is a key element of our growth. That will then lead to, secondly, using the CCaaS migration, which is still ongoing. Beth mentioned the big international wins recently. There is more and more and more of those, not using that as the endpoint, but getting them straight to the AI-powered platform, which gives us that exponential growth that we have got. The third is international expansion.
I guess what you've seen with NiCE is we've been really good recently at being, had some international wins. You can rightly expect that that will continue. Beth talked about it because the same platform that we've built for the North American customers, but these are not North America-only customers. You listen to Elisha. He's got in Japan, in Mexico, in Venezuela. They're all around the world, and our ability to be able to help international companies expand and then local companies utilize the same platform. Last but not least, I am excited about this, is the partnerships as a growth driver. What I can tell you is every one of those partnerships that we've signed, I've got direct targets on it.
Yes, there is innovation that Barry and the team are driving, and I care passionately about building one plus one equal three or five. Ultimately, it has got to drive exponential growth for us, and it is probably something that is a newer muscle for us. We have done some good partnerships in the past, but it is something that I think that will give us a huge lift. I have obviously not talked about inorganic, but there is clearly the flexibility, and I have talked about this before. We do have the flexibility, but it will be strategic growth. It is all about what will drive long-term shareholder value. When we look at those investments, it is very much about, and potentially inorganic moves, it is very much around the long-term growth. I think we are ready to seize that opportunity.
Look, we obviously have not talked about it much today, but I think we can talk in much more detail in October, as Beth mentions, around those midterm and that long-term outlook and what that growth prospects. It is fair to say I have high expectations.
John?
Hi, it is John DiFucci from Guggenheim. Bear with me for a little bit because there is a question here, but I saw a lot up there, and Tyler's used my long-winded questions. Really, my thoughts are around AI. I am not as close right now to NiCE as maybe some of the people in the audience, but I have been close, and I do pay attention. I was talking to Beth
about this in the break, and I mean, it was great to see you owning AI here because I remember in 2015 going to Israel on a bus tour and seeing a demonstration at NiCE and Barry, you probably were part of this. Actually, it was AI. It blew away the room, but it was not called AI back then. I guess my question, and it is a bigger question, even bigger than NiCE, but it is sort of twofold. We hear some people cover Salesforce and ServiceNow, and ServiceNow is in the back of my badge here too, so they are a partner.
About a year ago or a year and a half ago, they woke up and said, "Hey, we're going to be an AI player." NiCE has been doing this because you had to do it, because if AI became anything, it was going to kill you, and that is what you had to really work on a long time ago. It is not even just you, it is everybody in your sector. Maybe not everybody, but I think a lot of others too.
I guess my question is, is that advantage, the fact that you've been working on this is for Barry, is that advantage, is that something that matters today, or is AI just taken off in such a way that, like, you know what, I can become an AI guy if I want to because it's all out there, I just go out there and get it. On the go-to-market part, and Scott, I think this one's for you, is it really against your peer CCaaS vendors, or is it against everybody? Because I don't know, maybe NiCE isn't just a CCaaS vendor anymore. I'm sorry for the long-winded.
Yeah, and by the way, we're not a CCaaS vendor anymore. We're not, so that's clear. Look, when it comes to competitive AI solutions, you're asking kind of two questions, I think.
One is, can generic AI just win? Do we need the sophistication that we've got? Is that the question you're asking? Yeah. Yeah. And the history that we have. Yeah. Okay, and the history. Yeah. Yeah, that's what I mean by the sophistication and the history kind of thing. I think it's one question. There's another point, I think, which I'll tie it to, which is kind of the benefit of a platform, and I talked about it a little bit there. Don't underestimate that. When we compete, we're either competing against other platforms and maybe not CCaaS platforms or organizations that came from there, and they don't have that heritage of, we call it Enlighten, the CX-specific models. Now, for certain use cases, you're right, you can get away with it.
Like the order summary example I gave, the benefits of summarizing a contact or a call, a chat, or an email, if it's 90% accurate or 70% accurate, doesn't really matter if you're still reducing two minutes of call time. It doesn't matter. Other use cases, like the other one I showed here, the auto-discovery of what you can do, that really matters. That really delivers value. There are 30-40-50 AI use cases in CX. I would say that 10% of them, GenAI versus our heritage and what we do, it doesn't really matter at the end of the day, but the rest of them, it makes a massive, massive difference. The other thing I would say, that's when we're competing against other platforms.
The alternative is we're competing against AI specialist point solutions, and this is where the power of the platform really comes in. I'll incorporate in there as well as the build option where customers say, "I'm going to build my own AI solution that's specific to me." The benefit of the platform is insane here. The example Beth gave earlier, Beth, I'm going to do a tiny correction to what you said. It's not that their AI that they built themselves failed, it's that they built it, it was perfect for them, it worked really well, but the CIO said, "I don't want to maintain this." My God, as I said to Scott, a pet's not just for Christmas, it's for life. He's like, "I don't want to be in this business.
I want the software companies that that's what they do. Increasingly, we see organizations playing with homegrown solutions, but then realizing that actually that's not a business they want to be in. They want to outsource that to specialists like NiCE that do it. With the point solution as well, we come across, yes, you can get these amazing point solutions that are very good at a very specific thing, AI-based selling in insurance into a particular segment. Amazing. Again, because it's trained on specific data for that particular segment. The problem with that is you need five or six of those solutions. You need AI governance five or six different times. You get disconnected consumer experiences that work one way in one situation and a way in another situation.
You end up with an AI Frankenstack that's just like the CTI Frankenstack that we just got rid of. I strongly believe that our heritage around Enlighten, that differentiation matters for the vast majority of use cases and delivers value. The auto-discovery is a great example of that one, but also platform wins. We will go through another year probably of customers thinking you can build it, but as always, as every single wave that's come, we all know whether it was the internet, whether it was mobile, whether it was cloud, once you go through a year or two of build, customers realize, "Shit, I don't want to be in this business. I need to outsource to a specialist."
I would add two more things just to round that out.
The first is the pivot and the branding that I provided and the pivot that we've announced and owning AI. If we didn't have that historical capability, I would have been more measured, taken my time to do that. Be assured, I got up there and I made a bold statement to the market, "We will be the AI company that will humanize and create a reimagined customer experience." That is a lofty ambition, but one that I feel that we're able to step up to because of that history. The second is, I believe strongly the market will continue to see domain deep categories where you need the context matters and horizontal players where it's good, and that is where the build is because there are use cases across an enterprise where building yourself is okay. That might touch into our world.
There will be an overlap, and it's also okay. When you've got a consumer that has high expectation about the accuracy, the timeliness of the response, and then the fulfillment of what's going to, you can imagine in five years' time, no one is going to sit there waiting on a call. We think of it now because we've lived that life. Consumers will not, and that cannot be delivered through generic platforms. It's got to be domain. That is not just in the CX world. I believe that is in other domains as well. I would highlight that does not mean we've got all the pieces to the puzzle. There is a lot of work that we are going to do organically, and I believe there are also opportunities for us because of our financial strength that Beth and the team have delivered over a long period of time.
Again, strategic value, long-term shareholder return, that's what we're after. Before we go to Mark, one other thing, to your CCaaS question, how do you—
Microphone. Microphone.
It's an anecdote, but we had our ECAB yesterday, Executive Customer Advisory Board, big companies in there. The CIO of a Fortune 50 company who's part of our ECAB used the term, "And other AI companies like NiCE, we're not a CCaaS company. Our customers perceive us as an AI company."
Yeah, it was very intentional. Yes, by the way, go to market, quick response, it's a pivot of our company. When you've been living, breathing, thinking, working around CCaaS, and that's been your livelihood for 30 to 40 years, and you then become, it's not immediate.
Don't worry, on the go-to-market side through our partners, who we partner with, again, the signaling of partnering with the AI of Amazon, of ServiceNow, of Snowflake, that's intentional because of that pivot.
Can I have a question here on the web? Oh, I got to look at the one—so maybe this is for Barry or Scott, I guess. You mentioned data has to go through CXone Mpower to power the AI-driven outcome, but that is tapping into other systems of record. Can you or do you need to replace these systems of record to truly empower your AI self-service platform?
The answer is no. I mentioned the hubs when I did my presentation, and those hubs include an integration hub, which actually reach out into any kind of system of record, be it a Salesforce, a ServiceNow, an Epic, you name it, those systems.
We're leveraging that data alongside the conversational data to do that. Another thing we have, and we announced this yesterday, and it's going to hit a subset of the market, but we use Snowflake internally within CXone as our data lake. It's the technology that we use. A good portion of large companies also use Snowflake, and with that, we have new innovation, new technologies Zero-Copy, which basically extends the data model of CXone into all other Snowflake-driven applications in an organization. That allows us to seamlessly access that data and use that in fulfillment and for our AI as well. I have one more question from the web here. Scott, please explain why exactly NiCE has the surface area to win with agents relative to Salesforce, which seems to be the main panel in the front office.
First of all, you've got to break down agents. There is no doubt that there is a role that enterprises, and it's not just Salesforce. Salesforce, ServiceNow, Microsoft, Google, Amazon, and I think you'll see more and more of players like OpenAI and others where they will have AI agents that are easily created just like ours is, but it will be to serve a lot of tasks that cover a lot of different human tasks that happen within an enterprise. Now, there's a couple of things to remember. Number one, that if they are incorrect and it's inside the enterprise, it's okay. You imagine your own experience where you're using your own Copilot, and it's not quite accurate enough. It's not quite detailed enough. It's not quite context-sensitive enough. That is back to the models.
It's back to the question of the Enlighten models, the context-sensitive, the CX-specific insights that we have that are already in our models. When you're using an AI agent in the role of customer service where accuracy is of vital importance, you might be interested to know, by the way, enterprises are looking at the accuracy of human agents, and then they compare that to our AI models and our Autopilot and our Copilot and how accurate that is versus when the humans did it themselves. You would not be surprised to know that when they do that comparison, the accuracy that we have now with our AI is even higher than the human side. The generic players are an inch deep, a mile wide. We are the opposite. We are deep in domain. I want to come back to the comment that was made.
We are not a CCaaS-only player. We are a CX AI player, and we are not limited to the CX and AI because I do believe our total addressable market will expand. We are an unavoidable contact. That single pane of glass, why would you go to Salesforce for that? Why would you go to ServiceNow for that? Because they can't deliver all of it. If you say, "Oh, well, I want to use voice. Oh, well, I have to go back if I'm going to use my ACD." Does that mean there's no way a customer is going to say, "Well, I'm going to use a different platform"?
If we can be the best and own that and then be able to build out more and more actions, more and more flows, more and more tasks, then we go way beyond the contact center as we traditionally think about it into being an enterprise platform that delivers AI to ultimately fulfill customers' needs and intents, but we can go way beyond. Now, I haven't stated that anywhere, but you can clearly see the expansion opportunity. We've got to earn the right, though, and the way we earn the right is where we are domain deep in the best, and that is in CX.
Anybody else in the audience here? Catherine?
Catherine Trebnick, Rosenblatt. On the large enterprise, you talked a lot about large enterprise. Can you explain where you're headed with the mid-market?
Because it does seem the concentration of discussion on large enterprise that you might be abdicating that mid-market 500 seats. But before I do that, I have to say congratulations on the new branding. I do like it.
Thank you. Thank you. It was very exciting inside of NiCE, the reaction to the branding. It was. It was corona. Yeah. Amazing. But interestingly enough, I've been at pains to remind our team again and again, it's not just the branding. This is a statement of vision and intent of where we're going. Look, to answer your question, and I smile because Beth and I often talk about when you share one piece of information, the byproduct of what you haven't shared and what's the implication of it. We are clearly highlighting our strength in the large enterprise.
That has clearly been something that we believe will give us a huge uplift of growth potential because all of those large enterprises, and you've seen it, they're in their early days of AI rollout and adoption and expansion. We've just launched AI agent to CXone Mpower agent. The potential growth there on the upsell and cross-sell. That is why we highlight it because we see it, and it's a larger and larger proportion. Now, when it comes to the mid-market, interestingly enough, that's where the point that we made about the 110 additional partners is critical. Doubling down, and I guess I do have a lot of experience of doing the best of both, getting a volume, low-touch, no-touch capability through the ecosystem that is able to expand and scale. Our international expansion, you'd be interested to know, is nearly exclusively through partners.
Even here in North America, our ability to have a platform that is able to meet the needs of those customers, that is able to benefit, the platform actually helps in the bid market because it is easier for them. They do not want to have—they cannot manage four or five-point solutions. They do not have the IT shop. Helping our partners be more enabled to be positioning. I do not need to feed on the street in the sales side. They can do that, but we support them. That is a key element. Things like the AI Center of Excellence that we announced is a key enabler to support that partner community. We definitely see that.
I was just going to add as well. Of course, the small and mid-sized segment of the market is still attractive for us.
I can tell you, I actually had breakfast this morning sitting with one of our long-standing customers, and they said they have 650 agents in their contact center. They're distributed across the U.S., but one of the key reasons that they're here is to learn about Copilot, Autopilot, and proactive AI agents. It's applicable across all segments of the market. As Scott said, we emphasize it's more because the opportunity ahead is still less penetration on the large enterprise, but we have a significant number of customers. I'm sure if you walk around this afternoon and you talk to some of them, you'll find that the mid-market is also extremely excited about everything that we're doing.
That's exactly what I was going to say. I was going to say, this is the record conference we've had. We have 20% more attendees here than ever before. They're not coming from the high-end because the high-end only serve many customers. They're not coming from the SM small market because they don't travel to these things. It's coming from mid-market. The people we have here are from the mid-market.
Thank you. Okay.
Thanks for this, guys. RK Mahendran with HMI Capital. Maybe two questions. One on partnerships and one on AI. On partnerships, there's one notable name missing from the recent announcements. I'm just curious what that implies given their ambitions on agents and when we as investors, analysts, etc., can start to see some of that flowing into the financials. So I'll ask that, and I'll ask the next one.
Okay. I'm sure we don't need to guess who you're referring to. Two comments on this. Clearly, Salesforce, and you'll see if you're here tomorrow, you'll see the presentation with Barry.
We're already a great partner, deep integration. I think we do more with them in this space than anybody else. We've already got a great relationship that's already there. I can tell you, I've been very clear about these partnerships. They've got to be incremental for both of our customers. Announcing a partnership that we're integrating, I could have done that day one. Each one of these partnerships is very much about incrementality. There is more than more benefit, mutual value, and we've been working diligently with Salesforce on that. I look forward to sharing not just with Salesforce, but with other key partners. Do not read anything into it other than it's a critical relationship that we want to make sure that it's going to deliver true incremental benefit.
That is why I said, "Hey, we know there's more to come," because there is. When you see that, first of all, remember that every one of those partnerships has required engineering builds. We have announced it. We know what we need to build. Barry's team, whether it be with ServiceNow connecting to the workflow and the AI engine of ServiceNow using Q Business and Bedrock of Amazon or with Snowflake in Zero-Copy data, there is some engineering work to be done. It is a matter of months. We start deploying and rolling that out. I probably will not be calling it out separately because it will be a part of our AI growth, but I do see it as an accelerator. Obviously, inside the management, we have clear measures of success in growth targets with those partnerships.
Got it. Helpful. On AI, it was really helpful to see the chart, albeit no numbers, but of the volume going exponential. You sort of talked about how there is an interaction, sort of usage-based component in the pricing. I know it is evolving. How should we think about how much of that $208 million of AI that you guys disclosed is coming from that volume metric growth versus more per seat or fixed basis?
Yeah. I specifically called it out that actually the majority or more than half of that revenue we are showing you in AI and self-service is coming from those consumption-based models. Whether it is based on sessions or interactions, it is more than half already today. Of course, because we are in early days with AI, the consumption is going to continue to expand. You saw that in some of the data that Scott shared earlier this morning.
That is a great opportunity that we look at for the potential upside looking ahead as well.
Hey, guys. Billy Fitzsimmons from Jefferies here for Saman Samana. This one is probably for Beth. I will expand on kind of the metrics question. Several metrics were disclosed. Obviously, it stood out that AI and self-service are accelerated year-over-year. The 33% AI and self-service penetration stood out as well. That metric struck me as impressive and maybe even higher than I would have thought it would have been given how early the opportunity was. First, I just want to understand what is in that metric and those AI and self-service metrics in general. I know in the past you guys had talked about AI and digital. I just want to make sure if there are any nuances between those two or if it is just naming.
Then second, just want to understand for that 33% AI and self-service penetration, how is that measured? If a customer bought one AI product, would they be included in that? If that were the case, I'd imagine the spend penetration would be like a totally different number, right? There'd be a lot of opportunity to expand there. Just how should we think about the runway for AI and self-service adoption over time?
Yeah. You asked quite a few questions, so let me make sure I touch on all of them. I think I'll start with what's inclusive in AI and self-service and the revenue that we're reporting. I think first, it's important to highlight that we've been doing AI for many, many years. John asked about that earlier. What's not in that number is some of our machine learning-based AI.
This is really looking at our next-gen AI. It is inclusive of our digital channels, as you highlighted. That is what we called out last year when we talked about the disclosures that we had. Of course, with the 33% of our customers, it is actually being pulled together data that is down to a SKU level of what did we sell to a customer. Yes, it is based on if a customer is buying one self-service, an AI application, or digital channel, that is inclusive in that AI revenue. Given the stat that I showed, 25% of our new bookings recently were coming from those AI and self-service. We expect that to continue to have that strong momentum and continue to compound that growth that you are seeing. That along with, as what I mentioned a few minutes ago, around that most of it is consumption-based.
That's another element that most of our customers are still in their very early days of adoption of AI. As they continue to roll it out, you'll see that play out and come through in the growth.
Yeah. Maybe just to give you a bit of context on the timing, this is like any other major technology shift and change. If you think about it, and you just go back to what you heard from Elisha. Elisha has a lot of brands, a lot of service professionals, and they started with knowledge. They didn't start with agents. They didn't start with workflows and knowledge, and they only used Copilot for a certain location and for one brand.
What they're wanting to do, because you can't afford to get it wrong, you can't guide the agent incorrectly, the knowledge and the way that it's surmised and the guardrails that you put in there, there is a bit of work there. It's not technically the work, it's the accuracy. It's the change management. I do believe that we're not over that hurdle where it's then, "Okay, I get it. Now I scale it." We're seeing a lot of customers use, trial it, and they get success, and they are able to do so in a matter of months. The scaling opportunity excites us the most.
Because once you do it for one use case, your ability to do it with 100 or 1,000, let alone what you see with Mpower Agent and others, obviously that gives us a strong view that we've got a great growth appetite with our customer base that they will do naturally and that they will do it. They've already done it in other parts of the business. I'll just give you an analogy. No one in my engineering teams even thinks about using AI tools to help them be able to generate code and test and things like that. All of Barry's team, it's pervasive. It was one of the first things that was rolled out with AI. Now, getting them to use it really, really, really well, we're still on that journey. CX is the same thing.
Thank you.
Back there.
Hi there, Clark Wright with D.A. Davidson for Gileria. Historically, you have noted that there's been a 2x uplift when switching from on-premise to CXone. However, in the presentation today, you noted that they're directly now switching from on-premise directly to an AI-enhanced platform. What are you seeing now in terms of the uplift? Has there been any changes in that regard? I have a follow-up.
Yeah, no, there hasn't been any change. In fact, we've talked about we have customers that have seen an ARR expansion up to a 10 times multiple for an existing customer of NiCE. That's a legacy workforce engagement that's moving over to our CCaaS platform. We continue to see a great uplift as those customers move.
I think what's interesting, and I heard Scott say it just a few minutes ago, is that when you think about the opportunity in the market, we are out there to really grab those new customers, bring in the new logos, and we have a tremendous track record of those customers being customers for life. With the existing customers that we do have, most of those tend to be in the very large enterprise that have been doing business with us for many, many years. That is part of the customer base that looking as part of the cloud migration in the next few years will likely look to shift as all large enterprise makes that move. The multiple and the opportunity ahead still continues to be quite significant.
I don't think we were pretty accurate in the on-prem to cloud and the multiples and what we saw, but that was largely because we already knew what the revenue was. The multiples on the AI side, I think you'll see that evolve. You would have seen I presented a chart. I should have mentioned it. You would have seen agentic AI at 45% compound annual growth. I think that is grossly understated. Grossly understated. No analyst has gone out there and said, "In the CX market, what is the agentic AI growth opportunity?" because no one's really, we haven't yet seen it materialize. That's just on the agentic side, let alone on GenAI, which does have market data that shows high compound growth rates.
We'll keep you updated as we see more measurable expansion opportunity for a dollar of on-prem and what that means, but even a cloud as it is today versus AI as well.
Thank you. Additionally, last year, you noted that NRR for CXone was 113%. How does that compare today? How do you think about kind of the growth mix between existing spend versus new logo?
Sure. The NRR that we called out last year, importantly, the 113%, we have just recently started introducing NRR last quarter, the 111%. That NRR, just for transparency and for clarity, is representative of all of the cloud business across NiCE. Of course, CX is such a significant portion of that that the CX Mpower platform is driving the bulk of that NRR. We haven't broken it out separately.
We will continue to provide that NRR information, but we consistently have a very high gross revenue retention of all of our customer base. Our customer base, I highlighted a few times, we have a great track record of those customers continuing to stay with us.
We have time for two more questions. Thank you.
Thank you for taking the question. This is Nick Lee from Citizens Bank on for Pat Walravens. Beth, we saw cloud grow 12% last quarter versus 27% a year ago. Can you walk us through some of the factors that contributed to this slowdown?
Yeah, sure. On the face, it is comparing apples and oranges. The 12% growth that we are talking about for 2025 is organic growth. We did an acquisition of a company called LiveVox, which is focused around outbound. We did that right at the very end of 2023.
When you look at the growth that we had in 2024, that was inclusive of a significant portion of this acquired cloud. You have to take that into consideration. We provided some data last year around kind of sizing that so you can look at it on an apples-to-apples basis. That is the reason for the change, is that the 12% is organic for this year.
Thank you. Last question.
Hey, guys. This is Damian Coge n . I'm from Barclays. How are you guys doing? Just curious what the key driver is for the on-prem to cloud migration. And if the unlock is just the macro environment or is there any accelerator like AI that would drive that over time? I think it's natural to see a little bit of a slowdown in a tougher macro environment, but any sort of background you could.
It's a good question. Still, as it's always been, the most common driver is a burning platform, out of support, needing to then renegotiate a new three-year deal. Customers don't want to do that. That's still the underlying thing. I've got to get off Avaya. I've got to get off Cisco. I've got to get off my Genesys on-prem or renew for another three or four years, stuff that doesn't work anymore. Increasingly, we've got customers that also need to leverage AI technologies. They see others doing it in their market, taking advantage of that, and they can't do it on an on-premise as well. That is a big catalyst as well. That's basically it. That's very simple.
Is there a breakdown you guys provide maybe on the on-premise base from U.S. and then international? Is that international expansion opportunity primarily driven by the idea that there are more on-premise contact centers internationally compared to the U.S.?
No. The expansion internationally, I believe, is basically down to a great success we've had with our partner strategy in international, a big focus for us last few years. If you look at both of those big $100 million-plus deals that we closed, both of those got closed through partners that have been partners for NiCE for three years or so. In certain territories in international, we've got the machine working, and we've got partners working for us when we're not, and it works really well. That's the biggest thing. I wouldn't say in terms of the on-prem to cloud migration, there's nothing materially different between Australia, the U.K., Europe, and the U.S.
What I would add to that, though, is for our international team, the U.S. was quite advanced. I mean, we've seen a lot of they were often the first movers, the early movers. Our international, you're seeing more and more. We still have, I don't know what the percentages are, and Barry, you might. If I broke down that 35%-45%, where does international versus North America sit there? I think proportionately, we could probably provide it. Australia is actually the first. The first. The most established. Then it's probably the U.S. Then it's probably the U.K. Then it's Europe. They're all in the same kind of boundary. Yeah.
I do not want to miss the point that I would not call them laggards, but if you have been holding off because of whatever factors, whether it be macroeconomic, whether it be competing priorities, and you are now looking at it, those customers are no longer saying, "Oh, I just want to move to a CCaaS there or to the cloud." They are very much saying, "I want to go to an AI platform." They are viewing it very much about leapfrogging the innovation agenda that they have got. Because they do not have to do a two-step move, they can go straight to that. That is the beauty that they have, obviously, with the CXone Mpower. Thank you.
Thank you, everybody. That will end our Q&A session. I am going to ask everybody, next 20 minutes at 2:00 P.M., let us meet outside these doors, and then we will take you down to the innovation hall. Thank you.
Thanks, everybody. Appreciate it.