Concentrix Corporation (CNXC) Earnings Call Transcript & Summary

All right. Good morning, everyone. Thanks for joining us today. My name is Dave Koning. I'm a senior analyst at Baird. I cover payments and services. I'm very pleased to be hosting Concentrix management today. I think many of you know Concentrix is a leading customer experience provider. We have CEO, Chris Caldwell. We have CFO, Andre Valentine; and Head of IR, Sara Buda. And we'll be discussing Concentrix, the story, the evolution. We'll talk a lot about GenAI today. And I'll ask a lot of questions. But then Ravi is also on the call today. And if you have any questions, you can e-mail him. I'll be stopping intermittently during the hosting to have Ravi jump in with any questions from the audience. He has e-mailed everybody in the last 5 or 10 minutes. So you should have in your inbox his e-mail, you can just hit reply if you have any questions. And then one thing I just have to read a quick disclosure. Please refer to the event calendar published research or Baird's website for important disclosures regarding the companies discussed during this event. And that's all I have for a little preview. But why don't we kick off maybe with a little introduction to Concentrix since not everybody knows the company that well. Maybe share the story, the evolution and maybe why the valuation is discounted and what the Street maybe is missing right now.

Sure, Dave. Happy to do that. And really glad to be here with you guys. Always good to see you and Ravi. And I think it is important that we spend a little time introducing ourselves because we are seeing a lot of new interest from investors, certainly some people who are very familiar with the story and others who are new. And so Concentrix, we are a leading business services company, providing kind of fully integrated end-to-end solution that power our clients' brand experiences and their digital operations. Quite simply, we power a world that works. So from a scale perspective, nearly $10 billion in revenue, over $1.5 billion in EBITDA, strong free cash flow, as we pointed out in our recent guidance. We serve over 2,000 clients globally, including over 150 of the global Fortune 500. And so very, very proud of that. We're also very, very proud of the tenure with these clients. If you look at our top 25 clients, on average, they have been with us for 17 years. These are very strategic, very embedded relationships. We see generative AI, and I'm sure we're going to talk a lot about it on this call, but we see generative AI as a strong market opportunity for us. We have one of the largest installed bases currently in our business using our own IP, which makes us unique and differentiated, we believe, in our space. Clients choose to work with us because we're more efficient than their internal operations, and we can leverage our domain expertise and our footprint to deploy seamlessly and at scale for them wherever they want the work done to drive kind of global consistency, yet local intimacy, which they value very, very much. Both of those things are valued. In a relatively difficult macro, we're continuing to grow in 2024, and we expect to continue to grow in 2025, and that's a key message that we need people to understand. Frankly, our growth rate this year, if you look at it, midpoint of our guidance is about 2.5%. That stacks up pretty well against kind of the global consultancies, large IT outsources. And yet, as you pointed out, our valuation is very different from theirs. And so that's why we're really happy to be here today to talk more about the business and try to deal with some of those disconnects.

Yes. No, thank you for that. And one thing, too, to kind of kick off, we'll talk a ton about GenAI, like you mentioned. But maybe just start a little bit with recent growth in the industry. We look back over about 25 years or so. And over that period of time, on average, the industry has grown about in line with GDP. Some years faster, some years slower. The last few years, a little slower, the prior few years, quite a bit faster, right? So it kind of ebbs and flows a little. Maybe talk a little bit about why growth you think in the industry has been a little slower in the last couple of years.

Yes, for sure, David. So if you think about -- you've gone back 25 years, so that dates us all. But really, there's been 2 big issues within that 25 years. One was a financial crisis in '08 that you tend to see the slowing dip. And then COVID, where you saw sort of this slowing dip and then all of a sudden, this kind of very further economy where just a massive amount of consumption that was happening, which drove significant growth in volume within the business. The reality is now is that we're seeing a fairly tough macro, not only in North America, but pretty much globally. And what we see from a client's perspective is in tough macros, they tend to have less consumption of their products and services. So there's just less volume in the channel. You tend to see sort of less iteration of new products, new services because everyone's kind of hunkered down, making sure their financials are done and so less complexity. Every time you tend to have that, you tend to see a bit of a slowing down within our business. In a strong macro, you see, obviously, unit sales, consumption of services going up significantly. You see a lot more iteration in products and services to be competitive in the marketplace, driving a lot more complexity. And that all benefits from an outsourcing perspective of what we can deliver for our clients. And so literally, you've seen this kind of challenge coming off of the super growth of COVID, things kind of slowed down. But as Andre pointed out, regardless of the macro, we've grown in 2024. We're expecting to grow in 2025. And certainly, a stronger macro just helps us as we continue to kind of look forward into the future.

Yes. Great. And what do you think over the next 3 to 5 years, what do you think are kind of the puts and takes of growth? Like what are some of the key drivers to growth or headwinds in the next 3 to 5 years?

Yes. So David, we spent a lot in the last sort of 5 years building a platform to kind of take advantage of sort of where the market is going. And drivers for growth for us is the consumption of our services, not only the services we've done historically, but all these net new services we've built out, whether it'd be our IT services, whether it'd be a lot of the new services in banking from know your customer, to AML, to credit scoring, to deep analytics, domain expertise in both health care and BFSI. So we have brought all these net new offerings to the table that we're starting to see very, very good consumption, winning the right type of long-term business that we're very, very focused on that. We've also spent a significant amount of time building out a lot of our own intellectual property that are now starting to get the benefits. We've talked a bit over the last kind of year about new releases of software and products that we've developed. We're now starting to get kind of critical mass and scale and the usage of that, as Andre pointed out, one of the largest installation bases of GenAI in our industry, a lot of it using our own proprietary technology as well as partner technology like Salesforce and Microsoft and Google and Amazon. So we feel that our offering is very, very compelling and a big growth driver. That drives a secondary growth, which is taking share from our competitors. And as we talked about in our Q3 earnings call, in a significant amount of the RFPs and bids and consolidation deals that were going on, we won over 80% of them, which is pretty significant. And so we're winning net new share within the business. And then lastly, what we expect is generative AI, as we've started to see, is driving net new revenue opportunities for us on top of all of our capabilities that we expect will continue to power our growth into 2025 and beyond.

Yes. And why do you win in the industry? Is it because -- I mean, you're one of the 2 very, very large CX providers. You have full scale. I know you've got, I think, 1,000 full-scale GenAI implementations, I believe. You've won a lot lately. You talked in the -- but like why you guys and not others?

So a couple of things. One, if you think about needing a full solution and often people kind of confuse what full solution means. Full solution means sort of understanding the technology or implementing it and sustaining the technology that supports a process. The process can be everything from managing a mortgage, managing a claims, actually doing the end integration with the websites, the phone, mobile applications, having an actual call or chat. When you think about the entire process around that, we're one of the very, very few companies that actually have the deep, deep domain expertise and not only have the services aspect, but also the technical aspect to bring that to life for our clients. And as we talked about in our Q3 call, we won a very large transformation opportunity with a financial services company, where we are literally building their infrastructure, supporting the infrastructure, doing all the services for complete country and customer segment for the next 5 years. There was no traditional players we bid against. We were bidding against large IT consultancy companies, which certainly have the technical expertise to deliver, but they didn't have the deep domain expertise around what we were talking about and that end-to-end customer experience that was so important to the client to really outsource a very critical part of their business.

Yes. Okay. And when we think about automation and GenAI, how much work was done in the year like 2000, so like 25 years ago is still around? And trying to get -- kind of get into the idea of like how much was done that now is being automated and how much is still around agent-based? And maybe just talk through like how the evolution of how services get done?

Yes. So happy to do that. And thanks again for pointing out that Chris and I have been in this industry for 25 years. We're feeling older by the minute here, David. But you know what, I think if Chris and I who were in this industry 25 years ago walked up to this business today, we wouldn't recognize it. And it's largely around the complexity of the services that we provide now and how all of that easier stuff, and forget about 25 years ago, even 10 to 15 years ago, has been automated. And yet while things have been automated, the industry has continued to grow, and we have continued to grow and grow shares and obviously, offering net new services. So -- and I think investors kind of miss that automation has been and continues to be -- will continue to be a tailwind for this business. Clients turn to their outsourced providers. Again, think about those very strategic relationships we have. Who are those clients going to -- long-term relationships? Who are they going to turn to, to help them figure out how to implement the technology to drive the automation. Also, they're going to look to outsource more. So net new outsourcing and net new services for us in supporting the automation. And lastly, because of all the investments that Chris has talked about, we think we can gain share against our clients, our competitors as well. So if you think about where we are from a complexity perspective, we talked about at our Investor Day 2.5 years ago that we had about 13% of our revenue that was fairly -- it was with low complexity. That's down to 7% today, and we continue to drive it down. And yet across that -- even that time duration, we continue to grow.

Yes. And David, some real practical examples of that. Like if you think about 3 years ago and you asked us, how many people do you have doing data annotation and understanding how to do data modeling for LLMs. First of all, almost no one heard of LLMs 3 years ago. Now, we have a massive workforce that's driving that and not only implementing the LLMs, but helping to tune them and helping them to be more subject matter expertise around the main knowledge that we bring to the table. If you asked us 3 years ago, how many times are you installing and implementing products like Salesforce or Copilot or Gemini or any of these products, de minimis. Now we are driving that into our client base and winning and installing and building out those solutions for clients. If you asked us kind of 3 years ago, how deep is your analytics practice? It was very strong and good. If you ask us now, it's a very robust part of our business, powering a lot of the things that we're doing within our infrastructure. And same thing with software revenue. 3 or 4 years ago, we effectively gave it away for free as sort of an add-on. Now we're driving revenue with it, not only for pure software, but also for the services that kind of go along with it as we go forward. And so really, the business has changed and future-proofed us continue to grow as we go into the next couple of years.

And if we look at like the history and look at, okay, the IVR trees and then web self-service and just all the things that kind of changed in the last 20 years, each one probably took some of the interactions and automated it. Like you said, we've had automation for years. Is this a much different period? Or are we just going through, hey, every year, there's a little bit of automation headwind and then there's new products that you provide tailwind. Is there anything that much different now than kind of in that evolution of the last 20 years?

Yes. It's -- in some ways, GenAI has kind of consolidated a lot of that automation. So if you think about 2 or 3 years ago, people were talking about RPA, RPA, RPA. People were talking about kind of redoing the UI/UX of applications to kind of drive better user experiences. They were talking about sort of voice IVRs and driving a better voice experience. No one talks about that really anymore, right? It's now this new GenAI delivery, whether it's GenAI automated agents, whether it's GenAI cloud practices, whatever the case may be. The reality is that's just some -- what replaced all of the rest of the products and still has sort of the same attributes from an automation perspective. And in some ways, is driving this peripheral work like we talk about, like generative AI is not set and done. I think people sometimes confuse the consumer application of going on a ChatGPT and getting a really cool answer to your question. And what, let's say, a bank has to do or a health care has to do -- company has to do with GenAI, which is make sure that there's no data leaks, make sure that it's always giving an accurate answer, make sure it doesn't hallucinate, make sure that it answers the same way a million times a day. Otherwise, you're going to have a lot of difficult applications. That takes people in order to kind of drive that. And we're also seeing a lot of our clients using the generative AI to augment our humans with humans in the loop to make them more productive to deliver a better experience for customers.

Yes. No, that makes sense. I'll ask one more question, and then I'll let Ravi kind of jump in with anything from the audience. Are companies like is it harder for them to have internal centers? Like for years, companies have had internal centers and they also outsource them to you. Is the industry just getting more complex with all the different types of interactions? Is it harder for companies to have internal systems or internal agents, and it just becomes more and more likely that they outsource to you and others?

We're seeing a bigger trend to people looking at outsourcing even with generative AI. It's very interesting. When you think of a CRM platform, for instance, many people thought, hey, let's do that in-house. We need to own this. And they might have started that way. But then they soon started to realize, look, our outsourced partner handles 500 accounts that have this implemented. They know the best way to implement it. They know the most efficient way to implement it. Why aren't we leaning on them and their expertise to help us be more efficient. And then they're starting to say, well, we don't want to operate around the world. We are a primarily European company or North American company or Asia Pac company. This is our center of excellence. we do need to service people around the world. How are we going to do that? We need a partner to extend our business into all these different areas. And so that's why we see with the complexity, the regulations that are coming out with all the different types of interactions that are being driven by the new technology that the conversations with partners, a lot more engaged and what they're looking for is a true partner, not a transactional vendor. Lots of that business out there. We don't think that will stay around for very long.

Yes. Okay. And yes, I'll turn it over to Ravi on any questions from the audience.

Yes. We got a few questions so far. So I guess just first, what are customers doing at renewals? Are they essentially expanding seats, cutting seats? And then what's the net recurring revenue that you guys are seeing among your customer base?

Yes, it's a bit of a combination. At renewals, what we're generally seeing is clients are looking for how do they look at their cost of delivery and change it up, more technology, maybe different shore mix, maybe kind of looking at the services they're offering customers and saying, hey, we need to expand this, we need to shrink this. It generally -- it works out fairly even at the end of the day when you're working with the client. We're also seeing clients that are using the renewal to consolidate volume to their preferred partner, which, as I mentioned, we've been very beneficiary of over the last number of quarters by being that preferred partner for the clients. That tends to be the conversations that are happening. What we haven't seen is clients saying, hey, I've got a great budget to increase for next year. So that actually plays into a lot of our technology solutions that we've been offering clients and giving us a competitive advantage on the renewals.

Yes. That makes sense. And then just another one on customer ramps. You guys noted you made some investments in technology for a longer-term contract with the client. So they're wondering sort of when does that investment pay off? And then essentially, what is the risk that, that investment takes longer than usual to see an ROI?

Yes. So we were happy to make this investment. And here's why. First of all, the investment involved here is doing the upfront consulting work and then the build-out of the systems on which we will eventually run this program for a large financial institution client, for instance, I'll use that -- the example that Chris used before as kind of the test case here. What we were able to do, we are bearing that cost upfront. But because we are -- the contract is far stickier. It is longer term. So it's a 5-year deal versus a 3-year deal. And we are able to recover that upfront investment over that 5-year period, including our weighted average cost of capital. And so we've done that through a combination of a little bit of pricing, a little bit of getting more control of the transformation levers so we can drive more efficiency that we benefit from and putting a gain share mechanism in place. And so all of those things are part of that contract. It takes a while, but we do think we start getting to that run phase, if you will, in the second half of 2025 and start to see the impact that will have on both revenue and margins.

Great. Yes, that's super helpful. And then just in terms of Salesforce, so they launched Agentforce recently. So wondering just the GenAI tools, essentially, they seem ready to go and can be customized easily. So wondering how you think about Agentforce? And then are you competing against Salesforce? Or are you guys targeting essentially different customers there?

We're actually helping Agent -- Salesforce deploy Agentforce. First of all, you have to have a complete Salesforce stack, which we have a number of clients who have a complete Salesforce stack for the technology. Then you need to have to make sure that the data is all managed and curated and ready to go. So we're helping all of that. And then we're actually deploying the technology and putting it into the infrastructure. And so we don't see that as competitive at all. That's one of the revenue sources that we're seeing as new opportunities to grow, not only with Salesforce, but a number of other clients that we do. We have our own technology that is LLM agnostic. So there are clients out there who are looking for a broad-based ability to use multiple LLMs. And we believe that some clients will only go that way. So we have technology that complements that, but also interfaces into Salesforce. It also interfaces into Dynamics. It also interfaces into a number of the other applications that are out there. And so from our perspective, again, we are agnostic and are driving the best solutions for the client. And really, a lot of the software players see us as being able to offer value around their products versus just licenses. A lot of them have a challenge of saying, I can sell the technology, but I need help selling the solution and implementing it. That's where we actually come in.

Yes. Very helpful. And then maybe just another one. In terms of the type of pricing that you see with GenAI, is there any move to sort of an outcome-based pricing model? And then essentially, what would be the margins on that outcome-based model as well?

Yes, Andre and I go back and forth on this. A lot of things start off in talking about outcome-based pricing and then quickly go away because the client and the process is not mature enough to be able to do it. Where it's very mature, you can -- GenAI or not, you can do outcomes-based pricing, and we do have some great clients that have outcomes-based pricing, and we both benefit because we're both aligned to sort of to the metrics that make it work financially for both of us. But generally, it's not catching us on as much as we expected and where we would have thought it would have been this year. Andre, I don't know if you've got any other thoughts.

[ Longer term on it ], outcomes-based pricing has been talked about in the industry for 10 years, and the progress in moving towards it is extreme, extremely slow for the reasons that Chris alluded to.

All right. I'll jump in again, too now. When we just look very high level at GenAI, what are kind of the potential benefits and headwinds? Like if we think, okay, you had 100 hours of call volumes with the company pre-GenAI. Now we're stepping into a new world. How much of that goes away, but how much new work is there? And maybe how much complexity work is there that they bring more stuff to you? Like how should we think about the headwinds and the tailwinds? And maybe -- yes, let's -- we'll just start there.

That's a very broad question, but hopefully, I'll be able to kind of slice and dice and answers. The first thing that we're finding is in any automation, GenAI, any automation, you tend to tackle the lower complexity transactions. And that tends to take volume out, but not time out. So for instance, if I take 40% of the transactions out, I might only take 10% of the capacity or time out because what's left are the very complex engagements, the very unique type of cases, a lot of different things that are going into it. And so it's not a one-for-one, which people somewhat forget. People assume, oh my gosh, I can take 20% of volume out is 20% of revenue, not even close. The reality is what's left is a lot more complex and tends to be priced even higher than what sort of everything is because as you can imagine, you tend to price on a pool of work of easy cases and very, very hard cases. If it's all harder cases, you tend to charge more for doing those harder cases. So that's the first thing that people need to kind of appreciate around putting in GenAI. The second thing is that generally, when you open up a new channel and GenAI has some uniqueness because people tend to interact with GenAI chatbots and GenAI technology a little differently. If I interact with a human, it tends to be a sort of a back and forth. I'm not going to try and play too many games. I might mislead someone on my package really get delivered or not. But what we found is that with GenAI chatbots, people tend to push the barriers a lot, saying, I never got the package. I never was going to get the package. I've never seen a package from your company before because there's no human on the other side of the phone to say, wait a second, that doesn't sound right, right? And so we've actually seen where when you put some GenAI technology in, people then start to go other paths to contact the company, which is somewhat unique and interesting that is actually driving some additional volume that we see. We also have started to see some additional volume come through in some of our health care accounts where consumers are using generative AI to interact with the client, where they're able to do very large engaging letters around something that they might be upset with, something that they might be happy with. That takes a lot more time to process than what historically people said, I should write them a letter, but I just don't have time. I should do this, I just don't have time. Now it's very easy to interact with a company using the generative AI technology. So we're starting to see some more volume that's coming through that historically we haven't had to deal with. So it's quite a unique balance. Where GenAI does exceptionally well, right, is helping humans get the information they need to have a more engaged conversation, chat, interaction, whatever you want to call it, with somebody live. So instead of having to kind of look up stuff and instead of having to kind of find information or record or figure out how to deal with something, all of that is happening in an automated fashion in the background. And then someone is adding that human layer of empathy, that human layer of kind of filtering about what's important of an issue and what's not important of an issue with that consumer. And we're really seeing about 80% of all our generative AI interactions and investments and thoughts from clients going down that path. On the pure automation side, right, where it does really, really well is sort of, again, more basic interactions like, yes, you can get a greatly worded answer. Yes, you can find the content. Yes, you can pick up the pictures. Yes, you can kind of fill a shopping cart. There's just a lot more complexity to that for the most part on the vast majority of transactions that we handle every day.

And Chris, when we do that, obviously, we get a new source of revenue and not only implementing the technology, but maintaining the technology because it's not a static environment. And if we do it well, which we tend to do, the client rewards us with more transaction volume, maybe more complex work, et cetera. And we've given a couple of examples on past earnings calls. Maybe they resonated, maybe they didn't. But what we see broadly across our client base is as we automate, and this was true back in the days of machine learning and RPA, the more we automate, yes, those revenue for that transaction volume may go away. But we're still growing with those clients because they're giving us share gains. They're giving us work they've never outsourced before. And now we have a revenue stream from helping them implement and maintain the technology.

Yes. That's helpful. Like -- so what percent of minutes -- like it sounds like it's possible a lot of transactions get automated and tons have already in the last 20 years, tons of transactions have gotten automated, right? But what percent of minutes at this point get cannibalized by GenAI? And then how much extra work do you get because of kind of your examples of, well, humans actually interact more because they write letters with GenAI or you get more work because you're efficient or companies need you because there's more complexity. Like what's the dynamic there between you lose some like those minutes, but then you gain back some?

I would love to say there's a hard and fast rule. What we see in our client base is that generally it takes a couple of quarters after we automate something, it takes a couple of quarters to kind of grow past where that automation happened. It differs generally by industry, differs a lot by the volume that, that client is actually seeing within their ecosystem depending on how successful they are in the market. So lots of puts and takes. But generally, on average, clients that we automate are our fastest-growing set of clients within our business.

So net ends up -- GenAI so far for you has net been more of a positive catalyst than a negative one.

For sure. Yes. David, on a couple of fronts. One, we've taken share from competitors because we were one of the first people out there pushing GenAI. And when no one wanted to buy GenAI 2 years ago, we started building our own technology. And as we called out in many conference calls, we have it on literally hundreds of thousands of desktops within our business of varying GenAI tools that we built ourselves to be more productive, more efficient. So that's allowed us to be more competitive in the marketplace. We've been able to talk to clients from a consulting perspective. We've been able to do the POCs, which have been chargeable. We're now getting them into production and scale. So that's been beneficial to us. It's raised our profile of a lot of the clients that we speak with because they actually see us coming to them with practical solutions. And there are some clients we tell them, in some verticals, they're just not ready for GenAI. We say, look, you might look at this 2 years from now, this is kind of what's going on in the industry. And so it's provided us an ability to drive thought leadership within our client base, which we get reciprocated in other ways from other business and winning more mind share within the clients.

And what percent of your clients have already kind of started to use GenAI? And what percent do you still have to go in and help them automate?

That's a good question. I would say almost everybody within our client base is using some form of GenAI. And sometimes it's incredibly small. Sometimes it's something that they try to put in or POC. At scale, we have about half of our client base using it at scale. And I think investors misunderstand that a lot because they assume that all of a sudden, everything is going to disappear. But here, we're saying like over half of our clients are using generative AI at scale and yet we're still growing. That should give you an indication that there's a lot of robustness within our industry and our business about how we think about actually growing. And again, it differs around how clients are using it. Some clients as we've implemented it are for full automation, we've talked about it. The vast majority around to help kind of complexity of cases, complexity of research, complexity of data, kind of managing things better. And that certainly frees up some capacity, allows them to be more efficient, allows us to be more efficient. But again, as Andre pointed out, tends to win us more scale and more share within that client.

Yes. It's definitely -- you're right, a misperception. The investor community basically assumes that, oh, the minute GenAI steps in with a client, the revenue starts to fall off. And you're basically saying the minute they step in, the revenue actually gets better.

Yes. I mean, look, it's not for one for one right on the first day. I mean sometimes we're lucky enough to do it because we won a big deal and we're automating and growing at the same time. But absolutely, we're automating things and winning share and taking on more capabilities from it and selling new services to the client and new kind of auxiliary capabilities to the client. And that helps grow our business. And what also helps is drive a very deep stickiness into our client base so that as they think about their technology direction, as they think about where they want to grow, as they think about new markets to enter, we're part of that conversation. And that's really where we want to be in, right? That allows us to kind of help our clients be more successful and at the same time, grow our business.

Yes. Yes. Okay. And what verticals are more or less likely to use GenAI or a little bit earlier to start using it? And what verticals maybe are a little less using GenAI?

Yes. Look, telecom, consumer electronics, technology tend to be more pushing GenAI at scale. Banking, health care, insurance are using it in some of the back office. I think it's very early days, and they're really -- some of them just don't have the infrastructure and data ready to go. Some of them don't -- have regulatory issues to deal with. And some of them are just very, very apprehensive about having some of these things running to where it's many years down the way before they start to roll it out.

Yes. Okay. Now what's the benefit from using the same company for both the agent-based like model, the interaction like services, CX and the technology to implement, meaning like you relative to an Accenture. And the question -- the genesis of it is like Accenture is viewed as a monopoly that well, they're going to win all this GenAI work and they trade at almost 30x earnings, they grow a little slower than you, right? You grow a little faster than them right now. You trade at 3x, they trade at almost 30x. And...

[ You don't have to rub that in, David ].

Yes. I'm sorry about that. But it's just -- it's a little frustrating as an analyst to understand like what's the dynamic? I get their a monopoly. But maybe talk a little bit about why are you positioned relative to them and maybe you're both positioned well, but maybe talk why you're in the right position to win a lot of the work.

Yes. A couple of things. I think let's talk about the elephant in the room. The reality is people think that we are solely doing call center work. And the reality is far, far, far, far, far from that, right? We diversified our revenue very, very significantly. We have a lot of deep domain capabilities along a lot of complex environments that we operate for clients. And as we've talked about, we're delivering sort of these full solutions to clients. And so we do compete against them, have a lot of respect for them in a number of different verticals. And where clients see a benefit of having sort of the service delivery and the technically technology delivered by the same company is that we understand technology for good. We understand why we're putting in technology. We're not putting in bells and whistles for the sakes of bells and whistles because it's the newest, coolest technology on the planet. We're putting it in because it's going to drive a better experience because it's going to speed up optimization of processes. It's going to enhance the customer experience. It's going to maybe enhance an employee experience of our clients as well from some of the work that we do. And we know that because we literally handle hundreds of millions of transactions for clients around the world. So we know what their clients are -- customers are thinking. We know what the local cultures are. We know what their faults are in their systems because we touch them each and every day, and we understand sort of what best of class is. And by merging that with technology, it allows us to be more efficient for our clients. I cannot tell you how many clients we've worked with who have put in new technology and it has actually driven more volume, more complexity, not giving them the experience they want because they don't understand that last mile that we know very, very intimately.

Yes. Makes sense. Like I'll ask one more question, and then I'm sure Ravi has a bunch more from the audience as well. But talking a little bit about your IT services part of your business, it's about 8% of total. You made an acquisition maybe 3 years ago or so. How different is that than Accenture, Cognizant, that group, like in terms of what you do relative to what they do?

So a lot of it is quite similar and maybe a little more. But what we focus on is our deep domain expertise in the verticals we service. So if you're looking at redoing a core banking platform, that's not us. If you're looking at redoing sort of the credit collection systems, the mortgage systems, the UI/UX experience, the website, some of the digital operations, that is us. And not only do we have sort of deep domain expertise, but we also have the ability to scale. So we do a lot of engineering services that would be direct competitors, someone like a Globant or an EPAM or Endava or someone like that, right? We also have sort of the deep domain expertise around our platforms, whether it'd be Microsoft, whether it'd be Salesforce, whether it'd be Google, Adobe, all our platforms that we support and implement all around sort of the deep domain expertise that we have. So if you want sort of the experts around how to optimize your processes around deep engagement with customers and making sure that you're as efficient as possible internally to deliver those services, that's where we really shine with our capabilities that we bring to the table.

Yes. Got you. Well, I'll turn it back to Ravi now for a few more questions from the audience.

Yes. We just had a few more come in. I guess -- maybe on the GenAI economics in general, if we're thinking about the framework of moving from a little bit more of a people-heavy business, ramping up centers, that's sort of the legacy way to think about it versus now maybe a little bit more technology investment in GenAI, investing in LLMs, all of that stuff. How would that -- does that long term impact your margins? I guess that's the first question there.

Yes. So to be clear, we're really not investing in LLMs themselves, right? So we're investing and we are making pretty significant investments as we've talked about on an accelerated basis to productize some of the internal IP that Chris has been talking about. So as we think about our margins going forward, we're very confident that we can expand margins. First, we've got -- we're going to complete the work on the synergies from the Webhelp combination, $120 million of synergies that will have come through in 2025. Incremental will be in the $40 million range or so as we think about 2025. Then I think over time, we can rationalize some of that accelerated investment that we've been talking about in productizing our IT. So that can come down, and we'll start to see -- we've seen a lot of interest and acceptance of that -- of those products. We'll start to see revenue flowing through at very, very high margins. And then moving through some of the upfront investment phase of some of the transformational deals that we've talked about will also be margin accretive. Higher-value services just with more tech embedded, et cetera, and continuing to deploy the technology internally to become more efficient in our delivery. So all of those things have us confident that we can grow margins in 2025 and beyond.

Yes. That makes sense. And I understand the outcome-based, you said it's not really coming on as much right now, but wondering if there's any other sort of way to think about economics instead of just a price per seat basis, is there like a subscription model that you can essentially provide for customers? Or is there any way to separately think about how economics could change for the business going forward?

Yes. We're seeing more interest in gain share models, and we're quite supportive of that. That is kind of like, hey, if we can take out this kind of cost for this and you're putting this and we're putting this, this is how we're going to share the benefit of it. But it's more specific versus contractual for the whole contract. So that tends to benefit the economics of what we do. We're seeing more sort of shared investment models, I guess, specifically, we've talked about it around GenAI where companies are kind of going, we think we want this, you bring this. Hey, how about we share together and figure out a better way of delivering these things. And so we're seeing a bit of changes from that perspective. But overall, the traditional subscription SaaS revenue in our software space, you'll start to see the traditional kind of contractual types of models with bonuses and penalties based on execution, I think you'll continue to see. And then from the pure contractual gain share model, we'll probably see a greater growing, but that's -- I don't think this year or next year.

Yes. And then we got a question just about cash flow generation, normalized free cash flow generation. Wondering when the merchant integration costs should sort of settle down there? And then also going forward, what that means for essentially capital allocation decisions, buybacks, stuff like that?

Sure. And so you're right. Based on our last guidance, we're expecting $130 million of integration costs to come through. And those are all cash costs. So a pretty meaningful burden on free cash flow this year. We see those dropping off significantly next year as we complete the integration of Webhelp. I think certainly should be more than -- drop by more than half as we think about next year. And again, that will be additive to free cash flow as we go forward. As well we should see a little bit of a drop in cash interest as we go forward as we use the strong free cash flow to pay down debt, hopefully continue to get some relief on interest rates as well on the roughly $2 billion in debt that we have today that is variable rate. So all of that would be helpful from a free cash flow perspective and gives us confidence that we can see free cash flow generation go up next year and then step up again as we look into 2026 as we get through completely unburdened by the integration costs from the Webhelp combination. The other priorities for what we're going to do with the free cash flow remain unchanged. Obviously, we are committed to reducing our leverage, as we said we would when we announced the Webhelp combination. We think, though, that with the strong free cash flow, we can continue our share buyback program, which obviously, buying back our shares at the valuation day, we view that as quite attractive. You've seen us retire 1% of our outstanding shares in the last 2 quarters. And not suggesting it will continue at that rate, but you've seen us be active, and we certainly will have the flexibility to continue to be active as we go forward. And lastly, we do have a dividend. We just raised it in this most recent payment, and so we will continue to support that.

Yes. That all makes sense. And then we just had another question come in. Maybe can you review the Q4 trends that you guided to initially? And then what you are seeing going into next year as maybe improving or worsening factors kind of as we move into next year?

Yes. So what we saw as we prepared our Q4 guidance is that we did see clients, particularly in the consumer tech and consumer electronics space, really be impacted by lower sell-through of their products. And that sell-through creates volume for us. And so they adjusted their volume forecast down somewhat, and that impacted our guidance. We also saw some accelerated movement to get cost savings into this year, some accelerated movement offshore from some programs that, frankly, were planned to move offshore in the first half of 2025. That was accelerated a bit of a revenue headwind for us, a little bit of a margin headwind, too, as we kind of carry the duplicate costs, but that is something we work through pretty quickly. And we think that within 1.5 quarters, we'll get that work ramped fully offshore, and you'll see -- we'll get the duplicate cost out, and you'll see it be positive towards margins. Looking ahead to next year, even in the current macro, our confidence in growing comes from the strong new business signings that we've had in recent quarters, getting to revenue on some of the transformational wins that Chris has alluded to, ongoing demand for kind of our tech-enabled services, some early success in and momentum around our product sales, software product sales. And then we think on balance, look, this movement offshore has been with us for a long, long time, a little bit of an acceleration in Q4. We see that getting to a more normalized pace, frankly, as we go into 2025.

Yes. Yes. And then one more just came in. Kind of high level, what is your average tenure of your contracts? I know you mentioned some of your longest -- or your biggest customers being there for 17 years plus. Just wondering sort of the average tenure? And then what advanced notice do you need if a customer were to cancel?

So on average, our contracts are about 3 years in length, some a little shorter, some a little longer with sort of evergreen auto renewals that come on and that clearly drives the tenure up. We re-signed Andre, we might lose 1% of client base a year, either through our choice or their choice. So we have a significantly high sign-up rate that comes through every year. With most contracts outside of sort of the very, very long multiyear contracts, there's some ability for out with a penalty should a client exit or volume guarantees that can be wound down over a period of time if the client doesn't necessarily have the volume that is coming in. It really depends on sort of the investments we've made, investments the clients have made that how that mechanism works. Generally, you get decent visibility to it and you work with the clients, and that kind of is representative of how many clients we continue to deal with year after year after year after year within our client base.

All right. I'll jump back in. A couple more on GenAI. One is you talked about like how you feel like you can win in the industry with more sophisticated tools, scale, et cetera. Is -- this industry is viewed as highly competitive, right? Like it's very much service level like average handle time, like all those metrics are compared. And then if a client has 4 CX providers, they'll just move a little bit more here and there based on where the metrics are better. Have you created some sustainable advantage by some of the things you've created? And what prevents some of your smaller competitors from just creating some of the same stuff?

That's a great question, David. You're absolutely right. Historically, volume tends to be a little more fluid depending on how well a site is doing or how well something is happening. But with all of our investment in our own tools, we not only are at the top of the stack rank for the vast majority of our clients, so it's beneficial to send to us but also we can deliver a better experience or they're running on our technology that's actually running the whole process for us and some of the competitors and their internal operations, which allows us to be a lot more sticky with the clients. And so that fluid motion has kind of died down in the client base outside of telecom, which is, as you know, one of the verticals that we've been very kind of visible about talking about kind of shrinking that and focusing on the right business within telecom and have brought that down to sort of mid-teens from above 20%, not that long ago, while still growing our overall business intentionally.

Yes. And what about -- if investors, I think one concern would be. You've got a big group of clients. Some of them start to go through the GenAI transformation. There's some extra work. But like a year out, they find ways to basically like, oh, this is all getting so efficient and then the volumes decline maybe 6, 12, 18 months out. Is there any evidence of that? Or have you actually seen as they come on, it's like almost more and more just keeps coming to you?

No, we still see the volume tied very much to the macro, David. Like I mean, if the client is growing, they need more tools, they need more tuning on the LLMs, they need more changes within the agent that they're running. They need different configurations and technology on the background. And the complexity of the cases continues to kind of grow as a percentage as their business grows, right? Sorry, not grow as a percentage, but their complexity cases continue to grow as their business grows. And so we haven't seen that. And we have some clients who've had AI that we've put in place now for over almost 16 months, 17 months, 18 months, and that hasn't been the case. Where every company has been intentional about automation for us or a client has been done is very prescriptive. And we know it's coming. We know it's happening. We're generally the ones putting it in. And we're working with the client to say, okay, now that we've done this, what else do you need us to do and kind of look around. I think people also still to this day, kind of see from a consumer perspective as GenAI being put in and then it just works magically. There's a lot of things that are going on in the background to make sure that it works reliably and continue to drive it. And there's still a lot of areas, David, which I don't think people appreciate even with cost of generative AI coming down so dramatically, there's many, many places where it's still cheaper and a better experience to have a human do it than generative AI with power consumption costs, the LLM token costs, the technology and licensing costs and everything else that kind of goes along with it. That is not going to change anytime soon, which I don't think people necessarily understand even if prices continue to keep dropping down when you look at the total stack that's needed to support generative AI in a fully automated mode.

Yes. That's interesting. Let's see a couple of financial questions. One is, what percent of revenue, I guess, is tied to just agent-based type interactions? And I guess, as we go over the next few years, like does the amount that's more automation work, more IT implementation work, all that increase? And does that help margins? So it's really a margin question. Like is there some way to tie like, hey, 90% is agent-based today, it's going to move down to 85%, and that's a good margin thing.

That's an interesting way, Andre. The way I would look at it, David, is we look at each of our staff members moving up the value stack and therefore, becoming more and more valuable to our clients and being able to charge more for it. And then you've got the augmented revenue sources of the technology that goes along with it. And if you look at our total headcount, our goal is not to double our headcount to double our revenue. Our goal is that probably our headcount will stay relatively flat or increase as we continue to grow revenue, which, to Andre's point, helps our margin profile and certainly helps how we think about our cash flow going into the future.

Yes. So the margin drivers, scale, automation, some offshore shift, are those kind of the 3 bigger kind of drivers over the next 5 years?

It's a good question. It's certainly scale is one part of it. It's technology mix is one part of it. The domain expertise and services that we offer are one part of it. Clearly, we would do much greater margins on doing credit scoring for banking with our analytics and credit scoring group versus where is my package, which is part of that 7% type of work that we do. So complexity of that goes into it. And then really, offshore does have a portion that plays with it as well as when we talk about scale, the last part of scale is when a client has us like we have some clients that we deliver for them in 30, 40 different countries around the world, us managing that network for them and managing some of their internal resources as well to support that network absolutely drives a better margin profile generally.

Yes. Okay. And I'll ask one more and then we have about 5 minutes left. Ravi can ask anything else if anything else has come in. The cash flow, I think in very rough terms, I don't have my model in front of me, but something like 15% dividend, maybe 1/3 has been going to buyback and maybe the rest of debt paydown. So there's -- I guess, that question, just kind of what's your mix philosophy on that? And then where can leverage go like kind of over the next year, 2, 3 years? Like every year, do you try to take down 0.5 point or something?

Yes, that's actually a good question. So we were very -- as I said before, we're very clear that when we announced the Webhelp combination, our intention was to get our leverage down to close to 2x. And we've also talked about that at times being around our target for leverage. So I think in the near term, you will see us moving at a fairly rapid pace towards that with the strong free cash flow. Wow, yes, the dividend is going to use in the $85 million to $90 million per year range at its current level. And the share buyback, I think, will be in the same sort of range that you've been seeing us do, plus or minus. I think once we get to the target margin, I think then we have to think about our capital allocation priorities differently. Obviously, we have been acquisitive over time, adding quality assets to the platform, adding clients, adding technological capabilities. I think we'll want to continue to do that once we get our leverage down to a normal level. And in the absence of that, we could change the mix of how we allocate the free cash flow for a period of time because I don't think there's a need to drive leverage completely into the ground here.

Yes. Okay. And yes, we have a handful of minutes left. Ravi, did you get any more in the last few minutes?

Yes. We just got one asking high level about the potential policy changes. Any headwinds or tailwinds that you're seeing? I guess, how do potential tariffs on imports impact your clients on volumes that you serve?

Yes. Clearly, some of the products that we service are manufactured outside of North America, specifically the North American market. And if there's a reduction in volume, that would impact some of the volume that we do. If you think though, that is -- consumer tech is, let's say, 35% of our overall business. A big chunk of that is actually outside of North America. So it won't have a material impact on our overall business. The reality is that we have a lot of other verticals that would not be impacted by tariff, take banking, insurance, health care, telecom, all those others that -- and again, is a North American conversation that would have impacted. And probably high level, it's important to appreciate that 1/3 of our revenue is driven out of EMEA, 1/3 of our revenue is driven out of APAC and 1/3 of our revenue is out of North America. So again, you've kind of got -- sorry, 1/3 of our revenue is driven out of the Americas. So you have to kind of cut and cut and cut to get to what would probably be the impact of any tariffs that are going on. On the positive side, by the way, if you look at the EU and you look at some other countries that are coming out with generative AI legislation, we see that as, frankly, a very big bonus for us because it requires a lot more complexity, a lot more segmentation of data, a lot more regulatory approvals, a lot more security approvals, which plays into services that we offer to deliver those services across those regulated areas. So that we see as, frankly, a benefit with some of the governments that are talking about generative AI rulings.

Yes. No, that's helpful. And maybe just high level about the industry. Do you expect continued consolidation? And then maybe just touch on the benefits of scale in the contact center industry. Obviously, you probably have ability to cross-sell now with Webhelp and maybe just touch on that a little bit.

Yes. So first of all, we think there's going to be continued consolidation in the business because clients are looking for partners who can deliver globally across multiple geographies, multiple capabilities and that deeper and deeper domain expertise. We also see that clients are looking for investments in security, investments in regulatory compliance, investments in intellectual property that smaller players or mid-tier players just can't afford to invest in those areas to the standards that clients are expecting to manage their business. And again, if you look at our business, it's not contact center. That's one part of our business, but it's integrated into a lot of the technology that we do and the systems that we do and the other processes that we do. And the reality is that when clients are looking for a partner, we offer that full menu of services to them that makes us a very compelling value proposition for them.

Awesome. That's great. And maybe just one more. When you ramp up a client, what's a good trajectory of how you think about like in year 1, what the revenue could be as they come on? And then as you get to year 3 or year 2 or 3, what type of revenue increase could you see by adding more products or adding additional volume with a customer coming on?

We tend to see clients, and Andre can correct me if I'm wrong, generally take 6 to 9 months to get fully ramped from when the initial kind of, hey, let's work together to when it's kind of full at what we expected to happen when we first signed the contract. Revenue for the most part, tends to then increase over subsequent years, depending on the industry, depending on the other capabilities we sell into the client. And also frankly, depending on their demand of services based on the consumption of their products and services that they have in the marketplace. So clearly, if you have a client in one of our verticals that is going through explosive growth, we benefit very, very well from that. If you have a client that is more challenged in the marketplace, we tend to have revenue that's more stable and flat from that perspective.

Well, that is great. We are right at 9:00 Central [ time in the east ]. Good timing. So hey, thanks so much. I mean this was super helpful across many different topics. It sounds like you guys are doing well, which is great to hear. And yes, thanks again. Look forward to catching up soon.

Thank you very much for having us. I appreciate it.

Thanks for having us, guys. Good to see you.

Thank you. Bye-bye.

Thank you.