Hi, everyone. I'm Ryan Gilligan, Vice President of Investor Relations here at NICE, and we are thrilled to welcome you to our 2025 Capital Markets Day. Before we begin, I need to quickly show you this disclaimer slide. There we go. Just a quick note that the presentation does contain non-GAAP figures. Okay, now we can talk about the great day that we have planned for you all. In just a moment, we're going to pass it over to Scott Russell, our CEO, who's going to lay out his vision and our opportunity. Scott will then turn it over to Jeff Comstock, our new President of CX Product and Technology. After a quick break, we'll go to Philipp Heltewig, our Chief AI Officer and the General Manager of NICE Cognigy. We are very fortunate to have Nick Algeier from Lufthansa join us.
He is going to share his perspective as a customer. Then we will go to Neeraj Verma, who is our Vice President of Customer Service Automation. Lastly, we will wrap with Beth Gaspich, our CFO, who will provide a financial overview. After that, we will take another quick break. At that point, those of you that are in the room can step outside just there and grab lunch and bring it back into the room. Those of you that are on the webcast will have an opportunity to type in questions. We do ask that everybody save their questions for this formal Q&A session. With that, I will turn it over to Scott Russell.
Hey, good morning, everybody. Hopefully, you've enjoyed breakfast. Thank you for being here. Thank you to those who are online. It's very much appreciated that you've joined us for the next few hours as we go through an update for our investors in the financial community in the capital markets session. First of all, I do want to just recognize and welcome Ryan. He obviously has only just joined us. We put him under fire. He joined about a week before our first earnings and capital markets day. He is now an expert on the company. Very, very happy to have Ryan join us and lead our investor relations as a part of our leadership team. Before I jump into the agenda, I did want to just recognize you might have seen a PR this morning.
We've announced this morning a new COO, Arun Chandra, who has joined us, who is joining us on 1st of December. Arun comes from one of our customers, Walt Disney Corporation. Before that, he was an executive that worked at Meta, at HPE, running all of their operations. He also was the CEO of three tech companies. He comes to us very much as a partner to Jeff Comstock around how to make sure that we deliver the value when we, as a company, are an AI company the way we run, and then how do we deliver the value for our customers. Very excited for Arun to join as a part of our core leadership team, combined with Ryan and Jeff and others who have joined us recently.
Okay, so today, I'm going to cover, and if I can ask one thing, one thing, it will be to remember these three things. There are three things that we're going to go through for the entire day that hopefully will resonate with you and will recognize. First of all, we operate in a great market. I'm going to talk about why it's a great market, because I think that sometimes gets questioned. Why is the market that we're in a good one? Why would that be a positive? I'm going to go into a bit of detail about the trends that are in our favor in the customer experience market. Secondly, we will clearly highlight why we are perfectly positioned to capitalize on that market opportunity.
We will not only talk about the AI trends, and you'll obviously hear a lot about AI today, but we'll talk about we, as a company, whether it be our financial position, whether it be because of the acquisition of Cognigy with the AI capability, but also just the broader portfolio that we have as an organization and how that best sets us up to be able to meet the market needs, which is a great growing market. Last but not least, is clearly highlighting and articulating how that drives accelerated growth, accelerated shareholder value via some focused investments. We're going to go into the details of that. Beth will speak in a lot of detail about what those growth drivers are, what the outlook looks like, what the growth looks like.
We will be sharing midterm guidance for you and indicators of what those variables are. We'll obviously answer your questions later on in the day. Great market opportunity, which I will primarily cover, why we're perfectly positioned, and you're going to hear from the speakers, Jeff, Phil, Neeraj, talking about our technology capability and why we believe that we're best positioned, and then how that translates into accelerated growth, accelerated shareholder value, and via some focused investments. Before I do that, let me just give, for those of you who are not deeply knowledgeable about NICE and who we are and what we do, let me just recap. I get asked this question a lot this year. For those who don't know, I'm in my 11th month. I feel like I've been here for 11 years.
I feel like I've been in this organization for a long period of time. And that's in a good way, by the way. I feel like this organization was built for me and I was built for NICE. I get asked the question, why did I join? I wanted to join a leader. I wanted to join a leader in their respective market. I also wanted to join a company with the appetite, the drive, the vision, and the willingness to seize the opportunity that AI brings to its respective market. Leadership in its current space, but the appetite and the capability and the drive to be able to go and utilize, because AI allows technology and humans to converge better than any other. I've been in the technology industry for over 30 years.
We are the number one company in customer experience, and we'll talk more about that. 27,000 customers around the world leveraging our capabilities, 10,000 NICEs in different parts of the planet serving our customers. The majority of the top banks, the majority of the Fortune 500, or Fortune 100 and Fortune 500, leverage our company every day. While today is going to be particularly geared to CX, it's a worthwhile reminder that we are not only leaders in customer experience, we're also leaders in financial crime and compliance, protecting $5 trillion a day of transactions that happen through our Actimize, our financial fraud and compliance portfolio. We have 30 million pieces of evidence supporting first response, also for those who are in need with 991 recording and being able to support in the public safety and justice business.
We are a market leader in our own right, and we are well set up. As I say, we lead the market today as an organization, and clearly we do so in a very financially responsible and attractive way. The massive CX opportunity. Why do we believe this is such a great market? I think in the hype of AI, a few things get lost. Every customer we speak to starts with this. They start with what is the business value for their company. It could be triggered by problems or issues, but fundamentally, companies or brands that invest in customer experience outperform their peers. You can do any research you want. We obviously have done research from different categories.
Whether it be about increasing your top line, whether it be about reducing your cost, whether it be able to get operational improvement, whether it be about delivering a higher customer sat, businesses invest in CX, not because of AI, they invest in CX because it delivers business returns. Period. That is still true today. I'll talk about AI in a moment. Companies that invest in CX perform better than their peers. It's proven, and they're able to showcase that. For example, just a simple example, if you take Bose, they're a great company. You might use their technology in your car or at your home. They leveraged CX1. They consolidated all on our platform. They increased their customer sat by 30%. They reduced their operating cost and their call volumes by nearly 20%. They increased their CSAT, so they're happier customers, by over 20%.
They were able to achieve, through focusing investments on their CX platform, they were able to get a better top line. They doubled their sales during the holiday season as a result of having a better customer experience. Increased customer sat, reduced cost, better returns, and better revenue. I say this because sometimes we can get caught into the hype of what AI means in this market. Companies are investing in CX, and they will continue to do so because it's good business. It is good for their business to do so. It is a market that is naturally growing. We sit here today, call volumes, digital interactions, AI interactions are naturally growing. Consumers are interacting with their brands more than ever, and it will just continue to rise. Just as a natural market where businesses are investing, they're seeing more and more volumes coming through.
Our data reinforces that, not only because we're acquiring winning new customers, so our volumes, but if you take NICE in the last nine months, in the last nine months, our AI volumes have grown by 65%. Our digital volumes have grown in the mid-40%, and good old voice has grown in the mid-20%. Not only are we winning in the AI market, it's not at the expense of others, because the whole volume of interactions between consumers and brands are increasing. Here's the thing. That number doesn't include some of the things that you're now seeing in the market. You'll hear a little bit, and I know Nick will talk about the Lufthansa example, where, by the way, they were able to automate, I think, 16 million plus interactions on their Cognigy platform.
This does not even include where you think about how consumers are going to interact more with their CX platform rather than websites or apps, because they are just going to chat. They are just going to converse. They are going to interact on a conversational platform that will increase the amount of volume that goes through their CX platform. That does not even include here. Machine to machine, where you tell your agent on your phone, "Hey, I want to call my bank. I want to be able to increase my credit limit. And at the same time, can you contact my airline because I want to rebook the flight with my family?" All of those interactions, the volume of interactions are often limited by the time and capacity we have as consumers and as humans. A lot of those constraints go away. That is not built in here.
This is just pure volume of interactions that naturally is happening between consumers and their brands. If that market was great, it just got a whole lot better. It got a whole lot better. This is a massive market. The AI potential, the transformation that it brings, the opportunity it brings in CX has got a whole lot better. Why do I say that? First of all, you're going to hear a lot today about fully autonomous resolution. I think we've all had the experience when we've called in or we've chatted to our brand of choice, and they use words like this. Look, let me just put you on hold for a few moments, if I can, please. Then you don't know what's happening.
Right now, that agent is pivoting, and they're going and chatting and interacting and conversing with different parts of their organization to resolve your request. It might happen multiple times because they might have to operate with multiple individuals in the organization. Unless it's a knowledge-based request, such as a password reset or checking a balance initially on those easy automated plays, the ability for organizations up until the use of AI was pretty limited. You still needed a human in the loop. We can do that automated with AI. We can do it again and again and again at scale. As I say before, the Lufthansa example is a great one, but it's not the only one. I'll talk a little bit about our customer references. The second is our journeys are customer-driven.
Most enterprises, when they talk about AI, they're talking about how they can replace what humans do in the enterprise. Our focus is the customer's journey. It is a very different lens on how you use AI to deliver an outcome. When you're trying to use AI to complete a task that otherwise an internal employee would do, yeah, sure, that's useful. We do that as well for the front office. We'll get to that. Our emphasis is the customer journey. You do get full resolution in an automated way. All of the interaction needs, whether you ask for it as consumers or whether the brand wants to contact you to be able to fulfill an outcome, it can all be done automated, autonomously on the platform.
That going from reactive to proactive, using the data, the intelligence, and being able to drive a proactive model rather than a reactive one, which has historically been in the world of CX. I want to highlight the seamless human and AI experiences. Every company, even if they want to go to a full automated model, will not do it in one go. I would contend that there are many businesses, and I'll highlight this, that do not want to go to a fully automated model. They want you to contact and talk to somebody at the other end. Why? Ability to upsell and cross-sell. Ability to be able to provide different levels of service.
There are some businesses, particularly in life sciences and healthcare, that they want the first interaction with AI, understanding intent, understanding your needs, but they will intentionally take you to a human agent because your needs give them an opportunity to perform a task for you as a consumer better than what an AI would be because they want that human touch. Even then, even then, AI will play an active role because we will be supporting, assisting, using the same models, the same data to be able to converse with that consumer through the human agent. AI is a tailwind for NICE, or for this market. I'm talking about the market. It is a tailwind. There is no doubt. The market's great with CX. Interaction volumes continue to rise.
We have got a market shift where the value proposition back to the core business drivers and the opportunities that AI brings are continuing to expand our opportunity. I will talk about that in total addressable market terms as well. It not only gives us the ability to automate to automated self-service, but we also have the ability to do assisted service with humans in the contact center as well. Here's the thing. While we are incredibly bullish, and we are, we are incredibly bullish about the opportunity that AI brings to be able to automate low-complexity transactions. There is, and you can debate what low-complexity looks like, but there is a large volume of interactions that are happening with contact centers today that can be easily automated on the NICE Cognigy platform today.
Today, you can do that, and you can do it in a really scalable way. We also recognize, to my earlier comment, that some of the high-complexity transactions, the things that are really difficult to be able to serve and deliver, or, as I say, where they are intentionally wanting a human in the loop, there are still a lot of interactions, a lot of interactions that businesses will serve. This is not a market where it will replace all of the human work. Even if it ultimately gets there, the period of transition is a long one and a one where, to deliver the full CX needs, you need both. You know where I'm going with this. As I highlighted before, even in that human in the loop, those interactions will still be AI-assisted.
We do not see a world where AI does not play a role in the customer experience market. Either fully autonomous self-service, or human and AI working interoperably, or human-engaged service where your customer service representative is interacting, but it is assisted by the AI platform. This is a market that is fantastic. Great business value, great business value, interactions volumes continuing to grow. Clearly, a market where AI is giving a significant shift and boost and a tailwind and an opportunity for us. Ultimately, it is a market where it is not one or the other. Human engagement in the contact center, AI automation and assistance in that platform, end to end, it is a growing market from both perspectives. Why are we positioned to win? What is the strengths of NICE to be able to capitalize on this?
Let's just start with our current momentum. These are numbers today. 20 billion plus interactions, 20 billion interactions across all of the major industries, AI interactions, human interactions, combination thereof, that are coming into our platform every single day around the world. I mentioned the growth of voice. I find it remarkable because humans still like to talk and chat, and sometimes they want to talk to somebody, whether it be an AI agent, but we had over a billion minutes of voice, good old voice, in September. Over a billion minutes. There is a lot. That is with average handling time and call times continuing to reduce because businesses are better and better. They're getting the AI assistance to be able to make calls shorter, faster, better containment. We are a company that has had 40 years of experience in this market.
We understand customer experience better than anyone. We understand the intents, the needs, the contextual, the behavioral. We understand the skills that are required. We know what a good interaction looks like and what it does not. All of that knowledge, all of that experience, all of that logic, all of that capability is core to the platform that drives all those interactions, drives those engagements, and helps businesses deliver on the business promise that I mentioned at the beginning. We are not just getting started in AI today. You saw the numbers in Q3 that we reported last week. 49% growth in our AI, ARR, and self-service. Effectively, we are growing, and there are over 6 billion augmented interactions on that platform. Of the ones that are coming in, there is a significant number that are already AI-assisted or AI-delivered in a self-service way or in an assisted way.
Clearly, the traffic growth continues to rise and grow. As you could appreciate, for us to be able to continue to grow the revenue line, we've got to have the volume that is growing, and then we're able to monetize that volume. What's interesting about the AI, the traffic growth, is this. One interaction can lead to many, many AI sessions. Just think of the flow. You first call in, you have an interaction, you've got a need, maybe you've got a, well, an example that Neeraj is going to talk about later is a dispute on a credit card. You call in, and you're interacted by an AI bot. That is trying to determine, number one, what is your intent, what are you there to solve?
Number two, and usually it'll already have the pre-built logic about who you are, how you like to interact, not only the language, the style, the behavior, the sentiment that you have, and be able to change based on the tone and your interaction. That first session says, "Look, we need another bot that is specialized in dispute." It will go into details, and it'll be another AI agent that will be triggered. That AI agent then says, "Look, this requires someone. This is a complex one. We now need to get a human to be able to step in." They hand over to a human agent. That human agent then has a co-pilot where they're interacting, and there's another session. That co-pilot then gives assistance to the human agent about how to resolve the task, and then it gets handed back to an AI agent.
One interaction, which historically you would have seen in our commercial models, and Beth will talk about commercial and pricing, that would have been a user-based human where we had seat-based pricing. In the AI world, it's not one-to-one replacement because we're able to increase the volume of interactions, and not only are those interactions increasing, the number of sessions and AI services that we can run can then be multifold. That's what's driving a lot of that traffic growth above what you're already seeing in the momentum in the transition to self-service. That clearly translates to a financial model that is strong. Not only is the AI growing at 49%, we're nearly at $3 billion in 2025. We updated our full-year guidance and increased our outlook for 2025. Even more importantly, we're a cash, free cash flow, we're a cash generation machine.
Our ability to be able to generate real value for our customers is also translated into a financial performance where not only we're growing our top line, but we're then generating true value for the company that we can then transfer to our shareholders. We don't believe ourselves only. The market also recognizes us. Not many times where a vendor can get up and say they are the number one, undisputed number one according to all of the industry analysts. This isn't our data. This is theirs. Cognigy was ranked number one for conversational AI for CX. The NICE business was ranked both highest in vision, but also highest in ability to execute in the Gartner Magic Quadrant for contact center as a service. Clearly proven leader is the number one. Ability to execute and vision. We're the number one in intelligent self-service.
We're the number one in intelligent contact centers. You can see the list. All of these analysts spend their time and effort, and I know you interact with many of them. Their job is to analyze the industry and who is best placed to be able to serve this market, deliver on those outcomes, and advise companies who to go to. What does that mean? We get a lot of inbound demand from our customers that has already been pre-qualified, validated by these industry analysts. These industry analysts are constantly assessing the movement and the change and the role AI plays, and they are very deep in our industry and the knowledge on what we serve, but also what we need to be to continue to be the best.
If they were not proof enough, I guess you can just look at the world's leading brands that have chosen us. Every one of these companies, and there are 27,000. I could not put 27,000 names up on the slide. 27,000 companies that have chosen NICE as their brand of choice, their CX platform of choice. Every one of these has a business story of value. Whether it be an espresso that implemented our platform and they were able to increase automated payment completions, automated payment completions by nearly 30%. Every one of these scenarios, and I am so pleased Nick is here to talk about the Lufthansa story, but there are so many stories. Every one of them driving business value.
Every one of them are clearly looking for the outcome, the benefits, and the results, and every one of them improves the capability of our platform because we're able to then serve more needs, more data, more capability that we can then reinvest back into an AI-led platform. How do we do it? How are we the market leader? Validated by the industry analyst, validated by the customer market, and in total revenue terms, we are the market leader as well. We are the only company that has a purpose-built platform for CX. You're going to hear this a lot from us today. What do I mean by purpose-built? Number one, it is a platform that is only focused on CX. There are many other technology companies that want to cover tasks, workflows, cases, processes. They want to cover all of your different organizational things.
They're trying to be experts at everything that happens in the enterprise. In my view, that's not possible. If you want to go deep and you want to be rich and you want to have a great customer experience, you've got to have the logic, the knowledge, the capability that is purpose-built. NICE Cognigy was purpose-built for using AI in customer experience. Now, you can use AI for all sorts of purposes. You can use it in your supply chain. You can use it in HR. You can use it in your finance. You can do it in your operations. You can do self-service in those scenarios. They built it specifically for CX. You, of course, NICE and our history, whether it be in workforce management, whether it be in the contact center, the contact center as a service, our whole company was geared around CX.
We're the only organization that has the full suite of capabilities. The AI platform is native. I'll come to this later, but let me just say this. The other contact center as a service players do not have an AI-native platform of their own. They have to use somebody else's and embed it into their platform. They don't have their own. We do. The advantage that that brings us to win in this market is immense. Jeff and Phil and Neeraj are going to showcase how that will play out, why a combined platform where you've got native capability all in the one stack matters. It's also a platform that has all of that data, that 20 billion interactions per year. That's all resident. It's repeatable value. Every interaction adds to it. We get more context. We get more knowledge. We get more insights.
We can build AI CX-specific models that are solving for those scenarios, and it's transferable. We can do it by industry. We can do it by horizontal. We've got the ability to use that interaction data, that leading volume to our advantage to ultimately deliver a better experience for customers. It also is a platform that allows for that end-to-end journey. We can get better resolution. What I mean by that is we own the point of engagement. When it comes in, whether it's a chat, whether it's an email, whether it's a text, whether it's a call, whether it comes from an AI, ChatGPT, or whatever platform they're coming in, once it hits the enterprise, we are that platform of engagement. We are the customer engagement platform. It's a moat. It goes into there. Why does it go into that platform?
Where there are rules, guardrails, knowledge, systems to interact with. You're not going to let your consumer interact and have a poor experience. Not only does it need to be responsive and scalable and knowledgeable, but it needs to operate within the framework that the enterprise has set. That's not a static requirement. Businesses are constantly evolving. What they want their customers to see, what services they want to render, how they're going to deliver against that. That platform becomes the gateway of value for a company. Historically in the contact center, a lot of those calls were simply contained. Anyone can create a bot that says, "Hi, Scott. How are you? What's your request?" I say, "Look, I want to dispute something that's on my credit card." They say, "Fantastic. I've got that for you.
I will get a service representative to contact you shortly." Finish. That's not resolution. That's simply containing that first request, and then you sit there as a consumer waiting for that call to come back, that interaction to come back. With our platform, we take the call, we interact, we're able to resolve it real time, including performing tasks in the mid and the back office through AI agents, and we're able to deliver end-to-end resolution. We own the point of engagement, but we also then turn that intelligence, that knowledge we have of customers, and we turn it into autonomous interactions, autonomous actions. I can't tell you how important that is. Contact centers were often built because they had certain levels of constraints. How many humans, how many people could be staffed? They were never staffed for peak volumes.
When an event occurs, you can't possibly, which means you then they also manage the handling time. How long are people waiting and interacting on that call? They're constantly trying to reduce it. They're also then measuring that first contact resolution. Can they resolve it? Now, with knowledge ways that the agents can, but there's no if you're a large organization and you have got tens of thousands of different types of intents coming into your business, I don't care how smart the customer service representative is, they're not going to be able to be specialized. You needed specialization in your contact center. All of that knowledge and how that gets delivered, that has now been embedded and used from an AI purposes in our platform.
We use that same data, that same knowledge, what a human agent does, how they interact, what there is, and then we look to do it in an autonomous way. Every channel, every journey, every interaction, AI in the loop, human in the loop, it does not matter. It all goes into that one platform, and we're able to then learn and use that for our customer's benefit in either a human way or an AI way. It is built on the industry's largest CX data foundation that the market has. Clearly, we can then do autonomous delivery and be able to get a better outcome, higher resolution rates, reduced cost to operate, improved customer service through CSAT. That platform then gives ongoing value, whether it be in an assisted or an autonomous way.
I think you've heard us talk about this before, but this is why when we talk about the mid and the back office and the role we play, this is what we mean. Our platform, a contact center by its very nature, the platform was built to serve that first contact. It was the first brand guardians to be able to take that inbound digital, that chat, email, text, call, and being able to find a way to resolve it, usually through knowledge or specialized tasks. If they weren't able to resolve it, they would then contain it, so then there was a follow-up activity afterwards. With AI and the build of AI agents, our platform is able to build AI agents that can solve the tasks that you need as a consumer, and we can do it real time.
No, I'm not trying to pivot our company to be a mid and back office player. We're not going to be an ERP. We're not going to be any sort of back office enterprise HR finance systems. We will build and deploy AI agents that will perform the tasks that they do in the mid and the back office for customer journeys. Remember, we're purpose-built for a customer journey, which means I'm not interested in trying to replace what a billing clerk does. What I am interested in is that there is a task to be able to approve a bill to a charge on that bill back to Neeraj's example later. I want to be able to automate that task. We can do it from our platform. You know why? Because we interact with every business system that is out there.
Every CRM, every ERP, every back office banking system, insurance systems, every back office airline systems, all of them have already got the pre-built interactions. We've already got that connective tissue. We simply leverage that to be able to create AI agents to be able to solve the consumer journey. I highlight this because AI can be used for many reasons, but just because you've got an efficient back office doesn't mean you've got a great customer experience. In fact, I would argue there are a lot of businesses that might experience that pain because they're looking at how do they reduce the cost of somebody in finance or in payable or in credit or in claims rather than thinking about what's the customer's experience and how do I make it better and how do I use AI to fulfill against that. Again, we are purpose-built for CX.
What it does mean is it gives us adjacencies. Beth is going to talk about those adjacencies. We've lived and breathed in the world of customer service for a long period of time. We will continue to do so. We were largely contained to the front office. To deliver customer service, we're able now to move into the mid and back office to complete resolution, fulfillment of customer needs on this platform. We're also then able to move into sales, into marketing. Organizational internal boundaries are being blurred. Why did you have a specialized sales outbound team that was different than your customer service inbound team? It was largely the skills and the capability and the needs that you had to serve to. It was based on the human constraint.
If you think about it, if you're interacting, which is why, by the way, many, many customers still like to have human-assisted customer experiences. If you've called in about your pharmaceutical, you're making sure that your drugs that you need for your life needs and the delivery of that is on time, that if you're interacting, hey, there's also some herbal remedies or other products that you can take. You can proactively do and you can assist them. You'll be able to upsell on the same platform. Customer experience is not customer service. We say it intentionally. It is customer experience, fulfilling all of the needs of service, but then the opportunity to go beyond that into sales, marketing, commerce from a customer journey perspective. The last thing I'll say about this is it will increase our market immeasurably.
In interactions for those who joined me, you would have seen that our total addressable market is about $31 billion in the market that we operate today. We estimate, based on the opportunity that AI brings both agentically and conversationally, but also the expansion of the CCAS market, this market continues to grow. I'll come to our growth drivers in a moment. The ability to go into the mid and back office, we will more than double our total addressable market within the next three years. This is a great market of which we are well positioned to capitalize on. If we do our job well, we're able to then grow.
Last but not least, if you were not convinced up until now, we did some research, or more importantly, BCG did some research, and they were able to assess, and they went to NICE customers, non-NICE customers. They went to a broad section of different customers. It was a really interesting bit of feedback around how companies are thinking about who they will buy their AI platform from in the context of customer experience. This is not a who will I use for AI. It is who will I use for AI in customer experience. Let me just go one by one if you do not mind. 40% of the respondents said that they would buy AI from their CCAS provider if that CCAS provider has a world-class AI platform. Why? It makes sense if you think about it.
Why would a company that's already invested all of that capability and logic into their CCAS platform, if it's got an AI-native capability, one of the first things you'll do is you'll do human-assisted? When we first launched our AI capabilities, you could appreciate the majority of our early AI platform was assisted. It was Copilot. Because you were trying to make your human agent, which was under a lot of pressure, high turnover, really difficult skills to be able to recruit and then retain, high-pressure environment, inbound, unhappy customers. We're trying to make it easier for them, better for them. We're improving the productivity. Real-time assistance, contextual assistance, using that data on the contact center as a service platform. It expanded into auto summary and insights from the different calls. It changed how quality management was done.
Instead of getting a call recording and having to decipher and a supervisor, you would then be able to do real-time insights that made it easier for managers and supervisors to be able to assist and train their agents. Then naturally, as you could appreciate, our Autopilot then became more because we already understand what the interactions are. We already understand as a contact center platform what can be automated. Again, we've got great data, and we can already tell them using automated insights, "Hey, these interactions are purpose-built for self-service." Of course, that self-service autonomous play became part of that CCAS platform. No wonder 40% of the respondents said that they will buy their AI from their CCAS vendor. Now, NICE, up until the 9th of September, that was where we got our growth from.
Did companies come to us when we weren't their CCAS provider? No. By the way, they didn't go to any of the other CCAS vendors either. Because if you don't have an inherent AI platform to be able to deliver to it, then why would you go to your CCAS vendor? It's got to be, and what the market's now decided is you can't just have average capability. It's got to be great. It's one of the reasons why Cognigy was so important because we understood what their needs were. We didn't want to OEM a third-party product. We wanted to have our native capability to be able to serve those customers, those install-based customers, the best possible way. That meant we needed a significantly improved AI platform to do it. NICE Cognigy delivers against that.
The second highest amount of choice was to go to a CX or an AI-specific vendor. NICE Cognigy is one of those. And you know about the other ones. There's a ton of capital going into these companies in the market. You know them: Core AI, Poly AI, Yellow AI, Sierra, DECCA. All of these companies are built where a business says, "I want to go to an AI-specific player to do my AI service in the context of customer experience." Now we have got market leader in Cognigy with both conversational and agentic AI to go compete with that market. It does not matter whether they're contact center or not. In fact, I'll tell you this. We are going to aggressively go after every, every, every company.
We're going to be explicit to go after everyone that uses another CCAS platform because what a great opportunity for us with the richness of knowledge that we have to be able to have an embedded Cognigy is fantastic as it is. It's going to get a whole lot better on our platform, but we are not going to say to those customers, "You must use our CCAS." We are simply going to go to those customers and say, "Use our AI platform that is a market leader in its own right." By the way, the journey will likely get you there in consolidation because we will then have the ability to show a better-together story versus not. The first, the highest places where companies will go for their CX AI needs is the CCAS vendor and a purpose-built player. We are now able to natively capture that.
You can see there that, yes, this market does sometimes go to a hyperscaler. What they usually do there is they'll do that because they're trying to build themselves, and they want the inherent use of the underlying technology. Yes, they'll sometimes go to their CRM player. Here's the thing. We get the feedback. Why would you go to a company that is focused around the internal employee and usage of a CRM platform when the whole goal is customer experience from a customer journey? That's what we do. That's what we've always done. A platform that delivers great customer journeys is very different than a CRM platform that is about the business value that you can get from your internal business outbound to customers. Very, very different. To give you a simple example is Sony. Sony hit all of these identified measures.
They had a challenge. They were a company that has got a lot of different sub-brands, as you know, whether it be the TVs, different devices, PlayStation, etc., etc. They had a whole lot of inbound where it was not easy for consumers to be able to be navigated to get the right response at the right time, the right service. The first thing they did was they consolidated all of that customer experience onto CX1 and put it all onto the platform. Once they were on the platform, what they then did is they used all the data, which was primarily human-assisted interactions. You would call or chat or whatever with Sony. We used the data, and we identified the top self-service scenarios.
We purpose-built it on the platform so they were able to create all of those self-service scenarios natively and could interoperate with the humans that were assisting. What did they deliver? They delivered an increase in customer sat. They delivered a reduction in cost because they were able to contain more of those interactions via self-service and autonomous channels. They were able to increase their revenues because they were able to get better feedback, better responsiveness that was then able to get better performance from the top line because they were able to get more upsell through the same mechanism on that platform. Increased revenue, reduced cost, better customer sat. Again, I'll just repeat why we are poised to win. We are the only player in the CX market that is an AI-native CX platform. CCAS players do not have it.
I'll say that again. CCAS players don't have it. We are the only ones that do. We're also the platform that owns the engagement. CRMs don't own the point of engagement. It has to get into the enterprise before they can do anything with it. Our platform, that first point of engagement, is what we're able to contain no matter how and when those interactions. 24/7, 365, never down, being able to make sure we deliver against that promise. We have the CX domain expertise. Hyperscalers don't have that. Hyperscalers do not have CX domain expertise. They've got great technology. The same with the LLMs, by the way, the big AI players. We leverage those. We use the generative AI models. We love the large language models, and we're very agnostic about which ones we can use.
They do not have the, they do not understand the guardrails, the rules, the knowledge, the insights, the small models, by the way, to make it cost-effective as well. You do not want to be going always to your large language models when you do not need to. We are able to deliver upon that natively out of our platform. The human and AI engagement is going to be an interoperable one for many, many years to come. It is not going to be AI-only or human-only. It will be both. That is native in our platform. Those AI domain players that used to be NICE Cognigy and are now part of us, they could only do the AI piece. They cannot do all of the other voice and all of the other channels and all the other interactions that require human in the loop.
Enterprises, especially large enterprises, must have the ability to do both. They want the ability to do both. They need the ability to do both. The native platform that brings that all together is something only we can offer. I can tell you all of those seven-figure deals that we present each quarter, and it keeps on growing. Quite often, it is just a consolidation of fragmented stack. They have used different technologies for different pieces, and they consolidate on the platform. Now that we have got an AI-native platform, our ability to do that is better than ever. We feel that we are in a great market, and we are really well-positioned to win. What does that mean for growth? I will start by saying we have a number of growth catalysts, things that we see here and now that will drive our mid-term growth.
Let me quickly go one by one. The first is AI growth across every touchpoint. Clearly, yes, our NICE Cognigy platform will be taken to every install-based customer we have. You can be assured of that. Every one of those customers has AI needs. Some of them might have already implemented something basic, something advanced. It does not matter. We are going to make sure that we leverage not only that 40% preference, but all preferences because we have got a full platform, and we are going to take it into that market, and it is a large market. We are also going to make sure NICE Cognigy is a winner in the AI CX standalone market as customers decide that is the way they want to go. It is a market leader already, and we will continue to make sure it is.
We will also, number two, automate our agentic AI on our platform. Our NICE Cognigy platform and the agentic capabilities combined with the CX platform come together into one offering, a compelling offering using the data, the knowledge, the insights that becomes better together. We will then be able to provide better automation, better offering, something unique that the market is not able to do, and we are going to do so quickly. Jeff will talk a bit about what our focus areas are from an engineering point of view to capitalize on that opportunity. The third growth catalyst is the one that you all know very well, and that is continuing the CCAS jump balls. Our win rates are improving. They are getting better. We see that. We see our win rates, but also the market continues to move across. This is a good market.
About 40% has moved. You can debate about the stats on that, but about 40% of the market has moved to a contact center in the cloud. There is a significant amount of jump balls that we are competing for and win. Let me assure you, when we go for those CCAS moves, we're going to be using our AI capability as a differentiator from what we've already got. We've already got the market-leading platform. It just got a whole lot better. I think you can see my emphasis on international and partnerships is a key growth driver. International expansion is tracking really well. I've got to acknowledge that the company had invested heavily and continues to do so in sovereign clouds and capacity around the world, U.K., Europe, Asia. We're able to then get greater expansion with our market-leading platform.
Because international has not moved as quickly to the cloud as it has in the U.S., the opportunity for us to be able to go straight to that combined platform has become even more compelling. We are seeing that, such as what you saw with DWP, that originally chose purely on the CCAS platform and then quickly said, "Look, we will use all the AI capabilities that you have got at NICE to be able to then extend and then transform even faster in delivering to the U.K. citizens." Last but not least, as a growth driver, is going beyond the contact center. I mentioned going into the mid and back office and doing more customer resolution and automation from our platform rather than purely the front office. I also talked about going beyond customer service into sales, marketing, and other areas.
That growth driver, what you're going to hear is not really—this is more an engineering growth catalyst that will drive into a sale. So, it's not really a significant number in the midterm guidance that you're going to hear. But clearly, we see a tremendous opportunity, particularly with agentic AI, to be able to grow that with this platform. So, we're on the road to doubling our cloud revenue. You can see the numbers. We're going to go from $2.2 billion to $3.5 billion by 2028. We're on the road to $4 billion plus. And that road is based on those growth drivers, based on a great market, and based on our ability to compete and win. But there are a few things that we need to continue to invest upon. And Beth will go into more details about the financial outlook, including 2026.
First of all, we need to continue to innovate. This isn't a market where you can sit still. As great as our capability is, this is a fast-moving market. There is a lot of investment coming in. We've got to accelerate our innovation, whether it be about consolidating and bringing the Cognigy platform and the CX1 platform together to become a better stack, a more complete stack to be able to offer to our customers. Whether it be about innovating in orchestration or in automating resolution with the agentic capabilities, there is a lot of innovation needs that Jeff and the team will drive in the coming years. The second is to be able to seize on that growth catalyst about international. We've got to continue to increase coverage, capability, assets, service to be able to serve that market.
is not a case of just asking if we are going to open new capabilities in new countries. We have got to have a localized capability to be able to support that. It also means recruiting and enabling supporting partners who are going to be able to serve and be able to help us serve locally in those markets. Last but not least is you could appreciate the operational and the delivery foundations. Whether it be about reducing the time to deploy, we have made significant improvement in our deployment timeframes of AI. It was initially quite a long period of time. We have been able to reduce that, and we continue to invest in the capabilities to be able to serve and get deployments faster, more valuable, more capable in a shorter period of time. It is not that we are just doubling our revenue without investment.
We will be investing, and those investments will be able to deliver significant growth and shareholder return. Let me recap again. Number one, great market. Number two, we're in a really great position. Number three, that will deliver growth. I know that the expectations are there'll be a lot of interest in what that top line and margin and what that looks like. The growth is real. I think you saw in the sentiment that I talked about in the earnings call, I was very clear we needed to see the proof, the data, and the plan that we've built to achieve this is rock solid. We've looked at all of the things that could happen that could potentially derail or impact that. Does the AI movement happen as quickly? Have we got the ability to deliver?
What if the competitors step into different spaces and move into us? We've looked at all of those angles. This is a bottom-up rock-solid plan that we are executing on. I'm not giving a guidance from the midterm that has speculation. It is based on real data, real understanding of the market, real pipeline, real customer feedback, real analyst feedback, real knowledge of who our competition in the evolving competitive landscape. We believe that we will achieve that midterm outlook and top-line growth and double our revenue over the next four years based on all of those variables. We are ready to run. It is just about execution against the plan from now on. What we want to be able to do is be able to share with you on a quarterly basis. You saw, for example, in Q3, we shared our cloud backlog.
Our cloud backlog growing at 15%. We clearly are seeing both pipeline and the underlying proof points of a future revenue that is growing and is going all in the right direction based on executing against this plan. Hopefully, that makes sense. Hopefully, the context of who we are as a company, the market that we operate in, and the top-line growth that we're expecting to achieve is very clear. For the rest of today, what we want to do is quantify it. We want to take you through the real technology stack, what it really looks like. We then want to go through the financial models and what it really looks like. Obviously, at the end of today, we will come back and answer your questions.
We will have the whole group, all the presenters up on stage to be able to answer questions that you have. With that, I'm now going to welcome onto stage. He's been with us for six weeks, but it feels like six months. Probably not as much as mine. Jeff Comstock, Jeff joins us from Microsoft. For those who have not had the chance to meet Jeff, he was the leader of all of their CRM, their CX, his company. He led the build-out of a contact center platform ground up. He obviously led the CRM portfolio for Dynamics. He understands this space exceptionally well. We're very lucky to have him. He will be able to share with you the outlook and where we're going from a product and technology direction. Jeff Comstock, thank you.
Okay. Good morning. It's great to be here with you. Great intro, Scott. I appreciate that. You've got a lot of context on what I've done before joining NICE. I was with Microsoft for 25 years. For many of those years, like Scott mentioned, I was building CX-related products. I got the opportunity to scale those products through the last few platform shifts that we've seen in the industry. What became really clear to me at the very beginning of this platform shift, which is agentic AI, is that things are going to be different. CX is going to fundamentally change. How CX is delivered is going to fundamentally change. NICE is incredibly well-positioned in the face of this change. In this session today, I want to go through just a few things.
One is, what are the trends that we're seeing in the marketplace already underway? How is NICE positioned with those trends and in that context? I'll cover our product investment areas, like Scott mentioned. Right? How are we going to leverage that position of strength and our unique CX assets to lead and win in the agentic era? After my session, hopefully, that'll be super clear. Let's start with the trends that we're seeing. I don't think I need to convince anyone in this room that consumer AI has gone mainstream. Right? There's over 800 million users of ChatGPT alone, hundreds of millions of users of other tools like Gemini, Claude. What that means for CX is consumer expectations have skyrocketed. Right? As we know, unfortunately, most enterprises aren't meeting that expectation.
The gap between expectation and what enterprises are able to deliver is widening by the day. This is obviously a key impetus behind our acquisition of Cognigy, our market leader in agentic AI and conversational AI. You are going to see that in action when Phil comes up and demos the product. It is really helping our customers close that gap and then some in terms of those expectations. Second, the AI that augments the workforce in the flow of their work has really gone from market buzz to strong customer demand. Of course, NICE is really well-positioned here with specialized copilots out of the box. We are seeing a surge in usage, a surge in revenue from that. In Beth's session, you are going to see how that shows up in our revenues. The third big trend is platform consolidation.
In the CX space, it's highly fragmented, like Scott mentioned. It has been for many years. Unfortunately, enterprises have just had to, à la carte, pull together these best-of-breed capabilities, integrate them together in a very fragile way just to support a basic CX function. Now we're seeing a real big push by customers to accelerate that consolidation. Why is that? Customers even realize that to get the most out of AI, they need to consolidate and simplify their data estate. This is where NICE is really well-positioned with the CX1 platform. It's arguably the broadest and deepest CX platform on the market. As Scott mentioned, we're seeing customers, as they come on board to CX1, they're consolidating a number of other solutions onto CX1.
Now that story and that value proposition is just going to get that much better as we bring Cognigy on and we expand those platform capabilities. Next is the fact that agentic AI is really enabling us to leverage that point of customer engagement that Scott talked about and automate more of the tasks and jobs that need to be done to deliver incredible customer experiences across the enterprise. We can automate those tasks. We can orchestrate across the enterprise in a way that just really was not feasible before agentic AI. This is a huge focus area for us. There is a lot of strong customer demand. At the end of the day, they want more automation. They want better customer experiences. We are going to have a really big opportunity to expand our value proposition to our customers.
We're seeing a really strong demand for proactive experiences. Scott gave a really good example in his walkthrough. Right? The best consumer experience, if something is going to go wrong, for example, is for the brand to notify you that something went wrong, and here's what they're doing to resolve it. Right? That's just a better customer experience. We've got great solutions at NICE for proactive engagement, and we're seeing a surge in usage there. I'll also say this proactive engagement capability is going to play another strong role as we build out our agentic AI orchestration capabilities. Right? Because the direction we're going is supporting more holistic customer journeys, and that's going to involve proactive engagement, reactive engagement. We have all the component parts to that. We are really well-positioned from a platform perspective. We have a platform advantage.
Now I want to focus on how are we going to expand that? How are we going to leverage that advantage and differentiate even further? To do that, I want to show you the most common point of fragmentation in the enterprise. This is essentially every single enterprise contact center on planet Earth. Right? They have one system, one stack for human-assisted service. This is the land of CCAS. Right? This is where NICE has been a market leader for a long time and continues to lead the market. This is a simplistic view. Most enterprises have a collection of capabilities: digital channels, bots, virtual assistants. Right? This has long been a real big challenge for enterprises: fragile integrations. Each stack is operating on an incomplete data set. These challenges magnify as organizations try to advance with AI. Right?
As they're going on that AI transformation journey, this becomes even more of a problem. As they go to automate, they've got to automate it in two different stacks in two different ways. That automation drifts over time. Even though we're seeing from customers a really strong motivation to transform with AI, it's fragmentations like this that are standing in the way. How are we going to help in this particular situation and others and further differentiate our platform in the process? We're going to do that by bringing Cognigy onto the CX1 platform natively. That is going to have game-changing implications for our customers. Because now they'll have best-of-breed capability for self-service with Cognigy, best-of-breed capability for human-assisted engagement, and market-leading agentic AI capability. That's not where it stops. Right?
When we bring it on natively, what that means is we now have one data layer for customer engagement. Think about that data set. Right? That is every single intent, every customer need, all the steps that have been performed, whether by automation or through human-assisted engagement, and then the outcome that was achieved across the entire enterprise, across all interactions, across all channels. I will tell you that is by far the most valuable data in CX. This is the most valuable data. Right? We will have it in our data layer. We will apply our unique set of CX-specific purpose-built AI to derive insights. Right? How do we get better outcomes, including cross-sell/upsell, like Scott mentioned? Right? We will take those insights and we will feed a common AI stack that is supporting both self-service and human-assisted service. Right?
We're learning from all engagements, whether it's automation or human-assisted service. Everything we learn feeds the AI stack that then improves both self-service and human-assisted service. This creates a very powerful compounding AI learning loop that quite literally, with every interaction, it gets smarter, automates more as it goes. Right? This is our platform advantage taken to the next level. No other vendor has the best-of-breed capabilities, for one, as Scott mentioned, let alone all integrated into one platform powering this AI learning loop. That's what we're building. That's where we're going. I've talked about this platform advantage that we already have. Let me just briefly review the CX1 platform today. It is composed of three distinct product areas. On the left-hand side, and I'm talking about that white arc across the top, is automated experiences. Right? This is where self-service is.
This is where we just upgraded in a big way to best-of-breed capabilities with Cognigy. Then we have orchestration of workflows. This is classic CCAS, orchestrating customer engagements across all channels of engagement. On the right-hand side, the augmented workforce, another area NICE has led for many years in the WEM/WFM space. All of these product areas sit on the same stack. Right? They're powered by the same customer engagement data, the same purpose-built AI models, but they're independently adoptable. We sell them independently. They can be adopted independently. As customers adopt more, they get compounding value as a suite. This is very important because, again, as Scott mentioned, this is a very fragmented market. Customers want to consolidate, but we've got to meet them where they're at. They will start where they have the most urgent need, and they'll expand from that.
We have the CX1 platform as a distinct advantage that we already have and that we'll be building on. In terms of that fragmentation, meeting customers where they're at, the NICE team has done an incredible job over the years building adapters so we can land in any environment in the enterprise. Highly fragmented. Customers don't need to do a rip and replace. On the left-hand side, we have a ton of adapters where we can link right into the channels that they already have, right? Even other CCAS and ACD systems. On the right-hand side, we have hundreds of adapters and connectors to CRMs, other mid-office, back-office systems. This helps us bring in context from the enterprise. This will also be a key facilitator for us as we expand beyond the contact center and orchestrate more workflows across the enterprise. All right.
Let me kind of close here on our investment themes, our investment areas. First, no surprise, I've talked about it, is to accelerate Cognigy. Our focus there is very specifically to accelerate Cognigy's standalone value. Right? Phil and the team have done an incredible job getting Cognigy to a leadership position in conversational AI, self-service, and agentic AI. We're just going to pour a whole bunch of fuel on that, accelerate that roadmap, and extend that leadership position in the market standalone. The second vector of investment around Cognigy, as I mentioned, is to bring it onto the platform natively. I've already touched on how that is going to take the CX1 platform, take it to a whole nother level. Incredible platform benefits. That's also going to accrue benefit back to Cognigy standalone.
Because as it operates on the CX1 platform, we can easily bring in market-leading capability like proactive engagement that I mentioned earlier. All of a sudden, Cognigy will have a world-class knowledge management system and expert because we can just bring that along for the ride. And all the AI models that we have that are CX-specific also go right into Cognigy. In multiple ways, we are going to accelerate Cognigy standalone. It's a key land and expand for us. This is an area where a lot of customers want to start their AI transformation journey, and we're going to be there for them. The second place we're going to focus is extending our enterprise leadership from a platform perspective. We partner with some of the biggest, most complex brands on planet Earth, and we treat these customers as design partners. They need further feature functionality.
We build that for them, but we build those at the product level so other customers can benefit. The second place in the platform we're investing heavily is to ensure that our platform has planet-scale resiliency, availability, and performance. As Scott mentioned, we're getting tons more interactions on the platform. As we bring Cognigy natively onto the platform, that's going to drive more interactions. Of course, we're planning to execute many more agentic journeys on the platform. We're getting ahead of that, getting ready for that surge of volume. The last place, the third sort of big investment area, is this notion of orchestration. We have a number of capabilities in the platform we're going to leverage there. As you can guess, Cognigy, again, is going to play a role from an agentic capability.
We're making net new investments to enable these types of orchestrations. In a little bit, Neeraj is going to come up and really bring this to life for you and what that looks like when we're orchestrating more holistic customer journeys across the enterprise. We'll get more into that. I'll just close with this. Agentic AI, it is going to fundamentally change CX, and NICE is incredibly well-positioned. We have the specific and unique CX assets, and we have the plan to not only lead but to win in this new era. With that, I believe we have a break coming up, 20 minutes? Fifteen. Fifteen minutes. We'll be back in 15 minutes, and then Phil will come up and demo Cognigy live for you. All right. Thank you. A lot.
All right. Welcome back, everyone. Thanks for coming today to listen to us. Really good to listen to Scott and Jeff about where the future is taking us and the market opportunity. I'm here today to give you a bit more of a look at Cognigy and what we're doing at NICE, Cognigy, and the future that we're seeing as in the Better Together story of NICE and Cognigy together. Now, for those of you who don't know so much about Cognigy, we thought we'd put a quick slide up to give you an overview of who is this company that NICE acquired. We were founded in 2016 in the beautiful city of Düsseldorf in Germany. We have since emerged as a market leader in what used to be called conversational AI and now is called agentic AI for customer experience.
We're processing by now, so this was in 2024, but by now we're processing billions of user interactions every year. We have the highest recommendation rate in the industry, as you can see on Gartner Pure Insights. We're servicing brands all around the world. I think it's around 1,000 brands that we are servicing. We're also praised essentially by any analyst in the market. We're a Gartner Magic Quadrant leader. We're a Forrester leader. We're an IDC leader. We're an ISG leader, etc., etc. We're very proud that we have attained this position as a small startup from Germany growing into this market leader globally. Now, to give you a bit of background where we came from, this is getting a little bit technical, but we emerged in an era of the so-called NLP frameworks.
NLP frameworks back in the day 2015 were coding frameworks that analyzed user inputs and attached a label, a so-called intent to them. A customer would come in and say, "I want a refund for my ticket." The NLP framework would then detect, "Okay, refund request," and then would give you that result. You had to write code to do something with that. Okay, what do I do next? We saw a gap in the market where we saw that large enterprises do not actually want to write code to then determine what to do, but want to have a graphical toolset, what is now called a low-code toolset, to define the process that would happen afterwards. This is what would then be called a conversational AI platform. That is what Cognigy built and emerged as a leader in.
Over time then, of course, with the emergence of large language models, we first enhanced the conversational AI platform with large language model capabilities, such as generating these NLP models with large language models, testing, etc., etc. Last year, we created what is called the AI Agent Orchestration Platform. An LLM-native platform inside of Cognigy that makes use of all the enterprise features that we already had and still incorporated large language models. By the way, any large language model, not one specific model, to enable our enterprises to make use of this amazing technology. The next step for that is NICE Cognigy. That is what we are going to be talking about here today.
The platform is a holistic platform that allows you to design, run, and optimize AI agents at scale for large enterprises with various components, various subproducts like AI copilots to help human agents with agentic-driven experiences, knowledge, process knowledge, etc., in real time during calls and chats. We have a live agent product, which is a digital channel product. We have the AI Agent Platform that we're going to look at more today, Voice Gateway, to connect the Cognigy AI agents to any ACD on the market. This means whether it is CX1, whether it is Genesys, whether it is Avaya, whether it's 8x8. We can work on top of all of those ACDs and CCAS players. We have a very advanced Insights product as well, which is specific for insights for agentic AI for customer experience.
Now, very important, Cognigy is not just an agentic AI product, but it combines agentic AI with logic guardrails and pre-built system integrations that we have built out over the years because hallucinations remain a fundamental challenge in the agentic AI world. By embedding it into a structured workflow, as you will see in Cognigy, we actually get a handle on that. Plus, AI agents must be deeply aligned with the company's logic, data flows, and value creation. Agentic AI can't really do anything unless it is connected to back-end systems, to booking systems like in the Lufthansa case, to CRMs, to ERPs, etc. These AI agents need access to these systems, to the data in these systems, to the APIs in these systems in real time in order to be able to help a customer.
Now, I can tell you a lot about the platform, but we thought you're seeing slides every day. You're seeing videos every day on the internet. Why don't we do something different and just show it to you live? That is what we're going to do now. If we could please switch to the laptop view. There we go. Okay. This is essentially how you set up an AI agent in Cognigy. You can see you can pick an avatar. You can define a name. You can define a description, etc., etc., the style of how this AI agent speaks, and so on. This is what we call the persona of an AI agent. You can have one persona for all the things your AI agent does.
You can have multiple personas where you say, "Okay, let me hand this call over to another AI agent," and this other AI agent might have a different persona or not. What we then do is we assign this AI agent a job. Okay, so this AI agent, Jennifer, has a job here. She's a virtual service expert, and she has a job description and specific instructions and knowledge for this job that she needs to be able to do. We're also giving this AI agent so-called tools. In this case here, an authentication tool to authenticate a customer, a make payment tool, and an update account tool. These are the things that this specific AI agent can do. It is all done in this low-code platform so that enterprises can build these AI agents themselves. Now, Cognigy's AI agents are multimodal AI agents.
That means we can interact with them both in chat and also in voice. This is a chat interaction here in what we call the interaction panel. It could also be on SMS. It could be on WhatsApp. It could be on a website widget. We can just have a chat with it. Here we say, so the agent says, "Hello, Phil. Thank you for reaching out. My name is Jennifer. I'm a virtual service expert with Nexora. How can I help you?" I could say, "I want to make a payment." Then she says, "Got it, Phil. I can help you with that. First, we need to authenticate you. Shall I go and send you the verification PIN?" I can just have this chat with this AI agent.
In Cognigy, you cannot just set up the AI agent like this, and then it's kind of a black box, and you do not really know what's going on. You can actually look under the hood. That is what we see here. We can take AI agents, so in this case, Jennifer, and embed her in a structured workflow. We can do things before data hits the AI agent. These things are required in an enterprise context. You filter out PII data. You make specific settings, etc. We can also see the processes that happen if Jennifer wants to do a specific thing. For example, here we have authenticate. When Jennifer wants to authenticate a customer, she will say something specific, and she will say exactly that. That is important because LLMs, you cannot force them to say something in an exact way.
Here in Cognigy, you can. She will send out an SMS via Twilio and continue with the flow or with the payments. In this case, here we have integrated with Stripe. You can see we can do that in a very low-code way, in a very easy way. Now, as I said, these AI agents work in both chat and voice. What we're going to do now is we're going to call this AI agent, and we're going to use something in Cognigy we call live follow mode. I've now activated the live follow mode for this phone number that I'm going to use, and we can see what is happening in real time. Let's kick that off.
Hello, Phil. Thank you for contacting us. My name is Jennifer. I'm a virtual service expert here at Nexora. How can I assist you today?
Yeah, I think I have an outstanding payment on my account. Could you check that, please?
I can help with that, Phil. Before I share any payment details, I'll need to authenticate your account for security. Would you like me to go ahead and send you a temporary PIN via text now?
Yeah, please send it to this phone number now.
All right, Phil. I'll send the PIN to the number ending in 5008 that we have on file. Let me trigger that for you now. I am now sending you a one-time PIN. Please repeat it for authentication purposes.
Yeah, so that's 1167.
Perfect, Phil. You're all authenticated now. I do see an outstanding balance of $45 on your account. Would you like to pay that today using your credit card ending in 0115, or would you prefer a different payment method?
No, that credit card is perfect. Let's use that one.
Got it, Phil. And just to confirm, you're authorizing us to charge $45 to your credit card ending in 0115 today, correct?
That is correct.
Please give me a second while I am triggering your payment. All right, Phil. Your payment of $45 has been successfully processed to your credit card ending in 0115.
All right. What we saw here, we saw a real-time interaction with one of Cognigy's AI agents. This is not a pre-rendered video. This is exactly the type of experience that our customers can create with Cognigy. You could see you can use this low-code interface for debugging purposes, but this voice experience, the very natural voice, the very fast responses, whilst interacting with back-end systems, is not fake. This is reality. This is what we have deployed at scale for hundreds or thousands of brands out there already. We are going to have Lufthansa talk about that in a second. If we could please switch back to the slides. Great. Now, what we saw was the Cognigy interface, right? The Cognigy AI Studio that we can use, that our customers use, that enterprises use to build out these types of experiences.
What I also want to share with you is another experience, which is a front-end experience. This is the type of experience that can be created with the Cognigy platform. Earlier in the demo, we saw a chat, and we saw a phone call that was made using a regular phone through the phone network. It was connected to a CCAS and then against the Cognigy AI agents. Now, what we're seeing here is the next level of experience. It is a proactive outreach by an AI agent that knows that this customer has to renew their mobile phone plan. This customer is then sent a widget, and using this widget can then, using what is called WebRTC, communicate with the AI agent in real time. Here, the widget comes in via SMS. They click the URL. They go in, and then a voice conversation starts afterwards.
Let's take a look at that.
Hey, Ty. Thanks for taking the time. I'll keep it short. I analyzed your data consumption and noticed that you spent more time abroad this year, so it might make sense to consider our additional travel bundle. The rest of your plan seems to fit well, and I'm able to make you the following offer. You'd pay $95 per month instead of $120 for the same plan, including a discounted EU travel bundle. How does that sound?
Oh, that actually looks pretty good. Can I also pick a new phone?
Here you go. You can browse through all available phones and choose the one you like. Which model catches your eye?
Actually, I like the iPhone 17 Pro in blue.
Here it is, the iPhone 17 Pro in blue. Would you like to add it to your new plan offer?
What is the difference between the 17 Pro and the Pro Max?
Here you can compare them side by side. The iPhone 17 Pro has a 6.3-inch display and up to 33 hours of video playback, while the Pro Max has a larger 6.9-inch display and up to 39 hours of video playback. The best battery life ever in an iPhone. Which one do you prefer?
Actually, yeah, let's go for the 17 Pro in blue. I think it's big enough.
Perfect. Your iPhone 17 Pro in blue with the EU travel bundle is ready. Please review the terms and conditions, then sign to finalize your new 24-month contract. Would you like me to guide you through the signing process?
No, I think I'll be good. Just a second.
All right. Take your time. Let me know once you've signed so we can proceed.
Okay, there we go.
Excellent. Your new iPhone is on its way and should arrive on Friday. Is there anything else I can help you with today?
No, this has been great. Thank you so much.
You're very welcome, Ty. I'm glad I could help. Have a wonderful day.
I know everyone wants to erupt in cheers because this is exactly the type of customer experience that we would all like to have, right? Because it really is. What we are seeing here is not the future. This is reality right now. This is not a video that was produced by our marketing team in some video cutting. This is reality. This is a straight-off recording of a phone. Our customers can and are producing those types of experiences at this point in time. There is something very, very important that I want to share with you. Everyone can do a great demo these days. We have LLM technologies. We have speech-to-text technologies and text-to-speech technologies like ElevenLabs. We have the OpenAI LLMs. Everyone who can code a little bit of JavaScript can put a cool demo together and blow you away with the demo.
For the types of customers we are dealing with, enterprise customers, this is not near enough. There is a lot of stuff happening under the waterline that is even more important. Without these kinds of things, our customers could not even think about going live. There is omnichannel routing. We are supporting more than 30 different channels in the platform, whether it is voice channels through the contact center, whether it is voice channels on WebRTC, whether it is web chat, whether it is WhatsApp, whether it is WeChat, whether it is Line Messenger for the Asian markets or Kakao Messenger in Korea. We are supporting a large number of channels. You do not have to build this experience for each channel. You build it once, and then you deploy it on a channel, and it immediately works. You saw that earlier, proof in point.
I chatted with the AI agent, and then I called the AI agent, and the experience was equal. Plus, there is agentic orchestration. We have the tool usage, the planning, and reasoning baked directly into the platform. We did not tell the AI agent, "Okay, now you use the authenticate tool. Now you do this," right? We just provide the tools, and the reasoning capabilities of the LLM take care of the rest. We can build end-to-end workflows where we have a visual flow builder that you saw. We can have guardrails where we can make sure that the AIs actually stay on track. We can have fallbacks because sometimes LLM APIs are flaky, right? Maybe sometimes the response time of a certain API goes up, and then the LLMs do not respond in 200 milliseconds anymore, but they are responding in three seconds. That is too long for a voice interaction.
We measure these kinds of things in Cognigy and can then fall back to another LLM that is faster to maintain a very good customer experience. We have more than 100 different system integrations with CRM systems, ERP systems, etc., so that our customers don't have to build them, plus the contact center integrations. Cognigy is integrated into the NICE CX1 platform and will be even tighter integrated in the near future. We also remain a standalone platform, and I can't stress that enough. A large number of our customers at NICE Cognigy are using a very wide variety of CCAS systems, right? Be it a Genesys, be it an Avaya, we work just as nicely based on those contact centers through our integrations, both on chat and on voice. We have the handover to live agent.
We did not see that in the demos here, but the AI agents, if they get stuck, can also say, "Hey, I think I want to loop a human in," either by communicating with a human in the background and then providing this information to the customer, or by handing the full conversation to a human agent because human agents are not going away. When it is handed over to a human agent, we can use embedded agent assist. Then we have all these boring things that are still extremely important for large enterprises: versioning, auditing, who did what in the platform, role-based access control, data residency controls with sovereign clouds around the world, GDPR, observability, etc., etc. It is much more important to have a holistic platform rather than just a prompt wrapper that allows you to show a nice demo.
In an enterprise scenario, you will not be able to go live with that. To summarize all of that, how do we differentiate as Cognigy? We allow for hybrid AI agents. That means AI agents that are both agentic and based on traditional conversational AI, so more deterministic, especially important in regulated industries. We have AI operations and orchestration, the fallbacks, the observability, the orchestration of different AI models, native contact center connectivity to pretty much any contact center out there, multimodal experiences like we saw in the second demo. We can have chat experience. We can have voice experiences. We can have voice experience with multimodal widgets and graphical interfaces, which we believe, and by the way, many of the analysts believe, to be the future of customer service and customer experience interaction.
Of course, all the enterprise readiness with our certifications, security compliance, data privacy compliance, etc. That really led to us being loved by many of the largest brands in the world. If you look at Gartner PI Insights, we have the highest ratings in the world and the most ratings in the world compared to the likes of Core AI, Emilia, Sierra, and many others. We picked out three quotes here that make me, as the founder of Cognigy, specifically proud. "While others do the slides, Cognigy does the work, surpassing expectations and agentic AI boost teams with easy use." The other thing that is important when you want to make a customer successful, on the one hand, you have the product that needs to be outstanding, but on the other side, you have the company, Cognigy, that makes you successful, right?
This is something that we also, I think, share, envision with NICE. We do not just provide the technology. We provide the experience and the knowledge in the industry that makes our customers successful. Our secret, or one of our secrets to the success of our customers, is our customer obsession. We work very closely with our customers. In the end, we are delivering real impact. We could have an amazing product. We could have amazing customer obsession. If we do not deliver impact, it is worth nothing. We are delivering real impact. Call containment, depending on the use case, between 60-90%. Call containment means that the call does not have to reach a human agent anymore afterwards.
If it reaches a human agent, average handling time reduction of 25% or more through the AI capabilities, first that come before the interaction with the human, and then how we are supporting the human agent during the conversation with the customer. It is truly a win-win-win situation. CSAT improvement. Customers love it. Nick is going to talk about that in a second. Customers like being helped by the AI agents because they're being helped quicker. They do not have to wait on the phone lines to get done what they want to get done. We are seeing also a CSAT improvement here. I could tell you about this all day long, but you might think, "Yeah, of course, Phil is a fan of Cognigy," and I am. There are also other fans of Cognigy.
I thought I would like to bring on stage here Nick Algeier from Lufthansa. Please, everyone, give a hand to Nick. He will tell you firsthand about his experience with Cognigy. Thank you.
Good morning, everyone, and thanks for having me. My name is Nick. I've been with Lufthansa Group since 2014, and my teams and I, we're in charge of the conversational AI developments for our B2C space. If you ask me, one of the most exciting spaces to be around. Today, I would like to show three things. First, I would like to show how we are leveraging conversational AI to address real operational challenges at scale. Next, I would like to show how conversational AI has become really a mission-critical asset for us. Third, I would like to point out how our partnership with now NICE Cognigy is of strategic importance to us as we move into the next era of customer interaction. First, let me fill you in on who we actually are at Lufthansa Group.
We're around about 100,000 employees, and our mission is to connect people, cultures, and economies. We do that around about 3,000 times a day on one of our flights. We bring our guests to more than 300 destinations across 12 airline brands with more than 700 planes. All of that summed up in a year brings us to 130 million passengers, bringing us revenue of around about EUR 38 billion. That scale is great until it's not. Why is that? Our business is not a linear business. Our business is pretty peak-heavy. Let me take you on a little trip with a bit of a twist of a perspective because many people might not have looked at this challenge from that angle. It's July, beautiful summer day. You're at the airport, mid-90s degrees outside. The airport is full of people.
You see that couple about to go on their honeymoon cruise. You see that family about to depart on their beach holiday. You see the grandparents excited to meet their grandchildren. At the back of the skies, you see dark clouds. You can hear roaring thunder coming in. You see the first lightning. There is a thunderstorm right above the airport. What is happening? Unfortunately, due to the thunderstorm, the first flight is not about to go out. The flight from Frankfurt to Paris in this example can leave. What does that mean in turn? If the flight to Paris is not leaving, it can also not bring passengers back from Paris to Frankfurt. Unfortunately, the thunderstorm is sticking around for longer than expected. Also, your third flight and fourth flight of the day leaving for Zurich is delayed.
Unfortunately, due to the delays throughout the day, also the fifth and the sixth flight, they're not going to make it on time. Lots of passengers will have to contact us that day. That is not because they love to do so, but because they will have to. If you run the numbers here real quick in this simplified example, we're talking 500,000 passengers in a day easily. For our passengers, for our guests, this is a massive inconvenience, that situation. For us, that means suddenly 500,000 of our passengers need answers. They need support. That support can look very different. Some people might just need a new flight. Others might want to get their money back because there is no point in taking that trip anymore. Others might have missed their connecting flights, so they need a hotel. They need food.
They need compensation. It is a wide range of customer intents and requests that we will have to serve on that day. Let me show you such peaks. No matter how big the human workforce, there is no chance to handle that peak, that spike in demand reliably at scale. This was a summer day. Now imagine a full pandemic is about to hit you. We are not talking one airport closed for a day, but we are talking a country shutting down. We are talking an entire fleet of 700 planes on the ground for days and weeks, and your passengers reaching out to you. That is exactly the situation that we were in. That was the reality that we were facing, and that made us start looking into conversational AI before it became so fashionable.
We realized if the future for us continued to be so peak-driven, our customer support model, how it was, no longer had a real future. We saw this would not be a simple task, but we realized there are four key areas that we had to focus on. The first one is we needed to automate whatever we could. We had to reduce the manual contact volume. Second, for what's left, we have to prioritize. If your flight is leaving in an hour, you should be first in line because you have no time to wait. If you're contacting us for a seat reservation for your next trip two weeks out, you'll probably be okay waiting for a couple of minutes. Next, what we still need to handle manually, we have to look at efficiency drivers here and improve the time spent on each case.
Lastly, we want to gather insights into the first three areas and see what is working, what's not working, and how can we get a continuous improvement cycle going. We saw two things here. The first is there would be a wide range of use cases that we'd have to look at, starting from simple FAQs ranging to complex end-to-end rebookings, deeply integrated into our legacy systems, our booking system, and other legacy. Also, the more we lean into automation, the more important the scale would become. If you start implementing automation on several channels, that wave of contacts is going to hit you on all the channels at once. Scale really becomes important. The question for us was now, who would be the right partner to pick in that endeavor?
We sat down with our experts and were imagining what platform would work best for us. We came up with all sorts of requirements. The first attempt, 150 requirements. Then we cut it down to a short list of 30 requirements. I would like to point out five in particular that we then decided to focus on. First, we wanted to have an agnostic modular platform. We wanted to have the opportunity to replace certain components with others if we saw the need for it. We wanted to have a platform that not just our technical experts, our developers could use, but where also the business experts could contribute to creating the experience. As you saw earlier, we needed a platform that could handle the scale required. We wanted something where we could create the experience once and then implement it for all of our airlines.
Otherwise, we'd have to create one bot for each airline in each language. If you're serving four or five airlines in six different languages, you could imagine the work that would go into creating each of the experiences separately. Lastly, as I pointed out, we wanted to get analytical insights into what works, what does not work, and what needs improvement. We screened the market. We started with hyperscalers, and we saw that would not work for us. Very technical solutions is what we saw, but we realized we need to keep looking. The second area was CCAS providers. I believe it is fair to say, had we found something that worked for us, I would not be standing here today. The third area was specialized conversational AI vendors. There we were onto something. We did an RFP.
We looked around, and then out of that RFP, Cognigy came out as a winner. The first reason was Cognigy provided a platform that matched best with our requirements. But then we also saw that the team of Cognigy, they understood what challenge we were facing. They did not just understand the challenge, but together we've been working on that challenge, and they have supported us in tackling that challenge since then. Say this was the world before agentic AI. If you were to look into platforms now, you might have different arguments. I would like to have a look at the differences of the world before agentic AI and agentic AI and point out what I would still say is important to look at. First, how do these systems work? In a deterministic conversational AI system, you tell the system what to do.
In an agentic AI system, you tell the system what to do, but also you have to cater for what not to do. The problem here is there is an infinite number of options for the what not to do. The decision-making on the left-hand side is pretty clear. It's if, then else. On the other side, an agentic system has much more autonomous decision-making power and ways to decide, leading to outcomes. On the one hand, deterministic, so consistent and reliable. On the other hand, probabilistic outcomes. The decision-making was autonomous, but the outcomes can be inconsistent. The flip side of it is a deterministic system can feel rather clunky. It can feel rigid and scripted, while in an agentic system, the conversation can flow much more naturally. It can feel much more flexible and smooth.
If we were to look at platform selection again, I would say to those 30 criteria, I would add two more as very important. The first one is, as a business, I would like to have full transparency over what is happening underneath the hood. I want to be in a position to control where I want to use what. If this process is important to me, I want to have consistent outcomes. Three points that I would like to point out here. The first one is, if you ask me, is, sorry, I cannot answer that. Can be better than a rogue AI agent. Sorry, I cannot answer that is very different to sorry, I do not understand. I think the times of sorry, I do not understand, they are over. They are no longer accepted by the customer.
I can understand what you mean, but for whatever reason, decide that I do not want my AI system to answer that. That could be because it is a very sensitive topic or because I am not confident enough that the AI system has all the relevant information at hand to autonomously answer that question. Think back to that couple about to go on their honeymoon cruise. If their flight was canceled and they missed their cruise and now they ask for a refund, imagine how they would feel if we hallucinated their refund, gave them the wrong amount of money, and gave them the wrong timing of when they would receive the refund. I am not sure they would fly with us again if this is something that we mess up big time. For us, it is important to be able to blend these two things.
We would like to have the control over certain things and have a deterministic system where needed in place, but complement that with an agentic AI experience where it's meaningful. Fast forward to 2025. Where do we stand now? Actually, conversational has become the most popular channel throughout Lufthansa Group. Those experiences are the number one contact channel for us now. We're integrated in chat, voice, and agent assist. We're making use of generative AI in more than 50% of our conversations. Actually, in voice, we're beyond 90% of conversational AI use, and we're in the process of rolling out more and more agentic AI use cases. This year, we will probably end up with around about 12 million conversations being facilitated through the Cognigy platform.
We have had record days, like the ones that I pointed out earlier, where we had around about 400,000 sessions in a day and more than 10,000 concurrent conversations with our customers. Now we see that millions of customer contacts get prioritized. It is now a reality that if you call us and we can identify you and your flight is leaving in an hour from now, you will be first in line, and we will tell you that you're first in line because your flight is leaving soon. We see shortened AHT across both chat and call, and we're saving a good amount of money every single year. Where do we go from here? What we want to create is Lufthansa Group's digital go-to person.
I think it goes along to what Scott is saying, that companies are now not looking at ways to reduce customer contacts, but we want to find better ways to serve our customers. We want to embed our conversational experiences across different points of the customer journey. Overall, we want to create the best conversational experiences in the tourism industry. For us, that means practically we will embed our conversational experiences in more channels. You'll be able to find us hopefully soon, and also in messaging apps, more touchpoints on our website, wherever you're going to call us, etc. I also strongly believe that channels will blend. Phil also showed that earlier. I believe as multimodal experiences are starting to mature now, customers will think less in this is chat, this is voice, this is call.
We will see those experiences blend, and customers will want to interact however natural it now feels in that very moment. That can mean you might call us for a seat reservation, but why not select this seat on your phone? Because whenever you have seen a seat map, it is much easier to pick where you want to sit than having an agent walk you through your options. Think talking to the website and telling the website what you need to find out, and then the UI will serve you in a dynamic fashion. Lastly, this is not where our collaboration with NICE Cognigy will stop. Aviation will always face volatility and unpredictability. Dealing with that volatility and managing that volatility, this is really where customer value is created and where competitive advantage lies. For us, AI is more than nice to have.
How we handle volatility creates true customer value. AI enables us to manage unpredictability, to scale reliably, and to improve customer experience in a time where expectations have never been higher. With our partners like NICE Cognigy, we're not just waiting for this to happen, but we're actively creating the next generation of our digital customer experience, one that weathers storms even in the most literal sense. Thank you very much. Back to Phil.
Thank you. Thank you, Nick, for this amazing presentation. I think what this presentation really shows is that what we can create with our type of technology are true win-win-win situations. Now, what do I mean by that? Because usually it's win-win, but here it's a triple win. Firstly, the enterprises deploying our technology are winning because they are saving, as we could see in this example, multi-millions of dollars every year whilst creating superior customer experiences. That is the second win. The customers of our customers are also experiencing the benefits at scale because if such a storm hits and 500,000 people have to call the contact center, they're going to be waiting for hours on the phone lines before anyone can help them unless they have deployed AI agents.
The third party to win is us because we are the software provider and we are generating revenues in that way. I hope what you can also take away from this experience, this is not just about putting an IVR phone bot in. Yeah, we can do that too, and we are doing that. It's not just about putting a chatbot on the website. It is really about revolutionizing the customer experiences that can be created and not limiting the customers. With Cognigy, if a new LLM comes out tomorrow, you can embed that in your customer experience flow. It's an orchestration platform that allows you to be open. New voice models are coming out. New speech detection models are coming out, etc., etc. You can use all of that. You can use different models in different markets around the world.
What all of this means is that our customers can grow together with us. They can power all of their customer experience, whether it be customer service, whether it be marketing or sales, in one platform, and that platform is NICE Cognigy. We have spoken a lot about Cognigy and what we have done in the past and the position to which we have gotten. Of course, the really big news is that NICE has acquired us because otherwise we would not be standing here on stage today. The analysts around the world, industry analysts, really welcomed this. Many of those had said that an acquisition like this would happen, and some of them called it the biggest news in the CX industry of 2025. An industry leader as NICE coming together with an agentic AI and conversational AI leader as Cognigy, it is really the perfect union.
Now, one thing to highlight again, and I know I've mentioned a couple of times, whilst we will be integrated, and I'm going to be speaking about that a little bit more in a second, we'll also remain available as a standalone platform. Now, bringing those two together, what is that better together story? Why is it so unique? Why does it bring such tremendous value? On the one hand, we have Cognigy, which is agentic AI and orchestration of AI agents at scale. We bring a world-leading platform with all the systems integrations, etc. What does it mean financially? What does it mean GTM-wise? It means that now NICE cannot just sell AI solutions to their existing customer base, but the whole CX market. We have a standalone solution that can be adopted by anyone in the market, no matter whether they're using CX1 or not.
On the other hand, we have NICE, which is, of course, vastly larger than Cognigy, has a much bigger scale, a global GTM team in place all around the world. I have already visited our teams in Singapore, in Australia, all around Europe. Massive scale compared to what Cognigy has, 27,000 customers that NICE Cognigy can now be sold into. A massive upsell opportunity on that end. Those are the GTM portions. In addition, there is this amazing fit that is the product fit. Again, we have Cognigy with the AI agents. You can build, deploy, operate, and optimize. Then we have the CXone Mpower platform with all of their capabilities. We worked on this slide a lot, and we tried to do it justice of how amazing this integration is, but we could not really because it is much bigger than this.
If we wanted to show you how good this integration really is, we would need a slide that is 10 times the size. What do I mean by that? At Cognigy, we had all these plans when we were still standalone. What are we going to build in the future? We knew it would be about orchestrating customer experiences across the lifetime of a customer, not just singular interactions. It would be about not just handling inbound, but also outbound so that we can expand into the sales arena. It would be about intent mining, taking customer call records or transcripts and mining them for what customers actually want, and then building AI agents based on that. It would be about analyzing transcripts after the fact, so really in-depth agentic analytics.
When NICE approached us to acquire us, we saw that all of that was already there. The world's leading outbound dialer, the world's leading intent mining, all of these components are already in the CXone Mpower platform, and we are integrating with those now in Cognigy. We are going to create the leading CX AI platform on the market, and we are very well progressed in that journey already, and we can't wait to show you what's going to happen here over the next couple of years. I really believe that with those two companies coming together, we possess a strength that no one else in the industry has, no matter how big the competitor is.
All of them have some components of this, but really bringing it all together in one is extremely unique and is extremely exciting for us here at NICE and NICE Cognigy specifically. Now, I would like to introduce you to one of my colleagues here at NICE, Neeraj, who is going to show you a little bit about that future, about the progress that we have already made in integrating those two platforms. I could not be more excited than to hand over to Neeraj now. Thank you very much.
Thanks, Phil. I was going to use my Starbucks name. I like Nick better than Neeraj. No one pronounces it the same. Okay. Let's talk about what it takes to achieve consumer intent resolution. When we talk about intent, it's a really interesting story, and this is going to get deep and technical. This is what the consumer expects. When they call a business in, they expect they have a problem, they have a need. They expect that they're talking to a front office agent or a chatbot, and their problems are resolved in real time. The reality is slightly different. The reality is that it's a, we call this intent resolution journey. This journey goes across, and Scott's mentioned it, it goes across front, back, and mid-office workers. It's really disparate labor forces that exist in most organizations.
When you think about these disparate labor forces, oftentimes you've got back office workers that work completely different hours than front office agents that are 24/7. You get this asynchrony in resolution. The customer calls into the front office, they handle all the conversations, while someone in the back office might be fulfilling a request to actually complete the customer's need, which creates, once again, this necessity for a case management system to manage this asynchronous behavior. It creates a really frustrating experience. Everybody's called in, and they've, I don't know, you call in a business about a problem, and they give you a case number, and maybe they'll follow up with you later. This is the reality of intent resolution in a lot of organizations today. When you think about the unique advantage that NICE has, we are the customer engagement platform.
That does not mean we just sit at the intent's generation. We are not sitting at when the initiation of the intent happens. We sit across from the initiation of the intent all the way down to the fulfillment and resolution of that intent across all of these disparate workers. It is really important. We always talk about this concept of our data layer and this data lake that we are, all these conversations we are mining, all these screen recordings we are mining. The reality and the moat that we have created is we have a really, really interesting operational view of how an intent is resolved in every organization in the entire world. We understand exactly what each of these workers are doing, how they are doing it, what applications they are using, what tools they are using, and how long it takes. It is really, really important. It creates a really big moat for us.
AI agents are, you can, just like Phil said, you can create an AI agent for anything using OpenAI. The reality of AI agents that actually solve customer service problems is super complex, and our data allows us to get there really fast. I just talked about us gathering data points. Now we know what every single worker across the entire enterprise is doing to resolve a consumer problem. We're gathering tasks. We're gathering what tools they're using. We know the channels, and we know the integrations. What does that look like to you? What is an AI agent? An AI agent is just a description of a job. What job are they performing? What tasks are they doing? What tools are they going to be using to complete those tasks? We have all of this data on the platform.
Now we can translate the human workers across not only the front office. This is what everybody thinks about when they think about automation, automating the front office, customer chatbots. The reality is that we can expand our TAM and market all the way down into the full intent resolution journey. It's really, really important. Automating back office, knowing what those workers are doing, is going to be the future of customer service. Now, what happens when you do that? The reality is that as a customer engagement platform, we operate on top of all ecosystems. Everybody thinks about CRMs or case management systems as being the customer management platform, but really, intent resolution sits across all of those ecosystems and platforms. As a customer engagement platform, we understand the entire journey.
Once we add AI agents into this journey using Cognigy, it creates this intent resolution journey that's in real time. It's hyper-personalized. It allows us as NICE to really expand what we're doing today in the front office all the way down to journey resolution. We will see a quick demo. Let's talk about what this demo is going to be. You're going to see a consumer. We're going to join a call. We're going to join a conversation with a consumer and a business. This consumer is chatting with an agent, a front office agent, about a credit card dispute. They've seen something on their bill, and they're going to chat to an agent. You're going to see this customer is not, or this prospect's not on CX1 today. You're going to see what their agents are doing today to handle these types of conversations.
You could see, as the conversation goes on, this human agent's got SOP pulled up. They've got billing systems. They've got CRM systems. It's a pretty complex process. In general, agents have something like 8-12 applications open at any given time. It's really, really complex. This is the world today. Everybody thinks it's super streamlined. You have one application to handle everything, CRM. No. The reality of customer service is this. It's a really, really complex process. Now, if you looked at the SOP of this particular agent, the procedures that they're using to solve this consumer intent, their entire job is to talk to the consumer, take a couple of notes, create a case, and then say, "We'll follow up with you later." The conversation ends right here, and it's super frustrating, non-real-time resolution for the customer. Okay.
Now, several days later, that same case goes to a back office worker. This is really important. A back office worker has got a case management, ticket management system up. They've got a lot of tasks. Some are CX related. Some are not. This disparity in what they're doing is a big problem. They're going up, opening the ticket. They're helping us resolve it. Done. Now, if you really think about this journey, and by the way, when we think about the moat that we have today, just last week, I had my team pull up 1.8 million of these journeys across the top 10 banks in the U.S. This journey, on average today, across those banks takes three to five days to resolve. It's a big problem. This is not fake data. This is not made up. We understand these journeys really well.
It's a three to five-day journey across the front all the way down to the back office. Now, let's talk about how NICE solves this problem. We're going to join a CX orchestrator. Their entire job is to look at customer journeys, solutions to consumer problems from a journey perspective. This orchestrator is an AI-powered tool that's looking at the conversation that you just had and thousands of other conversations, putting together these disparate pieces in what we call, once again, the intent resolution journeys. Of course, there's a lot of front office journeys that we're already automating. Things like FAQs and, just like Scott said, password resets. There's a ton of these. Easy to create agents. We're already doing that today. A lot of journeys that are, they cost a lot of money. They cost a lot of frustration from a customer perspective.
Are journeys that go across the front, back, and mid-office. Now, this orchestrator, as he's analyzing these conversations, it's using AI agents to do analysis on screen recordings and SOPs to see what tools and integrations our agents and our workers are using to resolve this issue. I can see that the dispute charge is being highlighted by the orchestrator. Let's drill into the dispute charge. This is really interesting. This is the journey that we just noticed, the really disparate journey. You've got a tier one agent that's using all these tools, and then you've got a case being created by that agent, and then you've got back office workers executing those cases. The reality is all the data that we've gathered allows us to orchestrate and put this journey together, put together a picture of what's happening in the world today.
If you notice on the right side, we've got very specific descriptions of exactly what these agents and workers are doing to resolve the problem. We've been able to take conversations and screen recordings and tasks and integrations, convert them into descriptions of what an AI agent should be doing, which allows us to really quickly take the next step. It's recommending that, hey, I can analyze these screen recordings. I can analyze these SOPs that are already within CX1 and expert and empower agents in all of our call recordings. It's going to recommend something really interesting. You can see that it's taking this really synchronous journey across all of these workers and recommending that we create AI agents to replace the tasks that these workers are doing. It's really, really interesting. We have the descriptions.
We understand what these agents are now doing in the front office and the workers are doing in the back office. We can use those tasks and descriptions and data that's coming from the front office and the back office to create AI agents to replace the humans and the human-driven tasks in the front office and the back office with AI agents created in Cognigy. This is really a data-driven AI approach. Creating AI agents is, just like I've said, it's super easy. Type in a couple of descriptions. The reality of creating AI agents to solve customer service is really complex. It's this journey. Being able to translate that into AI agents that drive value across the entire NICE platform is really difficult. Now, I'll highlight and hover over a couple of these agents, and you'll see the tools kind of blink through.
This is what's happening. These agents are using these tools, and they're assigned these tools, and they're working together using the agent-to-agent platform. Now, let's click on SAM here. As soon as I click on SAM, it's recommending that I create a dispute intake agent. And this is directly within the Cognigy platform. I've taken the description, the system-generated description based on all the recordings and all the conversations and screen recordings we have from the front office. We've created this agent automatically within Cognigy using the right tools that are already available in the enterprise platform. The reality is that having this enterprise platform that we have within CXone Mpower, we have every integration that you can imagine.
As the human agent is opening up applications, we can monitor what they're doing, translate that to what an AI agent should be doing, and create the AI agent in Cognigy. This is the power of the platform combined with Cognigy. Now, let's go ahead and deploy these agents. This is how easy deploying an agent on the CX1 platform is. We understand the channels already. We understand what the journey looks like. Now we're going to deploy this AI agent across our platform. You're going to see a conversation. Let's talk about the original disjointed asynchronous conversation we saw that was super frustrating for our customers. Let's see that being replicated within the platform with AI agents.
You'll see the NICE reasoning capability of our agents right here and how our primary AI agent, SAM, is utilizing other AI agents across the enterprise in order to complete his tasks.
Hi, Stephen. Thanks for calling OneBank Customer Care. My name is Sam, your personalized OneBank AI agent. I see here that a recent notification went out to you regarding a potentially fraudulent charge on your credit card. Is that what you're calling about today?
Yeah. I just saw the notification. I didn't approve that charge. What do I need to do on my side?
The first step is done. I've initiated a dispute case for that charge. Have you had a chance to look at your other recent transactions to confirm that there are no other fraudulent charges?
I have. This is the only fraudulent charge that I see so far.
Excellent. To prevent any further unauthorized activity, I'm going to cancel your card now and order you a replacement card. Your new card will be sent overnight.
Sounds good. Thank you.
You're welcome. Now, let's make sure your automatic payments aren't interrupted. I'm just texting you a list of those charges. Can you please mark the automatic payments that you'd like transferred to your new card? Excellent. I've initiated the transfer for those recurring payments. They should update automatically over the next 24 hours. I've already confirmed with the merchant that they were unable to verify the authorization of the charge. A refund in the full amount of $286.58 has been issued to your account. I just sent you a text confirming what we discussed today. There's no further action needed from you. Is there anything else I can help you with today?
No, that covers it. Thank you so much.
My pleasure, Stephen. Thank you for being a OneBank customer.
Wow, what a great experience. The reality of the world is orchestrating these types of experiences takes an enterprise platform, and creating these types of agents is not a single prompt. It is a data-driven approach that takes the entire customer intent resolution journey, converts it to AI agents, and turns three to five days into one to two minutes. That is the reality of the world that we are living in today. CX1 is in the ideal position. You cannot expect case management and CRM systems to understand these journeys. They do not lie at the heart of the initiation of the intent all the way down to the resolution. It is really, really important. Having this data is what gives NICE the advantage. Now, we have been managing human-to-human interactions for 40 years. We understand how human agents work, operate, how they resolve issues, how they are optimized. I mean, it is really, really interesting.
We've got 40 years of data on this. Converting that and using that expertise to manage AI agents is just an obvious next step. I mean, it's really, really obvious. Here you see a unified agent management platform that manages intent resolution AI agents and humans working together. You've got AI agents that are reaching out to humans for help, and you've got entire fleets of agents that are not generic out-of-the-box agents. They're working together to resolve intent journeys all the way across front to back office. You can see how many operational dollars you're saving, how many agents you have deployed. It's really, really important. Some of the reality of having all this data is optimizing AI agents and testing and evaluating AI agents is really difficult. It's actually an open research topic. Are you going to use goal-based evaluations?
Are you going to write all the goals yourselves? Having this 40 years of data on what great conversations look like, what resolution looks like, allows us to create AI agents that actually resolve customer issues and not have to create them manually. I think in this sort of near-term future, you're going to see there's going to be a lot of adoption in AI agents. Creating those AI agents is difficult, and being able to automatically create some gives us a huge head start. Once again, the unmatched domain expertise that we have in resolving intents is really the key here. You cannot expect other organizations that do not live and breathe customer intent resolution to understand how AI agents need to solve problems. You see this across the market today. Our platform is purpose-built. We have tens of thousands of integrations across the platform.
Hundreds of millions of journeys flow across the platform every month. We understand these systems really, really well. We have thousands of AI models that understand what good conversations look like and what resolutions look like. It is unmatched expertise. Our sort of transformative approach, going from conversations and journeys to creating AI agents to solve those journeys automatically, is truly unique in the market. With that, I'd like to introduce Beth. Everybody knows her.
Thank you, Neeraj. And just in case everyone does not know me, I am Beth Gaspich. I am the CFO at NICE, and I am pleased to be here with all of you today. I am going to close out our prepared remarks, and I want to assure you as well that I am the last speaker before you and lunch in case you are hungry. I hope you have heard throughout all of our speakers and presenters today just how excited we are about the future of NICE with our CX AI platform and what it means in the next wave of transformation in the CX and customer experience era. Every time I see the demos from Phil and Neeraj earlier, it reminds me of just how I am increasingly impressed with the opportunity we have in front of us.
It's important to highlight that as real technology, NICE Cognigy is deployed in large enterprise customers, global marquee brands like Lufthansa Group that you heard about from Nick. It's really exciting. What I want to do is start back with where Scott began and started earlier today talking about the massive opportunity that exists in the AI market and how we are going to capitalize on that with our leadership and the strength of our assets. After that, you've heard from our product leaders, starting with Jeff. You heard from Neil. You heard from Neeraj. You also heard from a customer that our innovation together with Cognigy is perfectly positioning us to win in the CX AI era.
Now, in the next several slides, what I want to do is share with you how we're also perfectly positioned to win from a financial perspective and capitalize on this massive opportunity. I want to start with talking about building on our profitable foundation. NICE has a proven track record of strong financial performance. You can see that since 2020, our revenue has increased at a compounded growth rate of 13%, and we exceeded $2.7 billion in total revenue last year in 2024. If you look on our profitability and our operating income, you'll see that we're growing even faster at a 16% compounded growth rate over that same period. Importantly, to highlight the great track record of our operating margins throughout that time, exceeding a 31% operating margin last year.
This healthy financial profile really demonstrates the great best-in-class operating leverage we have at NICE, but as well as a testament to the scalability of our CX1 cloud platform and the ongoing financial strength and discipline we have at NICE. Furthering looking at the strength of our financial foundation and the health we have at NICE free cash flow, it's one of our greatest strengths. You can see over the same five period that I talked about previously, we had a 15% growth annually in our free cash flow, and we generated more than $700 million in 2024. Throughout those five years, we have consistently shown and delivered on very healthy, best-in-class, very strong cash generation from our operations.
Of course, it's important to highlight that, of course, during this time, this has allowed us to make strong, bold, strategic investments like the one that we just did of Cognigy. In the third quarter, we spent about almost $1 billion acquiring Cognigy. In this same quarter, we continued the strength of our buyback program, and in parallel, we paid $460 million of debt, ending the quarter with no debt on our balance sheet. We did all of this through this strong free cash flow generation and exited the quarter still with nearly half of $1 billion in cash. Our strong cash flow generation really sets us up well to step into and capitalize on this next opportunity in front of us, and really it demonstrates both the financial resilience and agility we have in our business.
Next, I want to share with you a little bit about how year to date we have consistently delivered on our revenue and profitability targets. You'll see that through the first three quarters of this year, in total revenue, cloud revenue, and EPS, we have exceeded the midpoint of our guidance consistently across all three, each and every quarter. The other important area to highlight here is the stabilization of our cloud revenue. You see that 12%+ growth the first two quarters, of course, coming, excluding Cognigy, and achieving that 12% in Q3, also excluding Cognigy, with a 13% growth overall in the third quarter, inclusive of Cognigy and our cloud. This growth that we're seeing and the impressive results that we're putting on the board here are really being driven by the success we're seeing in AI.
Our capital allocation approach as well uniquely sets us apart at NICE. I talked about the significant amount of free cash flow that we generate and the strength of our cash and our balance sheet altogether. It's important that we always continue to deploy our capital in a disciplined way. We've done that historically, and that is our plan looking forward as well. We're going to continue to propel the business forward, and we're also planning to deliver significant shareholder value. Our approach is really three-pronged. It's focused on these three different pillars, and I want to talk a little bit about each one. The first is strategic and disciplined investment, and we prioritize investing organically. The areas that we're focused on and will talk more about organically are around, of course, driving our product innovation, both fueling Cognigy as well as our other AI product roadmap.
The second is around increasing go-to-market efficiency, which, again, I'll talk more around the strategic partnerships and how that's positively impacting that growth and expansion, and of course, AI and cloud delivery. All of our organic investment is also allowing us, with that free cash flow, to continue to look at acquisitions as well. With respect to acquisitions, typically what we're looking at is technology tuck-ins. So we look for tuck-ins that fit and complement naturally our CX1 AI platform. While we are primarily focused on these types of tuck-ins, we also remain open to larger acquisitions as well, as long as they're meeting our strict criteria, both financially and strategically, and they continue to drive us on that roadmap and the overall CX AI strategy. Finally, the expanding share buyback program. You've seen that we've increased our buyback throughout the course of this year.
At the end of the third quarter, we had increased our buyback by 18% on a year-to-date basis, and we have a tremendous yield of a buyback yield of greater than 5% over the last 12 months. We continue to prioritize in addition to the spend we're doing organically as well as for acquisitions in our buyback program. Finally, our rock-solid balance sheet is something we're highly proud of at NICE. We continue to fund all of these capabilities that I've highlighted as well, those share repurchases. We announced a $500 million buyback program earlier this year, and we're also continuing to make that a mainstay of our program. These three pillars and the strength that we have in our capital allocation provide us with great ample flexibility to fund ongoing top-line growth as well as drive increasing shareholder value and returns.
We've built NICE both on resilience and flexibility, and we have an incredibly strong financial foundation that is going to allow us to continue to accelerate this top-line growth. Our plan is to accelerate that top-line growth and shareholder value by making some targeted and strategic investments in the coming year. These investments are going to be highly aligned with the growth catalysts that Scott talked about earlier today and that I'll share a little more about to ensure we win with precision and purpose. These are the five growth catalysts that Scott highlighted earlier today. What I'm planning to do is actually show you some financial metrics and data behind each of these that demonstrate the great positive momentum we're already seeing in each of these growth catalysts.
All of these five growth catalysts, of course, are overlaid by the incredible cross-sell and upsell motions that we have at NICE. We have multiple different levers that we are using to accelerate top-line growth. It is important to highlight as well that this cross-sell and upsell motion, we now have even a greater customer opportunity by also cross-selling into that Cognigy install base as well. Growing demand for our AI is our number one driver of our top-line growth at NICE. We are seeing rapid adoption of AI across all facets of our business, and you can see we expect Cognigy to further accelerate that growth. Customers are choosing NICE because they look for us to lead their AI transformations.
They see the positive impact that we have on the customer experience, that we understand the customer journey, and you can see financially quantifiable metrics and outputs like you saw from Phil earlier today, that this is real. Customers can measure the positive impact and ROI that they see from our technology. Our customers are choosing our CX1 platform, and this is allowing them to augment their workforce. It allows them to orchestrate workflows and, of course, to automate with AI and our Cognigy and AI, as well as our organic NICE AI as well, to automate those experiences using AI. Here, I want to share a little bit in numbers how we see the acceleration of AI in our business. From the start of this year, we began sharing with you our AI, ARR growth on a quarterly basis.
You can see throughout the course of this year, our growth is getting stronger and stronger, and this is even before the acquisition of Cognigy. In the third quarter of this year, just a few days ago, we shared that our year-over-year growth in our AI revenue grew 43%. That further increased to 49% when you add Cognigy in that picture. We remain confident that our CX AI ARR is expected to grow more than two times by the end of 2026 as we exit the year. We're on a great path. You can see that based on the growth. We see that in places such as our pipeline, the RFPs that are coming in, the recent bookings where Scott shared that we had AI in every single seven-figure ACB deal.
We see it in the numbers, and of course, we expect further growth acceleration from all of our AI offering beyond 2026. Earlier, Scott also mentioned that I would talk to you a little bit about the pricing model. From over the last several years, our pricing model has continued to evolve, and it has continued to evolve to really ensure we are monetizing the opportunities and this continued shift towards AI. I'm showing you here a very simple look of what our pricing model is composed of. There are primarily two key levers. They are users on the platform and sessions, which you can consider as interactions. We are monetizing across both of those. Regardless if you are a user human-led on the platform or if it is sessions, the pricing model will monetize and accelerate the growth from both of those.
What we've seen over the last year is the growth in the sessions is greatly outpacing the number of users. That means that the AI that's coming through and the shift to AI and automation, we see that in the volumes, which I'll share a little more with you. As customer automation is increasing, we're seeing the positive impact of that. You saw that in the prior slide with the increasing growth in our AI revenue that we're generating. Finally, regardless of the mix, whether a customer is using our Copilot to augment their users and drive a better user experience, or if they're fully agentic and using AI for automation only, our model is designed to accommodate and drive growth in any of these levers or a combination of them as well.
Here you see more evidence that our pricing model is working exactly as we designed it. We're seeing that the growth acceleration that Scott talked about, which has been phenomenal growth in the agent and agentless automated AI, is really taking off. The volume interactions that are AI-driven is really accelerating. You can see that there is a parallel growth as well in our AI revenue. The pricing model is working exactly as defined that as automation and AI use increases, we're seeing that direct parallel correlation into our AI revenue. Now I want to shift away from talking around the pricing model and talk more about what we're seeing from customers. You heard earlier from Nick from Lufthansa Group, but I'm going to share two other real customer success stories with you.
The first customer success story is a Fortune 500 media and entertainment customer, and this customer started their journey with NICE back in 2023. This customer initially adopted our core CCAS offering. That included also digital channels. As they stepped into using our CX1 platform, they created a really strong operational foundation for them with the customer experience. In 2024, you'll see that they came back and purchased additional AI capabilities. At that time, they purchased both our Autopilot capabilities as well as knowledge management, and that allowed them to automate the front office interactions, but also guide their agents in real time. You see that increase in revenue that we received as a result of that additional adoption during the course of 2024. They've had a great experience. This customer continues to come back to us.
Most recently, they have furthered that by now also deploying a lot of our CX AI models. Earlier today, Phil talked about the strength of the data, as well as Jeff, that we have at NICE. For decades, we have been delivering intent-based outcomes to our customers. We have proprietary data that allows us to drive the best resolution for our customers. Our customer, and this success story came back, took on that additional CX AI models in addition. You can see the great trajectory of growth we've seen over this period. In between 2024 and 2025, the AI growth in the recurring revenue for NICE was 71% year-over-year. The growth in the overall revenue and spend at NICE increased 45% year-over-year to $9 million. You can see that it's a great growth that we're seeing in the overall AI.
What's important to highlight about this story is the transition and the transformation the customer is going through. You'll notice that the seat count initially increased in 2024, but then as they adopted more of the AIs and it became further embedded in their environment, you can see that they actually successfully reduced the number of human agents in 2025. They were able to do this. They have reduced cost. We've delivered a strong ROI. Of course, it's a win-win for us at NICE as we've seen this great accelerated recurring revenue growth. I'm going to share with you a second customer success story as well. This is a very similar pattern where initially this very well-known Fortune 500 U.S. utility company came to NICE in 2023 and initially also purchased our core CX1 offering.
That customer very quickly came back and added more of our capabilities due to the depth and breadth of our platform. They quickly added virtual agents. They added predictive analytics, and they also added automated self-service. What you've seen in this customer is that they have been able to maintain a steady level of agents. In their normal environment and in their business, they would have continued to add more and more human agents. With the adoption of our CX1 platform and our AI capabilities, they have used our AI and agentic AI capabilities to be able to maintain that steady number of seats. They have done that with feedback to us that they've improved their customer experience because they're getting a much faster resolution and they're much happier in terms of the interaction with the agents, which are now being guided in real time.
Once again, you can see our economics worked perfectly as designed. You'll see that on the AI ARR, we've seen a 40% increase year-over-year. In the total spend, we've seen a 26% year-over-year increase. This is just in the last 12 months. This is just the beginning for us. We're seeing many, many customers that are on this path, initially adopting the core CCAS capabilities, then starting to get more and more familiar and adopting our other AI solutions that are allowing them to really drive ROI and further our ARR spend with NICE. The other growth catalyst, the second growth catalyst, is about the migration of the CX customers from on-prem over to the cloud.
Scott talked about that this morning, that the estimate is that there are 15 million human agents on a global basis, but it's still expected that there's only about 40% penetration of those customers moving that have already moved from on-prem into the cloud. This leaves a massive number of 9 million seats that are still available and that we will use to continue to cross-sell and upsell and bring in those new logos onto our platform over the next several years. There's a great runway ahead that this will continue to inject growth into our cloud. This is an area where at NICE we shine. I've shown you some of the strength that we've had in those customer success stories. You can see we have an increasing number of 1 million plus ARR cloud customers. We operate in the large enterprise more and more.
We are bringing on further large global brands. This will continue to be a tremendous growth catalyst looking forward for NICE. Today, these 444 customers already represent more than 50% of our cloud revenue. The next growth catalyst, you've heard us talk about quite a few quarters over the past year. We've seen enormous success in our international business. We've won some great deals, both large deals over $100 million in TCV, but many other smaller deals where we continue to gain momentum and grab further market share. You can see this reflected in the 36% compounded growth we've seen from the third quarter of 2020. Our international revenue and contribution from the cloud is now 57% of our total international revenue. More and more of this business is coming in the cloud internationally. This momentum reflects the opportunity that we have as well.
When I shared of the 9 million seats that are still yet untapped to move to the cloud, the biggest areas of opportunity there both are in the large enterprise and as well as the opportunity for us internationally. The next growth catalyst is accelerating our ecosystem. This year, we have both added many new strategic partnerships. We have also expanded a lot of key partners that we work with at NICE. You can see that 73% of our new large enterprise CX1 ACV year-to-date at NICE was led by our partners. We're seeing really great success. Through our integration with AWS, Salesforce, Snowflake, and ServiceNow, we're expanding our reach, we're accelerating our time to value, and we're driving strong customer success together.
This will continue to be a great growth catalyst as we continue to strengthen these partners and use this opportunity to tap into further market share across the globe. Finally, when I turn to our last growth catalyst, we want to talk about expanding beyond the contact center. First, I'll start with where we work today. Today, automating customer service is orchestrating workflows, which is really our bread and butter at NICE. This perfectly was demonstrated by the demos that you saw earlier from Phil. Regardless of whether it's human-led or fully automated, automating customer service is the core of what we do. Fulfilling customer service intents is really the next frontier for us at NICE. We have spent years mastering interactions in the front office between our customers and consumers. This was demonstrated in the example that you saw earlier today from Neeraj.
It shows really just a glimpse of what's possible and how we're continuing to extend the customer journey. Finally, when we look beyond the customer service, this is the TAM that today is sitting in front of us, but we have an immense opportunity for new incremental revenue. This new incremental revenue is not included in any of the financial models that I will share with you. This is an opportunity that exists going beyond customer service. This is areas where you've heard Phil talk, for example, around proactive outbound capabilities and allowing customers to do more selling to their customers. This is really just an area that is the tip of the iceberg that we haven't yet even really started to tap into.
The combination of all of these different areas provides us with incredible confidence and excitement about the future as we start here in the core of our growth, where we're already seeing great indications of that accelerated growth, moving into the mid and back office where we can naturally extend to the orchestration of workflows, and then, of course, the further capabilities outside of the customer service arena. I talked earlier about our cross-sell capabilities. This slide really shares just the strength of what we do at NICE, but it also demonstrates the intense depth and breadth that we have in our platform, which is now even further enriched with the addition of Cognigy. You can see that our customers routinely come back and buy more and more from us at NICE. We're seeing that our CX1 customers are increasingly adopting multiple solutions.
The deepening adoption we have of all of these solutions and our advanced AI capabilities in our platform is one of the reasons that we feel highly confident about the growth path and the acceleration that we expect to see looking forward, and also will provide long-term durable growth given the stickiness of how deep we are actually embedded within those enterprise customers. To summarize for the growth catalyst, you can see we are actually executing on multiple levers. There are multiple growth catalysts that we have in our business, and we have many ways to win. We are seeing momentum across all of those, and we will continue to fuel them, which is what I am going to talk about a little bit more now.
With the strength of our financial foundation, the healthy free cash flow generation we have, and this immense opportunity that's in front of us with the wave of CX transformation, we are extremely well positioned with our AI and our CX AI platform. We plan to make some strategic targeted investments. Here is what it looks like for 2026. In 2026, we plan to spend an incremental $160 million to seize the opportunity that's ahead of us now. You've heard about how excited we are, how confident we are that at NICE, we have the assets, we have the market leadership, and we have the financial foundation to fuel and win that opportunity. The time is upon us, and we expect to seize it by really making these strategic targeted investments. These investments are designed to unlock that accelerated future revenue growth.
I want to walk through each of these a bit. The first is the investments around cloud and AI delivery. The second is around R&D, where we've grouped these together, where we plan to spend about $95 million incremental spend in 2026. Also in the go-to-market area, we plan to spend an incremental $65 million in 2026. Let me talk a little more specifically about where that spend will be going. In 2026, under the cloud and AI delivery, we are going to continue to fuel the delivery of the cloud by optimizing compute, which is needed for the very large enterprise customers where we operate, also to expand our regional infrastructure and resource capacity. I showed earlier that we've had great growth internationally, and we're seeing continuous momentum with a market that's still largely underpenetrated.
We've invested in the past couple of years for sovereign clouds. We will continue to do so and further fuel it from here. Finally, we will also build AI centers of excellence that will help us drive that important time to value for our customers. This is all part of the delivery. When we look on the R&D front, for R&D, we purchased Cognigy. We're very excited about what agentic AI means to the inclusion of our platform as well as on a standalone basis, and we plan to fuel it further. We are going to continue to invest in Cognigy as well as our AI roadmap for our other AI capabilities at NICE.
Finally, on the go-to-market expansion, I've talked a lot about the success we've had this year with the partners that we've either brought on board or that are new for NICE. We will continue to focus on those and use that go-to-market to further enhance and increase our global reach. The last thing on the go-to-market is around the AI-first sales strategy. We are going to bring in tools and more focus on subject matter experts as part of our go-to-market that will also allow us to seize this opportunity. This is all incremental revenue spend on top of what you would typically see for us. In a few minutes, I'm going to show you how that all comes together. First, I want to share what the impact will look like on our gross margin.
In 2025, you see where we plan to exit 2025. The expected and estimated impact in 2026 is about a 200 basis point impact on our gross margin. It's important to highlight that once again, these investments are strategic, they're deliberate, they're intentional, but they're also time-bound. While we will make these significant investments in 2026, we will also continue to invest in the course of 2027, but you'll see those margins start to recover in 2027. You'll see in 2028 that we're returning back to our expansion path. This is going to shift a little way from the internal investments that we plan to make. I want to focus here around the operational rigor at NICE. One of the areas that we pride ourselves on at NICE is the rigor that we have in driving operating leverage.
Today, we are already using AI tools internally at NICE across all of the different domains that you see here. We have deployed them in highly measurable ways. What is important to highlight is that we are going to maintain that continued strong rigor in our business, that operational excellence that we have been known for. While we are making those strategic investments, we are also putting in more and more AI-embedded capabilities within our company to drive ongoing operating leverage long-term. Next, I am going to move to our operating expenses, where you will see that we have increased go-to-market and R&D investment to accelerate our top-line growth. First, I want to highlight it is important to note that the percentage of revenue that we are showing here for OpEx is actually inclusive of R&D capitalization.
We wanted to show you the full spend that we have on the R&D front, but that should be taken into consideration that the portion that you see on top of these bars is what we're capitalizing and putting back on the balance sheet. This represents the full spend we have. The plan that I talked about, where we're planning to target these deliberate investments, are showing up here. You can see that a big area of expansion and these deliberate investments first is in our sales and marketing, where we expect our expense ratio to move from 20% to 22% in 2026 and to continue to invest. That sales and marketing ratio will increase to 24% by 2028.
For R&D, you can also see this combined spend of about 15% of our total revenue estimated for this year, and that we will continue to further fuel that over the next few years as well. These investments are going to allow us to really seize the opportunity to make these intentional investments to be able to drive that growth momentum we see in the top line and to really capitalize on the opportunity both with Cognigy, but more broadly, the CX AI transformation that is happening. Now I want to bring everything together financially and what you should expect. We expect revenue growth to grow from high single digits, which is what we expect this year, and the 7-8% total revenue growth that I mentioned just a few days ago as we reported our third quarter earnings.
We expect to see that growth acceleration continuing to expand first into 2026, where that range is now moving from 7-9% in 2026, and to further accelerate to an estimated double-digit growth in 2028 between 12 and 14% on our total revenue. This will all be driven by all of those growth catalysts that I talked to you about earlier. The cloud revenue growth is going to be driven by those cloud, those different catalysts that I mentioned, and of course, with AI being the number one key driver of that growth. Our expectation for cloud revenue growth is a range of 12-13% for this year in 2025. We just increased our cloud revenue expectation coming out of third quarter.
I want to reiterate that expectation is that we have maintained an expectation of a 12% cloud revenue this year, exclusive of Cognigy, and incremental growth as well coming and the Cognigy acquisition that we just recently closed. Moving into 2026, we expect Cognigy to add about 150-250 basis points of incremental growth in our cloud. During that time, we have seen already the stabilization of our existing cloud revenue, where we also see many indicators of accelerated growth. With an expected outcome by 2028, we will see a 17-19% expectation in our cloud revenue growth. The investments that I highlighted earlier are intentional, and we feel that we want to move now to take time during this wave of transformation to really capture this growth and fuel it here at NICE.
Now, beyond the top line, I also want to share with you what it means for our margins. The margin profile with this incremental investment injection, we'll expect to see a shift from where we expect to outcome during the course of 2025, with operating margin this year of about 31%, a free cash flow margin estimated to be about 19-20%. We expect that additional investment to result in a 25-26% operating margin during the course of 2026, and an 18-19% impact on the free cash flow. Again, you'll see that return to profitable growth and the operating leverage in our margin coming out in 2028. That will happen gradually throughout the course of 2027 as we step into 2028. Earlier today, I mentioned the strength of our financial foundation and our healthy free cash flow generation.
Throughout this time, while we make these incremental investments to drive acceleration in our top line, we also plan to continue to return to our shareholders through our buybacks at least 50% or greater of our annual free cash flow. In any scenario, we are going to maintain a balanced approach where it matters, keeping a disciplined yield return profile that we expect to yield significant shareholder value as we capitalize on this opportunity and drive this accelerated growth. The cloud backlog is something that we just shared for the first time in the last quarter and a few days ago. We have seen that the cloud backlog increased 15% year- over- year. Excluding Cognigy, that same cloud backlog grew 13% year- over- year. The expected duration on that is approximately 24 months. That gives us really great confidence of this acceleration of our underlying business.
Of course, that will be amplified further by the addition of Cognigy. AI is driving the next phase of our cloud revenue growth. This year, of the $2.2 billion cloud revenue that we expect to achieve for 2025, AI represents approximately 12% of that. We expect, with AI as our number one growth driver, that AI will expand to approximately 30% of our cloud revenue in 2028. We expect to achieve $1 billion AI revenue in 2028. $1 billion of AI revenue over such a short period is, of course, a major milestone both for NICE, but frankly, the industry as well. We are extremely excited about this path and the indicators of growth that we see that we plan to capitalize on and further fuel.
Before I wrap up today, I do want to summarize and share what our value creation playbook looks like at NICE. It is composed of these four different areas of focus. The first is our cloud revenue growth, where we continue to be laser-focused on that acceleration. We are perfectly positioned to win in this CX AI market opportunity. Our assets and our short-term investments that we plan to make will allow us to get to this 17%-19% growth by 2028. The second focus area is our operating margin. I have shown you our great track record that we have at NICE of driving great profitability. We have the operational rigor that we will use to continue to drive that leverage in our model and deliver on this expectation with return to margin expansion in 2027 and rising margins in 2028 and beyond.
The third is the health of our free cash flow. With more than $700 million in free cash flow that we generated in 2024, we run a very healthy and profitable business. We have tremendous retention rates from our customers. You can see just how sticky our customers are. You can see how many different solutions as part of our CX1 platform our customers are using that are very embedded in their day-to-day and their mission-critical customer-facing opportunities with those consumers. For free cash flow, you can see we expect the margins to be 20-21% in 2028. Finally, capital return, where I've mentioned that throughout this period on an annual basis, we expect to deliver at least 50% of our free cash flow back to our shareholders through our share buyback, where we have a $500 million share buyback program in place for us to utilize.
Bringing this all together, we have the perfect playbook, the financial foundation. We are positioned to excel and win in this next era. Together with Cognigy, we could not be more excited here at NICE to drive forward and be the market leader in this enormous transformation that is happening in customer experience. Ultimately, leading with strength, investing with purpose, and delivering value. Thank you. That concludes my remarks. We are going to now break for lunch, and then we will come back for Q&A. Thank you.
Hi everyone, we are going to get started in just a minute. Are we good with the webcast?
Okay. Go ahead, Scott.
Before we kick off, maybe I thought I would just say a few words because you have seen the whole morning in the presentation. You have seen, I guess, the overall vision that we have, the product capabilities, the financial outlook.
I just want to set the tone and set the scene. We are in growth mode. Now, I sort of opened that up this morning, but when you're in growth mode, when you've got a market that has got a growth opportunity, we are unapologetically driving long-term growth. Unapologetically driving long-term growth. The vision for this company is clear: long-term profitable growth. The time is now. This isn't a market where I can sit there and say, "Oh, let's acquire Cognigy and let it run as it is and try to capitalize." We've got to double down. A lot of the investment that you saw is on the back of seizing the market opportunity. A lot of it is related to making the acquisition of Cognigy and the AI capability live, real, winning for this company over the long term.
We are clearly pivoting ourselves in a growth mode of this business. It is about the long term. We're driving that long-term growth. When it looks at the investment, and Beth had the opportunity to present it, it's not a long window of time where we're that additional $160 million. We're not going to go continually in that going down the margin. We're going to lift the margin back up, and it's going to be scalable, profitable growth as we continue to expand. We need to seize the opportunity now. This isn't a company where I can say, "Oh, let's just try to continue at the same margins and expect the growth profile." There are a lot of companies investing in the AI market in this space. We have an opportunity to lead and dominate this market, but we need to move.
That is the reason why you saw the implications on 2026. I hope you would have confidence that through that, we are able to then drive long-term growth. I would remind you, even without those investments, we are already back to organic growth of our core, accelerating that growth. We are re-accelerating our core growth. You add the Cognitive and the AI impact, and that growth becomes more long-term sustainable. I wanted to highlight that because I did not have the opportunity in the beginning. You can see the product capabilities where we want to go. You can also see what we are trying to drive in terms of that high-growth balanced business. The last that I will say is you will see the margin profile is a very favorable comparison to most of our peers.
We will drive that margin growth as well as the top-line growth in a more balanced way. Hopefully that answers, I guess, a couple of questions that I had immediately after the session. Happy to answer any more in that regard. I'm not sure who's controlling the.
If everyone could just raise their hands, we'll call on you and we'll get you a mic. Why don't we start with Tyler?
Hey, good afternoon. Thanks for doing this. Maybe a multi-part on the big highlight, obviously, investing for growth. Scott, you came in earlier this year, I guess. Feels like, as you said, longer than a year. NICE is a company that has historically been a very financially disciplined company, right? Beth showed the chart of kind of consistent expanding operating income. We've seen really good EPS growth, CAGR.
What specifically were the biggest areas that you kind of came in and said, "Gosh, we're massively underinvested here"? I guess for Beth, I mean, as I look at the free cash flow margins, last year, I think you did about 27% free cash flow margin.
Even as I look—26%, but close enough. Yeah,
mid-20s. Even as I look to 2028, it does not look like free cash flow margins are getting back to that. Just help us understand the free cash flow dynamics there too. Thank you.
I'll answer the first one on the investment and what was missing. It is probably not about missing, but it is more about emphasis and capitalizing on the opportunity. Three parts. The first is clearly a native AI capability.
If you look underneath our platform, we worked with a lot of partners that we needed their AI platform embedded and underneath to be able to deliver to an AI experience. The market's pivoting. I've presented the numbers. If anywhere between 30% and 80% of that volume of interactions can be delivered in an AI-native way, as this context, as the center of engagement, the customer engagement platform, we have to own it. We have to be at the center of it. That is a core part of what we offer. That's not something—that was a core capability. And Cognigy was the perfect mix. I would highlight Cognigy. If you heard Nick, scale, it's a platform. You don't need forward-deployed engineers. We don't use those terms because it's already built. An analyst, you could sit on there, create agents.
A low-code platform is a great platform view. AI is one. The second is international markets is not just about sovereign clouds, but it's also about feet on the street, partner ecosystem, our own teams to sales covering those markets. That market is not well-penetrated from a CCAS to the cloud. We've got to move faster and more aggressively to be able to seize on that opportunity. We were already doing it, but you do not want little increments. You've got to go and bolt because once they've moved to their CCAS cloud platform, it's pretty hard to change. The third thing that I would say is the underlying core AI platform when you combine this together in the CCAS platform.
We believe it needs the opportunity to accelerate that the Jeffrey's going to drive, that native AI CX platform in the way the market is going to require that requires that as well. Winning the AI, including in a market where we have zero participation, that drives tremendous growth, international markets getting the size and scale of the coverage. Last but not least is a core stack that we're able to then expand into those adjacent markets. Core platform going into the front, middle, and back office, fulfilling journeys, going into sales, going into marketing, those expanded areas. I think the way I would best describe it, Tyler, is we are really driving to capitalize on the market opportunity, but I can't sit and wait. If we sit and try to hold on and then manage it, we didn't buy Cognigy to run it as it is.
We bought Cognigy to be able to really double down our investment in it to continue with leadership so we can—because the market will move. It will not wait for us. In fact, we will lead the way, and it will actually cause our competitors to have real challenges because we will be able to dominate where we currently do not participate. Beth, do you want to answer that?
Yeah, sure. Let me take the free cash flow. As you can see, I mean, we consistently generate significant healthy amounts of free cash flow. When you look at the free cash flow we generated in 2026 relative to what we are seeing this year, first of all, I would say last year we generated more than $700 million in free cash flow. On a year-to-date basis, over the last 12 months at the end of Q3, we had almost generated another $700 million.
Continuing to generate profitable cash that we invest into the business. When you look on the free cash flow margin as a % of revenue in last year relative to what we expect in the current year, 2025, there are two key factors which are different year- over- year. One, of course, is Cognigy and the impact of that from the date of close through the end of the year. At the time that we announced the close of Cognigy, we mentioned that we do expect that it will be dilutive in the near term. Over the first 18 months is the estimation based on our financial models, that by the time you get to around the 18-month mark from the date of close, you will see a return to accretion coming from Cognigy, both in EPS and cash flow, by the way.
That's one of the factors you see in the current year. The second item that you see in the free cash flow impact this year is a non-recurring one-time item. If you look back on the second quarter, we had an issue of a tax matter that settled. That was a one-time non-recurring item in the second quarter. Those kind of items are you can't anticipate the timing of the close, but that happened during the course of Q2. It was actually quite a favorable outcome. You can see that because if you look at our effective tax rate during Q2, it was not changed. We had more than adequately reserved for an expected outcome there. However, it did impact the free cash flow impact this year.
If you normalize those two impacts, Cognigy together with that one-time non-recurring item, you'd see a similar margin to what you saw last year. I think more importantly, as we look forward, we're all very excited and sitting here about the opportunity that's ahead of us. We want to fuel it. I've talked about the investments that we plan to make to seize the opportunity. When you look at the expected free cash flow margin, even during the course of this fueling, where we're fueling the roadmap of Cognigy and our AI roadmap and everything we're going to do to capitalize and seize this opportunity, it's an important part that we feel is highly necessary to drive that accelerated top-line growth and durable sustained growth over the long term.
Okay, let's go over there If you do not mind stating your name and your firm, please.
All right, wonderful. Thanks so much for doing this. Rishi Jaluri, RBC. Really appreciate all the details and greater transparency. Definitely something I think a lot of us will go back and really appreciate as we rebuild our models. Maybe two questions. Scott, obviously, I can understand why today is very CX-focused, but I would have to imagine there is probably a lot of excitement around the FCC business as well, especially if you think about the maybe—I do not want to call it opportunity, but the risk that AI agents create more fraud, more complexity in the financial systems, which probably creates an opportunity for you and maybe even leveraging AI within the FCC platform to make it easier for firms like ours to stop fraud and really get just more insight into what is going on.
We'd love to hear a little bit more of your thought process there. One that would be, I think, both for Beth and Scott is just around the gross margins. Totally appreciate that you're bringing down gross margins in the near term. It makes sense. Totally get that. It seems like you have confidence in that bouncing back pretty quickly in 2028, in spite of the fact that your AI mix shift is going to go up pretty dramatically, right? Obviously, cost of inferencing might be coming down a little bit here and there, but maybe walk us through what gives you confidence in gross margins rebounding back, especially as that AI mix continues to go up and up and up, and we're going to keep having more complex models that just add more drag. Maybe walk us through that. Thank you so much.
Yeah, great questions. You're right. We put a great emphasis on CX today. I mentioned the FCC or Actimize, as most people know it, and the public safety business at the beginning. Obviously, the CX business, more than 85% of our revenues, is the primary. We are very focused on capitalizing that opportunity. Having said that, we are the market leader when it comes to financial crime and compliance. We have a great track record. The beauty of that business, as you describe, is AI from a machine learning point of view is entrenched and deep and really domain-specific. It's actually very strong in a highly regulated way. Businesses are clearly the bank, the financial institutions are working with us and using the generative models, but there are two factors that continue to play a role.
I think in the usability of what we're seeing in that space is companies where you've got your investigators are using AI to be able—their generative AI, our platform in more on the investigation and streamlining and shortening cases in the investigation. Explainability matters. I like the example that Nick used in the CX1, going from deterministic versus probabilistic. You can't use probabilistic when you're talking about fraud. It has to be something that's explainable. What we see is the continued investment of that platform. It's a great growth opportunity. As you know about the Actimize business, it's got a lot of runway in terms of the on-prem customer base and moving over to our cloud. It's a real growth driver. It's a great business in its own right and opportunity there.
Beth, if you— Yeah, for the gross margin, I would highlight a few things. I think, first of all, it's important to highlight that we are in the early adoption stage for many of our customers, both as you think about a lot of the very large international customers we've added over the last 12 to 18 months, as well as many of the AI customers, that they're in the early phases of adoption. We've already embedded a lot of those fixed costs without giving the benefit of the ongoing ramp that we'll see in the cloud that's going to be further accelerated, that growth trajectory as we see ongoing adoption of AI. We showed you a lot of that in the interaction volumes today. We see the signs out there. They give us the confidence that we've made these investments.
On the other side of that, you're going to see the cloud growth accelerating and driving accretion to the margin. Similarly, on AI specifically, we see that our AI solutions and Cognigy are going to be accretive to our gross margin. We've done a lot of the AI development of our AI internally over the years. To some extent, it's at some cost. Of course, we're continuing to invest, but a lot of the AI is our own technology. The AI proprietary Enlighten models is an example. A lot of the AI other solutions that we've built. Today, we integrate with a lot of the LLMs, but the LLMs are agnostic. It also gives us the flexibility that if we find that we're partnering with a third-party vendor that is more expensive, then we have the option to switch. We're not wedded to any particular vendor.
I think the combination of those things, the top line where we have early days in terms of the cloud revenue contribution together with the accretive nature of what we offer and the meaning and value that it brings to customers from these AI capabilities and more agentic AI, while they're getting a strong ROI. They're willing to spend more with us to save significantly more on their end.
Take a question on the end there. In the second row, please.
Thank you. Sorry for being inconveniently placed. Samad Samana from Jefferies. If I think about the disclosures around where AI revenue is in terms of cloud today, just back of the envelope—I apologize. I'm doing this in a notebook. It's about $260 million of AI revenue that goes to $1 billion in 2028.
Can you help us maybe understand how much of that is NICE pre-Cognigy? How much of that's Cognigy, given that you gave us the $85 million and mark for 26? I know, again, I'm not trying to make you do back of the envelope math without a pen and paper up there, but just help us think about those building blocks and have one follow-up or many follow-ups.
Yeah, I would start by saying that the Cognigy contribution, we expect it to add about 150-250 basis points of contribution to our cloud revenue each year on an annual basis. So that's a contribution from Cognigy. We broke it down this year as well in 2025. It was about a 50 basis point impact positively to our cloud growth in Q3. We expect it to be about 150 basis points to our Q4. That's for 2025.
As I said, extending to 150-250 as you look forward to next year. We expect it to continue to amplify our growth. This, of course, is already on the stabilization of the cloud you have seen on the core CCAS side.
Yeah, I would just add a couple of things about Cognigy. It is pretty hard by the end of next year. It is going to be pretty hard to distinguish. There is no other AI platform. It is the Cognigy platform embedded on the CX1 stack. There is no separate. We are using the agentic capabilities, the conversational capabilities, the models, the Enlighten models, the CX data, all those things obviously leveraging on the CX platform, but it is one AI platform.
What I would say is if you think about where our growth to get to that billion, it will be CX customers, CCAS customers of NICE, where we cross-sell Cognitive capabilities at scale. We will win standalone with Cognitive as a standalone player where we do not get reached today. That is all incremental. Then where we're able to go into the mid and the back office and do more things with our customers, AI capabilities beyond pure conversation, more use cases, more scenarios. Those three under the AI banner are the growth drivers. It is accelerating. As you can see, our total growth rate going up to the 17%-19% in the cloud, but the AI growth rate obviously way more than that. You do not really distinguish post once we've gone through the acquisition and the integration and we're investing.
A lot of short-term impact on that margin is very much around the acquisition of Cognigy and seizing on the opportunity in front of us. Once you get through that, it is NICE CX. One follow-up. Everybody describes AI as a gold rush. I think that NICE is showing their actions are behind it, right? Whether it is the acquisition of Cognigy, now accelerating investments for next year. Just maybe help us think through, though, what are the guardrails that you would put around in terms of seeing should we accelerate even the investments you announced today based on what maybe the proof points of success are? Would you overshoot it? On the other side of that, what are the guardrails on the other side? Just help us think about your framework around this as the anchor point going forward.
Yeah, so first and foremost, it's easier to chase the puck. If you noticed in the short-term guardrails, we presented five growth catalysts. Four of those are very clearly associated to the midterm guidance that we gave. AI everywhere, automating AI creation, CX market jump balls, international and strategic partnerships. They are clear and using those and having multi-pivots. Whereas I would argue two years ago, it was the CCAS move that was our primary growth driver. We have clearly put more arrows in that we can fire. The contact center beyond the contact center is really exciting. What Neeraj presented, the opportunity for us there is immense.
That is more about an engineering roadmap, but we're not getting ahead of ourselves around trying to chase those end-to-end scenarios because there is a big CX in customer service and automating that without needing to go too broad or too wide. One of the guardrails is be clear and sharp in our purpose and being maniacally focused on that. As these new things come up, we're able to then absorb. As the market, this will continue to evolve. We focus on those CX journeys, which means those adjacencies will be very purposeful in investment and very purposeful in outcome that we expect to deliver. The other one is on that international expansion, as much as I'm very excited about, we're not going to go into every country in every geography where we don't need to, where it has a high cost to serve.
We're really targeted about which market. Obviously with Cognigy, we haven't really talked about it today, but we get a tremendous uplift from the go-to-market that they have with their European presence, their relationship, and their knowledge. If you go into Germany, when they think about an AI market-leading company, Cognigy is at the top of the list. We're able to use that in the European market and be able to expand there. Same when we go into Asia. It won't be to every single country. We're not going to 50. It'll be really purposeful on the big market. I guess what I'm highlighting is our growth is really purpose and targeted. Our innovation is really purposeful and targeted.
We are delivering value that we can resonate as a market leader, as a CX market leader, rather than being an AI player that can do a lot of different things. My analogy is many AI players are trying to be an inch deep and a mile wide. We are deep in the CX market. I believe you will find in the different domains, you will get leaders in each of those domains, whether it be in HR or finance or supply chain or procurement. There will be leaders that will go deep in those domains. We will be deep in the AI CX market.
[audio distrotion]
Great. Elizabeth Elliott, supporter for Morgan Stanley. First question, just on the hybrid pricing model strategy. The growth in sessions outpacing users makes a lot of sense.
We're seeing AI drive a ton of acceleration, more interactions. Two questions. One is, are there any implications for the revenue visibility as customers shift more towards pricing per session rather than pricing per user? Second, one of the pushbacks that we often hear from customers is, as we price by session, by interaction, costs can quickly get out of control. Is there anything that you guys are doing in terms of guardrails or holding customers' hands in order to help manage the transition between users and sessions?
Yeah. It's an evolving one. First of all, we wanted to present on the pricing because it is a changing buying mix. When you do those seven-digit ACV deals, the proportion that is under the AI that is session-based or interaction-based versus seats.
Customers are being very purposeful about buying where they're ultimately able, if they're going to get efficiency, productivity, maybe even seat reduction, they're reinvesting back into the AI. Again, with Cognigy, we are able to win in both scenarios. The perfect one is you keep on increasing seats, interaction volumes increase, and you have the AI sessions or not. We've proven and we know with customers, such as the media company, that even when seats reduce, i.e., we can see that reduction, the AI sessions more than compensate because you're able to do more of those service requests. You're able to have more of that. You're able to then drive that. The mix, and I expect that mix shift will continue where customers will initially look to use productivity and keeping their seat count similar. That's what we've seen so far.
Ultimately, once automation really kicks in, the seats will reduce, but they will correspondingly increase their AI side. Now the second part of your question is about the value and maybe those we're able to, one of the reasons why you use a CX-specific platform, you can go to an AI, an LLM, start doing this natively yourself and realize your compute is out of control. Phil presented and Jeff presented observability, analytics, insights about what the ROI is, what your cost to serve, but also what the return is. We're able to give them that observability in the context of CX. They can thoughtfully manage that without blowing out their costs. It's one of the reasons why you would come to a vendor like us. Last but not least, I will say this, we are very optimistic about value pricing models.
It's early days. Because we've got the data about where you get value, and for example, on Proactive, we know we've got examples with customers where they will pay a significant amount more per session on some of the outbound scenarios because of the value drivers that it will bring. They'll spend a lot more on those, and then they get the return on it because it's increasing top line, not just reducing cost. That gives us future opportunity to be able to use pricing power, using our knowledge about the value drivers and being able to step up to that. This market, most customers ask us for a value driver proposition, but they don't necessarily want to pay for it or price it in. They want to monitor it under a consistent pricing model, but that could evolve over time.
Let's go over here, please.
Thanks, Ryan. Tom Blakey from Cantor. Thank you, by the way, for the great presentation and lots of data here to digest. Maybe a multi-parter, but I'd like to hear from maybe Scott, Beth, and the panel. The platform consolidation examples that I think the gentleman here at the end gave was great in terms of that credit card fraud example. You're turning three- to five-day resolutions into minutes, but there was a lot of different software packages that had to be touched there. Along the theme of platform consolidation, is there enough investment here with this quick hit, if you will, investing in calendar 2026 where margins start expanding again to maybe shift in terms of your competitive landscape more on the workflow side and those adjacent market side? I'd like to hear from everybody about, are you now competing against ServiceNow and Microsoft?
Is there enough investment here that Beth's outlined very articulately to compete at that level going forward?
Yeah. Thank you. Jeff. Maybe I'll just kick us off. I think our focus initially is going to be just that CX workflow perspective. Helping customers from that point of customer engagement and automating the tasks and orchestrating across the enterprise. It isn't about taking out those applications. It's about orchestrating on top, right? That's how we're going to start. That's where customers want us to focus. We'll just see how it plays out, right? Does it mean those other systems just have less value to those customers? Do they replace them? We'll see how it works out. We're going to be very CX-driven in that regard, if that makes sense.
Yeah. What I would add to it, though, and I guess this is very important because it's very purposeful from our point of view. Those systems in the mid and the back office are largely built around human usage internally. Whether it be use cases, processes, transactions, it has a flow of activity that is driven by what a user or a department or an area does. We are very purposeful around a customer journey. That doesn't mean that the Lufthansa example, they were really clear about what journeys, what scenarios, what use cases, and what it took to fulfill those to achieve the outcome. You extrapolate that out and what the platform is able to do, we can be super sharp on building out those journeys, those AI agents on our platform that interact with those platforms.
It may mean that you do not need to raise a case or you do not need to initiate that task because it is automatically happening from our platform. I say that because it is an evolving mix. As Jeff mentioned, and you know this better than anyone with your background at Microsoft, other vendors have got their own focus around whether it be CRM or finance or billing or those different platforms. They are looking at it from the internal usage. Our focus is from a customer journey. That will distinguish. When that all comes together, I suspect what you will see is our focus around strategic partnerships. For example, with ServiceNow, we are building collaboration with ServiceNow because if you are going to build in internal workflows that can be used for customer service, we will use it.
We want to help customers capitalize on that back office investment and being able to seize upon that, but we'll do it from a customer journey context, not from an internal workflow or ticketing context. Hopefully, that makes sense.
Let's go here on the end, please.
Hi, Alexa Rosha from Guggenheim. Thanks for hosting this today. I know you started to disclose AI, ARR earlier this year, but if you could just provide a little more color in what actually makes that up. I know it obviously has to do with AI solutions, but I'm just curious if there's a list of products, just trying to better understand how do you define AI, ARR. Thank you. Do you want me to cover it?
Yeah. We are very specific. Ultimately, it is a set of SKUs that are AI products, either through a direct interaction or a product that is used as a facilitation of that interaction for customers. Very specific, very clear and transparent and tangible that we can then track the growth. As we're tracking the volume of usage, whether it be an autonomous one or it's an AI agent that orchestrates that journey or it's an AI product, for example, Copilot that is helping assist, any of those are our AI products. It really does cover automation, orchestration, and assistance or augmentation. It covers that portfolio.
I would just further clarify that it is CX only, the AI that we report, and it's advanced AI as well. We have some kind of early machine learning-based AI that is in CX that we exclude from that. It is what we characterize as more automation-based AI.
Okay. We will take this next question from the webcast. The question is, on expanding the contact center, can you do this alone or do you need to make any acquisitions in this space?
Sorry, what was it? Expanding the contact center. Beyond the contact center. Beyond the contact center. I sort of come back to the point of the question that was asked around where we were going to invest. We are very purposeful around seizing the market opportunity within the current portfolio. Cognigy was a big acquisition that opens a market opportunity that is immense for us, both in capturing the automation play on where NICE does not play, the obvious install-based opportunity in bringing that together, but then under Jeff and Phil's leadership, being able to bring it under a single platform.
That platform allows us to go into the market adjacency, into the mid and back office, and that platform allows us to go beyond customer service into sales, into marketing, into other scenarios. Having said that, the amount of innovation that's happening in this market that can be leveraged either through partnerships or through potential acquisition, Beth made the comment that we would look at primarily tuck-in technology, tuck-in acquisitions to expand and adjacencies that would support us. We're always open to that as long as it delivers long-term shareholder value. If it's going to be large strategic ones, it really needs to be justified. We've got a very high bar of threshold about the return that would bring. Everything that is being presented does not assume inorganic moves.
It is very much around our current capability, obviously, with significant organic investment that we've talked about to be able to seize upon it.
Let's go right here, please, in the second row.
Sam Brandeis, Woodbush Securities. Thanks for all the color and insight today. I have a two-part question regarding a competitive landscape. This summer, Salesforce and ServiceNow announced a $750 million joint investment into Genesys. I know you guys made your own deal with ServiceNow and actually tightened with Salesforce recently. What does that joint investment signal to you? And how do you guys view your positioning versus Genesys specifically? Scott, I know you touched a little bit on hyperscaler competition. How do you see those players, I guess, evolving into 2026?
As they talk about AGI and ASI, if they were to achieve that, one or multiple of them, would you view that as a risk to you guys and overall industry?
Okay. I'll let you sort of cover the second part. On the first part on the investment, look, I think there's a couple of things to note. It's not just the investment that Salesforce and ServiceNow made in Genesys. I would argue the amount of the investments that are going into CX, AI players and pure players that are trying to, it is significant. It's a great market. This is a great market. I can't stress that enough. This is a growing market in its own right, even without because you've got increased demand of volume of interactions.
You've got a volume of, again, I'll just go back to Nick's presentation where he talked about, "Hey, the way consumers interact with brands, they use websites, apps, mobile." The convergence and the opportunity for us to be a primary engagement platform drives volume irrespective of the competitive landscape. Number one, I think those investments, and I can't speak on behalf of those two companies, but I think it recognizes that this is a growth potential. From a competitive landscape for us, all I can tell you is at the same time they were making those investments, they were aggressively pursuing expanded partnerships with NICE because of our market leadership, our platform. They need our reach, our engagement platform, and they don't want to be limited. I guess we use it.
We're very happy about the competitive scenario because let's face it, we've got a native AI platform. The others don't. You're going to have to use either ServiceNow or Salesforce's AI platform in the CCAS space. We don't have that concern. We're able to give the complete end-to-end on our stack, including integrating natively with ServiceNow and Salesforce and Amazon and the others, and we're able to do so in a proactive way. I believe it actually strengthens from a competitive standpoint, but we keep a close watch on that. I'll comment on the hyperscaler and then I'll quickly get on the hyperscaler side. Look, we collaborate with the hyperscalers. I mean, Amazon, we've got a great partnership. We've got hundreds of millions of pipeline being generated through the Amazon relationship because they've got a reach and a go-to-market that they can bring.
Our Cognigy business actually works openly on the Microsoft, sorry, the Azure, on the GCP or on the AWS platform. It has a public cloud opportunity to be able to work there. What I would say is this: those organizations clearly want to be able to advance their technology on a large number of use cases and scenarios. We are obviously aware of the depth and the breadth that you need to be purpose-built in this market, which requires very targeted investment that goes over and above what you are sort of making publicly available. I guess what I am available across all of those different spaces. We are using precision and focus to be one of our competitive moats. Speed is a choice, but also speed is a moat.
If you're able to be deep and targeted and specific, then even if those platforms or those companies are getting there, you're able to then still get competitive advantage. Do you want to talk about anything specific?
Yeah. On the AI side, clearly, these models are advancing very quickly. We see those as very complementary to what we do because those models need to execute on a CX contact center platform that has compliance, security, observability end-to-end. Again, massive advancements. Just even OpenAI has the real-time GPT or GPT real-time model that's voice-to-voice. Super complementary, right? We'll drop that in where it makes sense, but that's going to run on top of that substrate that is hardened for contact centers and CX. We keep on top of the latest AI.
Phil and team are already working with the GPT real-time model to see where it makes sense to use those kinds of frontier models. We'll just continue doing that. We'll stay up to date. We'll stay on top of the frontier, and it'll be complementary to what we're doing.
At the front. Hi, Willow Miller for Arjun Bhatia, William Blair & Company. Thinking about one of your more historical and ongoing growth drivers, migration to the cloud, can you help us understand what the ARR uplift from on-prem to cloud looks like now, especially if a migration includes AI or even Cognigy at the get-go? In the past, I believe you framed the uplift as two to three times.
Y eah. Make sure I just understand your question. The ARR uplift that's already embedded now?
Right. In the past, you had mentioned when a customer goes from on-prem to cloud, you could see a two to three times uplift in ARR. Yes. Thank you. I'm curious to hear how that could trend now as customers think about AI with that migration to cloud. Yes.
Okay. Thank you for clarifying. Yes. That's correct. When we look at the transition and the migration we've seen of large enterprise, as they migrate, first, if we talk about our own install base that were prior and previously on our legacy on-prem solutions, as they've migrated over to CX1 and our cloud AI platform, I would say on average, we see about that two to three times uplift of ARR for a typical customer. We actually have quite a few customers that have given us 8, 9, and 10 times uplift.
If you recall, I showed you a slide earlier today about the number of applications that the customers have deployed, the solutions that are embedded within CX1. They're all seamlessly integrated. You can see on that slide how I show that routinely, each and every quarter, we've seen that customers are kind of cross-selling and buy-selling. That just demonstrates the opportunity that we have there. During the break, actually, someone asked me as well about some of the customer or two of the customer success examples that I shared earlier as well. They noticed that not only did we see this great expansive growth in the AI, we also saw it in both of those customer success stories on the non-AI portion. Once again, it's demonstrating that our customers are coming back. They're large enterprise.
They have complex needs, and they're buying more and more of our capability. When we first introduced CX1, it was the seamless integration that was native to the cloud of workforce engagement and analytics and CCAS, the ACD. That all still exists. Many customers continue to buy that, of course, and the core capabilities we have. Our AI capabilities, Copilot, Autopilot, Cognigy, of course, the Autopilot, all of that is now further incremental revenue that they're buying. Further deepening of their relationship with us and extending the customer lifetime.
Just to be clear, we haven't actually modeled out, well, what is it with AI?
I think it's a little early to be able to say exactly what that ratio is because what we've seen with AI, and Beth mentioned this earlier, most customers start with a number of use cases, and then they build out. You get natural expansion as more and more scenarios. They start doing auto self-service. They do with assisted. They'll do orchestration journeys. It's not that they implement everything all in one go. It's an incremental build-out. We could probably relook at that, and maybe we'll get further update. It's actually a great question.
Yeah, it is a great question. I mentioned it earlier, but also you should recall that our AI is still in fairly early days when we're talking about our AI. Many customers are still in the infancy in terms of adoption. It has a natural expansion path.
Of course, you've seen that those interaction volumes we've shared are just growing tremendously. That'll continue to feed into our AI/ARR revenue.
Okay. [audio distrotion] We have time for one more question. If anyone has one.
Hey, guys. Thank you for taking the question and for the great presentation. This is Nick here for Pat Walravens from Citizens. Scott, you mentioned earlier today that you were working on decreasing the time to implementation. I was wondering if we could get kind of a baseline for, say, timing on that right now and maybe ideally what a potential goal is in terms of implementation.
Y eah. I'll say two things. First of all, we have an internal metrics. We don't show that, but we have increased or improved our time to deliver. I think it's maybe 30% or 40% we've already increased. We've already improved. So improved. Yeah. I'm sorry.
Improved. Importantly, though, it does really change depending on what scenario. A Copilot where you're doing assistance using the data on the platform and you can do those cases have a much shorter timeframe versus maybe when you're introducing your self-service for the first time, you're putting the checks and balance, the guardrails. Cognigy has a proven method and way of being able to roll out. We are very focused on the assistance that we provide for them. The first deploy use case might take a little bit longer, but after that, they're rolling out and they're implementing themselves. It does depend a little bit. There's not one model. I guess what we are focused, though, is to make sure we assist customers to be able to do it as quickly as possible. What are the issues that we're actually doing?
We've already got the pre-built connections. We've already got the data. We've already got the knowledge. We already know what the agents do. Putting that into that platform where you build it, where you've got the same or better quality level, you're testing to make sure that the AI agent doesn't hallucinate, delivers on the service, can manage to, at minimum, equal or better than what your human agents were doing, those sort of things you need to make sure that you configure and get it right up front. Once you do, you then go and scale the deployment. This year we've seen a significant reduction because we did invest in the AI Center of Excellence, and I spoke about that. We have seen an improvement by about 30%.
Maybe I can give more information in the future in future earnings calls about what we're seeing there. It's definitely a focus.
It continues to be one of the areas that we highlighted that we're going to fuel the investment as well. As you said, Scott, we already did a lot of improvement there with around a 30% improvement, but it continues to be one of the things I highlighted as part of the incremental spend looking ahead as well.
Thank you, everyone.
Okay. Thank you, everyone. I appreciate your time today. Those who are online, I really appreciate it. Hopefully, you got a lot of value. We're in growth mode. It's exciting times, but I look forward to answering any other questions following on from this event. Thanks, everybody.