Amdocs Limited (DOX) Earnings Call Transcript & Summary

All right. I think we're out set here. So I'm Richard Cohen. I'm the Wells Fargo software team. Today, I'm delighted to have with me, Anthony Goonetilleke.

Very good.

Practice. Group President, Technology and Head of Strategy at Amdocs. So thank you for coming here today. Appreciate you.

Thank you for having me.

Yes. And so I guess maybe a good place to start just in terms of level setting the conversation. Talk to me a little bit about Amdocs, the story, just kind of what does this company do? And what problems do they solve?

Yes. So we are a company very focused on customer experience and very focused on the vertical of telecommunications. So we've been there for quite a while. We provide a set of products and platforms that really sit at the core of telecommunications. And we don't do hardware. So it's purely software, all the way from soup to nuts. And we really focus on taking the complex and simplifying it, right? So I always joke around and say, if you're Zappos, it's a cool job. You just have to pick a shoe, get the shoe size and get it delivered. If you're a phone company, it's like you need a phone number, you need to transfer your phone number to the device, you need a handset. You need to add it to a family plan. You need to make sure there's promotions, there's discounts that get applied. Is it a yearly contract? Is it -- there's so much complexity that goes into the back end. But we try and really abstract all of that and really provide a seamless customer experience to the user.

Great. And so I think a good pivot point from that is just in terms of the business model, I think Amdocs has a pretty unique business model in terms of the software and services combination. Can you tell me a little bit about that business model?

Yes, sure. Sure. So we have -- obviously, we have very strong recurring software revenue model. Our customers are Tier 1 customers, right? Like these are not companies that are in and out these -- I can mention any one of their names, and you would know them. They are the top-tier customers in any country. And so we build products and software, and we invest in R&D. And then we also take accountability to implement them. So one of our philosophies is you get a product company, they throw the product over the fence and then you get a services company that gets it and they're trying to implement it and then there's a problem in the project and they're pointing fingers at each other. We like to take accountability to not just give you the products, but also implement it, make sure you hit your business outcomes and your KPIs. And so we feel that, that accountability model is very unique, and that's something that our customers appreciate because at the end of the day, we're there for the long term. We have long-term contracts with our customers. We don't necessarily lose customers, right? And it's not like an industry of 50,000 customers. It's a fixed set of customers. And so we want to make sure we're delivering value, but they're not just buying software from us, they're buying value from us in terms of what are their business outcomes are.

That's great. And I guess with kind of that vertical focus, I'm curious, when you go out there and look at the telecommunications providers that aren't using Amdocs, what's the alternative?

Yes. The alternative is, I would say, a mixture of like 17 different things. So for example, some house maybe in-house, then you may get this company, that company, try and like stick it together with duct tape and glue and kind of hopefully that you will work. We spend a lot of money investing in research, investing in R&D, trying to bring like the best of what others do to the telecommunications industry, also understanding that at the end of the day, our industry is mission-critical, right? I always make the joke that if -- obviously, we know about all of the Internet issues from today, right, with the Cloudflare stuff. But like your Gmail is down or WhatsApp is down, you'd be like, oh, WhatsApp is down. But if your phone service is down for 4 hours, I mean, people's lives are at stake, right? So when you're building software for the industry, you need to take like that responsibility is very, very high to provide a very high-level mission-critical carrier-grade level of software that delivers value. So I think when we build software and when we think about it, all of these things kind of go into it, and that ties back to your initial question of the accountability model, why we think it's very important.

Yes. That's great. When we think about, I guess, if we were to bifurcate the business, there are 2 customers that make up a big chunk of the business. I think it's around 50% of the business. And then there's a lot of customers that make up the other half. When we think about, I guess, the growth drivers or the differences between those 2 types of customers, I guess, first, what are the differences in the growth drivers? And then second is, does that change over time? Do those smaller customers -- can they look a lot more similar to those larger customers?

Yes. And that's always the goal, right? Like when we come in, we might come in and a customer might choose to buy a single product or multiple products. And we always want to add value. And because we have a wide portfolio, we always have an opportunity to sell more and increase our footprint, right? So if you look at customers like we have PLDT in the Philippines, right, and we have Globe in the Philippines that just started off small and kind of started to grow as they got value from us and bought more from us. So even though I would say, look, I mean, you talk about kind of this bifurcated couple of customers, but together, they're like revenue of $200 billion. So they're not small in a way, right? And we've been with them for the long term. We've had -- obviously, we've been loyal to them. They've been loyal to us and the business has grown over time. And we feel like also there are many other opportunities around the world. Just this last quarter, we announced some deals with PLDT, with BT. These are customers that are very strong in their country, in their region. Brazil, for example, growing economy, growing customers. Japan, we had some announcements in the last couple of quarters. So we feel like definitely, like if you look at our top 12 customers, about 50% of them are coming from outside of North America, right? So that is obviously the goal, and we think that's an opportunity for us to expand and grow clearly.

Great. Okay. And so if we switch gears a little bit and talk about just the overall growth drivers of the business. One of them is cloud. I'm curious, cloud, I think it's 30% of the business, growing double digits. How should we think about how you guys define what a cloud customer is? And let's start there, and then I'll have a follow-up on that.

Yes. So if you think of telecommunications companies in general, they were the kings of data centers, right? Because they -- back to what I said before, they were running mission-critical stuff. No one else could run it for them. So they had their own data centers on-premise and they used to run software in them. Fast forward 2025, there is a bifurcated approach of moving some of these workloads to the cloud. So we have very strategic partnerships with AWS, with Azure, with GCP, and we help our customers move these workloads to the cloud on a couple of different ways. One way is migrating to our new stack, which is cloud native, cloud-enabled, which runs on public cloud. So this is one way to do it. Another way is if you have legacy stuff, we think of -- we call it refactoring technologically, right, because maybe you don't want to put a huge investment in it, but you want to move it to the cloud. So there's multiple different ways that you can take these workloads from on-prem and then migrate it to a public cloud. And so this is where kind of the growth and the revenue comes from. And we're still -- I would say, we're still maybe in the -- towards the end of the first quarter to use a football term in terms of the migration to the cloud. So I think that's still a multiyear journey. And right now, the focus is on IT workloads, but we also have network workloads, which is an opportunity. So I think this is just a multiyear journey, which we've done some very interesting acquisitions like a company called Astadia, which helps migrate and modernize mainframe systems to the cloud. And so we think this will be a good growth pillar on an ongoing basis for the company, both from a product perspective and also our services perspective.

All right. You stole my second question. It was going to be what inning of cloud are we in? But that's great. And so I guess like kind of the second addendum to that is similar to cloud, your earlier stages on the curve with AI. So AI has been a big focus for the company lately. I guess let's start on the product side. Tell me a little bit about amAIz that's kind of live now, and you have some new things coming next year. So talk to me a little bit about the product side.

Yes. We kind of split it into Horizon 1, Horizon 2, Shuky, our CEO, very, very early on, I think it was like almost like a month after kind of OpenAI publicly launched a couple of years ago, we said, look, like we are doubling, tripling down, and we want to be the leaders in the telecommunications space. And we never look back. And some of that includes disrupting some of what we do internally ourselves, right? But it's better for us to do it than anyone else. So it's clearly an opportunity for us. Some of the results we see, so we've been running more than a dozen POCs around the world. And I've been in technology for 20 years. I have never seen technological results like this. We created a partnership with NVIDIA. So we're working with the Agentic framework to include it in our amAIz platform. And in Horizon 1, some of the results we saw, just give you a few examples. So if you take a call center, telecommunications providers have some of the biggest call centers in the world, right, probably next to financial services. And they generally measure 3 key KPIs. So the first one is called average handling time, meaning how long am I on the phone when I call, right? The second one is called first call resolution. Do I resolve your problem the first time you call me? And the third one is around what we call TNPS, or Transactional Net Promoter Score. Basically, it means are you happy with the service I provided you? Yes or no, would you recommend me to anyone else? All the KPIs boil down to kind of these 3. Generally, a technological project or transformation gives you like maybe a 7% uplift on a good day, a 10% uplift. Like we are seeing 40%, 50% uplift on all 3 KPIs. And now we're starting to see in the last 2 quarters, customers going from proof of concept to go to production. So we think there is definitely something there. Obviously, we don't play in the BPO space or the labor arbitrage. So we are good with bringing technology that really adds value and moves the company forward. So this is really an opportunity for us. And as we look towards Horizon 2, we believe there is an even bigger opportunity to help our customers kind of use their systems of record and then put what we call our cognitive core to modernize their transformation and go faster. So think about recognizing revenue faster, launching new products faster by using generative AI. And because we have this -- so we've built like a very deep taxonomy in the telecommunications space and in [ ontology. ] So these are things that are very verticalized in terms of generative AI. So today, a lot of people talk about horizontal capabilities, which are good and nice. But when you're delivering results and value to an industry, you want to be very verticalized and understand what you do, right? So it's not just good enough for you to take an order, but you need to know that you can provision it, you can deliver it, that the number that you use is the number that you're going to get. So all of those things are being taken into account. And so we are very bullish when it comes to generative AI. We think it's an opportunity for us. There's really no downside. And especially the strategic partnerships that we have with NVIDIA really kind of helps propel us in that space.

That's great. So when we think about -- kind of you mentioned some of the ROI that customers are seeing with some of these pilots. And so when we think about how that translates to your revenue, how should we think about the uplift from a customer that's now starting to use a lot of the AI technology that you're providing?

Yes. Look, I think as we go into 2026, this is the first time we're seeing some of these opportunities go from proof of concepts to production, right? So this is when we're just starting to see kind of some of the early Horizon 1 revenue start to hit. But we really think that customers are going to buy differently in the future, right? So it's not about saying, well, I used to buy this, I'm going to continue to buy that and then just buy a little sprinkled Gen AI on top, right? No, the customers' behavior is going to change, right? I am going to have a workforce with several capabilities that are being delivered by agentic capabilities, not just my labor. So when they're buying from us, now they're thinking differently and buying differently. I was just talking to -- with a customer this morning around the network space. That's another big one for us. There's so much data there, right? Service assurance where you can be just proactive. And a lot of this work was done by labor, done by BPO and outsourced work and things like that. And we think that should start to impact, obviously, our revenue starting '26. We're already starting to see some early signs of this revenue. But for us, it's all an upside, right? Because the core system business in terms of transformation will continue to be there. This is a system of record. It's not going to disappear. Generative AI is not going to replace the customer. The customer is the customer. But then you have these capabilities of being able to deliver things faster, better and also do that labor arbitrage for where you were using 10,000 people, maybe you only need 3,000, right? I know it's probably not a popular term these days, but it's happening, right? And people are going to be retooled and looking at different ways of doing things.

Yes. No, that brings up an interesting question, which is internally, I think a lot of the margin efficiency has been kind of utilizing AI internally. So what kind of things are you guys doing internally on the AI front? How have you seen that impact your own cost structure? And how does that kind of inform how you go to market with your own products?

Yes. I'll give you an example. We have a platform called connectX, which we launched in record time. We have about 15, 16 logos on it. It was kind of almost built Gen AI native, right? So it's smaller teams, it's faster to market. When it comes to our core bigger systems, we're delivering more value. There's a -- I don't want to get too technical, but there's a term called PIs, which essentially these are like project increments, which are releases we bring in and functionality we bring in, we're doing that faster, better. And obviously, that has an impact in terms of how we can deliver stuff, right? So we are early, right? It's not about just dipping our toes in. I think about 70-odd percent of our employee base is trained in one shape or another, and that will only continue to increase. I think there will also be a bunch of people that won't make the cut, right? And we're starting to see that, and I know it's a tough message, but there will be the ones that do and the ones that don't. But when it comes to software development, we are A to Z from all the way from design to build, to test, to delivery, to operations, we think generative AI is going to have an ongoing impact. We just saw Gemini launch today. The results are phenomenal, right? And we -- there's many great platforms out there like Cursor and things like that, that we embrace. But people do also forget that there is 2 types of software at the end of the day. If I go like to a 40,000-foot macro view, there is what we call deterministic and nondeterministic, right? The deterministic software still has a place, like enterprise carrier-grade, mission-critical systems. When you call 911, you want to make sure that, that call is being made, right? You don't want to best effort. So there are still a place for deterministic software and nondeterministic software as well.

And how -- I guess, on that front, I'm curious, like I think there's been a lot of conversation around the limits of what the current state of generative AI can do in certain pieces. And so I'm curious like when you guys were thinking about productizing this, what did you find was working really well? And you talked a little bit about the data advantage that you guys have. You guys have a lot of access to the telecommunications data that kind of runs through your own system. So I guess the combination of a couple of things, how does that data shape some of the products? And how does kind of the limitations you've seen so far, the things that it's really good at and things that it's not so good at inform also the product?

Yes. I'll start maybe from the end. Look, I think you can't have generative AI without data and vice versa, right? So I think the fact that we are the systems of record that we know -- for example, there's a term in telecommunications called ETF or early termination fees for different -- if you terminate a bundle or an offering or whatever, right? We know that there are 230 database entities that impact early termination fees, right? So we can program this in and we can educate the system, right, tell our agent that any of these change, it will have an impact on proration. So you're smarter, you're focused, you're not guessing, right? So this is like a very simple way to kind of explain how the data connects to the logic of, at the end of the day, kind of what you deliver. In terms of limits and what we found, look, I think the most amazing thing, and I don't think this is a surprise to anyone, is really the transition between kind of NLP, natural language processing, to technology, right? The fact that someone can go there and just type something in, in their own way, and it will translate. And by the way, we're doing it in -- we have agents running in Arabic. We have agents running in Spanish, and it translates across languages very, very well. So we think that is kind of one of the massive, massive advantages. I wouldn't necessarily say, look, there's a limit. Things are changing every week and every month. I think obviously, we are living in the golden age of speed at the moment. I think every week, I pride myself that I wake up and try to read for an hour, 1.5 hours every day to catch up. And even that's not enough anymore, right? Like there's so much going on. And so I think it would be unfair for me to say, hey, like here is the limit. I don't think anyone knows what that limit is. I think we're just going to get more and more value that's being extracted from it. And we will learn how to use it. We will learn how to have deterministic software, nondeterministic software. The models will mature more and more. I think we're in for a fun ride for the next 5 years. And I think if I think about what we believe we can do and how we can help our customers, it's really an opportunity for us. I mean there's no real downside.