All right, welcome to our next session at Citi's TMT conference, and this is a reminder, this is for Citi clients only. This session, I'm pleased to have EXL Service, and from EXL Service, we have Chairman and CEO, Rohit Kapoor, and CFO, Maurizio Nicolelli. Thank you, guys, for doing this. Appreciate it.
Thanks, Ryan. Thanks for having us.
Yeah. So maybe for those that are newer to this story, could you give us a short elevator pitch on who EXL is, who you compete with, what's some of your competitive advantages in the market?
Sure. So EXL Service is a data analytics and digital operations and solutions company. We principally help our clients run and manage their operations, as well as adopt analytics and AI into their businesses and into their workflows. We've got 55,000 employees globally situated. We work with 750 clients, across the globe. We work with clients across multiple industry verticals, but the predominant industry verticals for us are insurance, healthcare, and our emerging business unit. Our entire belief is that we want to be able to help our clients by leveraging data and AI and creating value for them, leveraging data and AI.
We believe that the interplay between domain knowledge, the understanding of how to manage the data estate of our clients, and apply AI into the workflow creates a compelling differentiation as compared to other players in this space. We look forward to sustainable growth journeys with our clients and building up long-term and meaningful partnerships with them.
Got it. And at your Investor Day this year, you kind of touched on the company evolution over the years, and you touched on data and AI was there. But, I guess, what are some lessons learned through other kind of evolution cycles of the company, and what are you doing now to position the company more towards kind of data and AI?
So we've had multiple pivots along the 25-year journey that we've had at EXL, and we've made multiple pivots along the way. I think the important lessons for us have been to get conviction around some of the emerging trends early on and to be able to position the company to take advantage of those trends that are taking place. That necessarily involves change, and it involves a change in mindset. But having clarity on and conviction around what these trends are, what implications it's going to have to our clients and to our business model, I think that that's been the important element. And then we've been very disciplined in terms of systematically investing in these capabilities to kind of position ourselves well.
So, for example, we believe data management is going to be an explosive, you know, market opportunity, and companies that have strong data management skill sets will position themselves well to serve their clients, both in digital operations as well as in data analytics. And we've been investing in that space, you know, for the last several years. We've done three acquisitions in data management and built up a really nice competency set out there, and we'll continue to kind of build out more and more capabilities there. We also believe that, going forward, AI is going to be embedded in every single workflow. And in order to embed AI into every single workflow, you necessarily need to understand the workflow, which is the domain.
You need to be able to have an iterative and an agile, you know, implementation methodology that can allow you to be able to get to higher levels of accuracy, low levels of latency, and a low-cost structure. And you need to be able to know how to be able to use AI in an optimal fashion. So acquiring those kinds of skill sets and training our workforce to be able to deploy this at scale for an enterprise, that becomes critical. And so these pivots that we've been making, I think for us, these are very deliberate pivots, which are with conviction, and, you know, they do require change and anticipation. So we try and kind of stay ahead of that.
Got it. And we'll get more into AI in a little bit, but I feel like one of the more unappreciated things from investors at EXL is you've been honestly one of the only IT or BPO companies the past two years that's consistently beat and raised every single quarter. So I guess, what would you attribute this to? I mean, this question is for Maurizio, too. Is it more just solid execution? Is it you're setting the outlook in a more of a prudent approach to set up these beat and raises?
I think when we take a look at our guidance, we've been very prudent in really setting that guidance to a reasonable level. Obviously, we goal ourselves at a higher level at the end of the day, and we've been doing that for many years now, right? We've probably been doing that a good three, four years now. So we set those targets at a level that are our goals. You know, our guidance is a reasonable level that we'll set at. You know, keep in mind, when we enter into a year, like this year, our guidance for the start of the year was 9%-12% growth for the year. You know, right now, our guidance is right around 10%-12%, right?
So it's right in line with our guidance that we set for ourselves at the beginning of the year. So we've been very prudent, really, in setting that guidance and then executing to it, because if you're an investor, that's. It's really important to see that the management team is executing at least to their guidance that they're giving to the street. Now, you know, we will push ourselves further than that, but that's. We feel that's really important.
... I guess maybe shifting gears towards demand broadly, how would you characterize the overall demand environment, your current pipeline, the state of that? And I guess in the past two years and going forward, have you benefited from enterprises potentially focusing more on cost, maybe with your core BPO offerings maybe seeing some elevated demand?
Yeah. So we think the demand environment has been fairly stable for us, and it's been really helpful for the kinds of services and the kind of capabilities that we've got. We've seen our pipeline continue to be very strong, and we've seen conversion take place, you know, at an appropriate pace, you know, as such. We are also seeing larger deals come into the pipeline, and we are seeing more integrated deals come into the pipeline. And when I say integrated deals, I mean digital operations combined with data analytics, combined with digital, combined with AI, all being kind of infused and kind of being kind of put together.
So that's actually providing us, you know, to have a strong growth rate, which is a double-digit growth rate for the company, and, you know, seeing some sustained momentum in terms of our ability to build and grow our business. We're also seeing growth, and the demand side kind of be fairly even keeled across the three industry segments that we have of insurance, healthcare, and the Emerging Business Unit. And so that's pretty, you know, consistent across. We're seeing a greater amount of demand, really in in Europe and in U.K., where the growth rate of our business has been a little bit faster than the company average.
And then we have huge opportunities in other parts of the world, where we currently have not been able to take out our products and our capabilities. So there seems to be an adequate amount of you know demand for that. And then finally, on data management, there seems to be an enormous amount of demand that's there. Predominantly getting prepared for how to use GenAI, because in order to deploy AI or GenAI, you necessarily need to have access to good quality data, and therefore that's it's a big driver for growth as well.
Yeah, for sure. And you mentioned seeing larger deal sizes. You're seeing kind of normal kind of conversion cycles from pipeline to revenue. How about ramp times? Are contracts and projects ramping how you would think, or is there some kind of delayed decision-making there? And then also to just touch on win rates, have they those remained kind of consistent?
So I think the implementation of the digital operations businesses that we have won is fairly consistent with the past historic you know precedent that we've had. No real change you know in that. On the data analytics side, certainly it's much more project-based you know ramp-ups that we see out there, and there you know there can be variability in terms of ramping up and kind of scaling up that business. So that that's been there. And the other question you asked was?
Win rates.
The win rates for us, I think, you know, have basically gone up a little bit, but stabilized, you know, at the current levels. We tend to see a whole multitude of competitors. You know, when we are competing for deals, we compete with some of the large global IT services companies. We compete with some of the large digital engineering companies. We compete with some pure-play data analytics and AI companies, and we compete with some of the large integrated players as well. So it's a whole slew of competitors that we have, and our win rate is, you know, has been a very healthy win rate that we've had, and it continues to be pretty stable.
Got it. Maybe jumping back into data and AI for investors again that may be newer, can you talk through maybe some of your capabilities in AI, analytics, just data in general, and in terms of what, as you get more into those capabilities, does your kind of competitive landscape change a little bit? Like, are you not competing as much against maybe core BPO companies, and maybe it moves up the stack a little bit?
Sure. So, data analytics and AI now represents 51% of our revenue. So it's, you know, a majority of our business, and that's the part of the business that's, you know, seeing a heightened amount of demand. Within data analytics, it would get broken down into four service lines. It would include payment integrity, which works that we do in the healthcare business. It includes data management, it includes analytical services, and it includes marketing analytics. So those are the four service lines that we have. Frankly, marketing analytics has been challenged for us, and it's, you know, been a headwind for the growth of the business. But we've seen great growth across payment integrity, data management, and analytical services.
The kind of work that we do on AI and the solutions that we are building on AI, that's actually the most exciting part of our business. We are seeing a significant amount of traction around data extraction and, you know, a capability that we've created, which we call as XTRAKTO. That's something which is basic and something which is necessary for all clients to be able to implement. We're seeing a fair amount of activity around conversational AI and the ability to leverage AI in customer service and, you know, in engaging with end customers and then we are seeing a heightened amount of demand take place around code conversion, and we have a capability around Code Harbor, which is to convert legacy code into more modern and new-age code.
In AI, the part that we worked on in the last few months has been to launch our Insurance LLM. Because we do so much of work in insurance, we have access to derivative data of our clients that we've used to train a basic LLM model and create an EXL proprietary Insurance LLM model, which is built on the NVIDIA stack. And that Insurance LLM model is outperforming all the other models that we've tested it against, and we are looking at deploying that and putting it into production with some of our clients. So these are some of the, you know, areas that we work on in data analytics and on AI.
Frankly, you know, we are continuing to innovate out here, continuing to create new industry-specific solutions, and to be able to take that out to a broader set of customers.
Got it. And you touched on some of the IP accelerators you've made, invested in over the years in terms of, like, code, code conversion, the insurance LLM. Is this an opportunity for you to differentiate more in the market? Is it something that, especially on the kind of accelerator front code conversion, or is that becoming more kind of table stakes, like investors are expecting this type of capability with kind of the vendors they're working with?
No, I think it creates a enormous amount of differentiation. You know, the code conversion capability that we have, which we call as Code Harbor, we've now got clients who want to leverage that and license that capability from us. So it allows us to be able to make a one-time investment and to be able to benefit from increase in revenue and an increase in profitability across multiple clients. This is something which also requires a fairly high level of investment. So one of the things which is happening in our business is we're gonna be making more and more investments in terms of creating some of these proprietary solutions, which are gonna be much more useful for our clients, so that they can accelerate the deployment of AI in their own environments.
It's a source of differentiation, and it's also a source of creating greater pull-through revenue on digital operations. Because when clients are choosing a partner on digital operations, they not only want to get the cost, cost efficiency, but they want to work with a partner that is going to transform and modernize their entire operating stack and be able to elevate their performance. So we're finding that's helping on the win rate as well as in our ability to be able to have greater amount of pull-through revenue.
And you mentioned you have some clients wanting to license Code Harbor. So I guess, how are you charging for some of this IP? Is it an opportunity to maybe decouple revenue and headcount growth a little bit down the road?
Yes, it certainly creates an opportunity to be able to do that. Right now, there's a lot of experimentation and trial and error that you know, we are going through because we don't have the same level of maturity that we have on FTE-based pricing or time-based pricing. So we are kind of you know, crafting out things which work for the clients, work for us, and we are testing this thing out. But you know, certainly for us to be able to get license revenue, it depends upon the speed at which you know, our clients will use this and ramp it up. And there'll always be some level of service element to it as well, because there's also the implementation that we have to do alongside with the licensing.
Got it. And focusing a little more on GenAI in particular, you mentioned this IP, but is there other capabilities in terms of GenAI that you have beyond this? Maybe talk through some use case examples where you've worked with clients on GenAI, potentially solutions, and have you seen things move beyond kind of more experimentation-type phases yet into more production phases?
Yeah. We've used GenAI in you know something which we call as Smart Agent Assist, and that's something which we've deployed enterprise-wide across multiple you know clients. It's something which you know is able to augment the work of the human as they interact with customers and prompt them in terms of what the next best action is, as well as to be able to prompt them in terms of regulatory compliance and being able to you know to kind of have a 360- view of the customer. So that's something which you know is working very well.
I spoke about XTRAKTO, and again, that's something which we've you know, deployed you know, across the a number of clients across the enterprise, and that's something which helps us to be able to port data from one system to another system or from one application form onto a system. That's been you know, really useful, and there are other you know, cases that we're kind of building up. You know, we've used something in the retail industry that allows you know, our clients in the retail industry to be able to proactively look at trends that are taking place and adjust their supply chain accordingly, so there are at least fifty different use cases where we are deploying GenAI and these types of solutions.
We've also created, you know, the ability to use GenAI internally for ourselves and use EXL as, you know, a test bed for being able to deploy GenAI.
Oh, that's great. And I mentioned this on the last session, but, and investors were quick to put BPO in the loser bucket of GenAI in early 2023. By everything you're saying, it seems like it's TAM expansion, more of a net positive. But is there any, is the path, I guess, from the near term, medium term, long-term impact from GenAI, does that vary at all in terms of how you think of near term versus long term?
... So on the TAM, we think, this expands the TAM very, very significantly for us. We think, our TAM goes up by three times. There are a number of things that we wouldn't have done previously, and we wouldn't have participated in, which we are now participating in. Two, you know, in terms of, the near-term headwinds, sure, there are places where there will be cannibalization of revenue, but we think that that cannibalization of revenue is a net positive for us because our penetration rate is so low, that the more we cannibalize, the more customers want to give us business to increase work.
Got it. And, I guess one of the other kind of adjacent areas from AI is maybe clients might be more willing to accept more non-FTE kind of commercial models. So, I guess could you give some color on where EXL is in terms of client adoption of non-FTE, and as clients' views of, I guess, what is kind of normal productivity giveback? Has that changed at all as technologies like AI have become more upfront?
So the non-FTE commercial construct dialogue has certainly increased and gotten amplified very significantly. However, we've not really seen that result in actual client contracts being signed that way. We're still seeing clients sign client contracts the traditional way. The discussion around productivity gains and the benefits of AI, that's something which is certainly being, you know, baked in, in terms of the contracts that we are signing up with them. But keep in mind a couple of things: AI can be helpful in terms of improving the customer experience, and AI can be helpful in terms of improving the productivity. Our clients want both. They do want the end customer experience to be improved alongside with the productivity and the cost to be taken down.
We haven't seen anything meaningfully change in terms of productivity benefits over the last few years. I mean, in the past, RPA used to be deployed to provide for productivity benefits, and today we, you know, we are using AI to be able to deploy and provide a benefit to our clients on productivity, and I think we are still in that period where the amount of productivity benefit that you can get using AI is still to be well-established both from a client's standpoint as well as from our capability standpoint, so there's nothing very different that's you know happening out there.
Yeah. Yeah, I know clients have been hesitant to adopt non-FTE models throughout history because they want to keep those savings. But, is there an opportunity longer term, you think, for to push more towards non-FTE?
Absolutely. I think we're gonna see more of that take place. As we develop more solutions, I think it's gonna go into a more non-linear format, and it's gonna be a lot more outcome-based driven commercial constructs.
That leads to a benefit on margins, too.
That should be a benefit to margin.
Got it. Got it. Kind of shifting gears, looking at analytics. So you talked through the four components of analytics, but is there any way you could kind of rank or tier them in terms of which ones you have the largest exposure to? And I guess which ones have been growing the fastest recently.
So when we look at the four parts of analytics, you know, our biggest piece is analytics services. And the growth there was a little bit stunted about a year and a half ago from the banks slowing their spend, but that, we've seen, has reversed, and we're starting to see some good trend there in terms of growth. We're coming out of that. So that's been a very good positive. If you look at our payment integrity business, that has done extremely well for us over the last two, three years.
That has grown in the high teens over that period of time as we continue to get more business from new customers, clients giving us more claims to process through our engine, and then also the rate of inflation of medical claims continues to rise year after year, so when you couple all three of those, you know, it has really helped our growth rate. Data management has also been another big area of growth for us, and you know, Rohit talked about the big opportunity that we have in data management, and that's grown very well in the mid to upper teens for us within analytics, and so that's gonna be kind of the third-largest piece out of those four at the end of the day.
And so you have, you know, the largest piece is starting to come out and starting to show some very positive growth momentum. Payment integrity and data management have been growing very well over the last two, three years. Then the last piece is the marketing analytics piece, and that's the piece where we've seen, you know, a considerable amount of headwind over the last year or two. But that's starting to stabilize now, going forward, and also, it's become a much smaller piece of the overall pie in our analytics business. So, at its size now, it really doesn't move the needle anymore. So you're starting to see that momentum in analytics, and you're also seeing it in the growth rate, in our analytics business, right? So you saw the slowest amount of growth in the first quarter.
You saw the second quarter go from 4.9%- 6.5% growth. Then, if you take just our guidance for the second half of the year, you're gonna see an increase on that second quarter growth rate, right? So you're starting to see that momentum back in analytics, and that complements that very steady double-digit growth that we have in digital operations.
Got it. And on the marketing analytics piece, I know there was more insurance headwinds, I think, in 2023, in terms of a couple of states pushing back in terms of allowing rate increases. Is the headwind's still primarily in insurance, and market analytics in general, what will it take for that trajectory to improve and accelerate?
It's starting to stabilize, so we're seeing a bit of stabilization there. It's still very tiered to insurance at the end of the day, and it's tiered to the client subset that we have within marketing analytics, which is kind of the mid-tier insurance firms. Now, they are starting to sell policies within those states that they had pulled back in, you know, which we just haven't seen. We haven't seen them starting to market and significantly using our analytical services just yet, right? That could be an outcome in the next year or two.
Is there an opportunity to push that into additional verticals beyond insurance over time?
Correct. And we have been working on diversifying that, both into banking and healthcare.
Okay. And on the data management side, you made an acquisition recently in ITI Data. Could you go over perhaps the strategic rationale of the acquisition, opportunities it provides, any potential synergies you see from it?
Sure. So the acquisition of ITI is very complementary to the rest of our data management business. You know, it-- what it does for us, it brings us into particular vertical areas, you know, particularly in banking, that gives us a bigger moat or bigger opportunity within banking. Also helps us, you know, it also brings a certain level of capabilities that either augments or additional capabilities within data management that it really helps us now going forward. And it-- what it does also, it also helps us with adding those capabilities for us to be able to cross-sell. Now, when we bought Clairvoyant back in December of 2021 , there was a big opportunity there to use their capabilities to really cross-sell into our client base.
ITI is very similar, particularly within banking, for us to really build out and continue to outgrow that, really grow that business overall. One, it's a big augmentation to an area that we need more capabilities in. Two, it really helps us cross-sell and really build up and continue to grow data management well into the mid- to upper double digits in the teens growth rate.
Got it. And you mentioned analytics improving throughout 2024 , but I guess going forward, how do you expect... Do you expect it to continue to accelerate? I mean, I know 2025 will have easier comps in the beginning of the year, but just in terms of thinking potential, like medium-term or forward growth rates for analytics, the opportunities there, particularly as some of the headwinds, like within market analytics, continue to stabilize going forward.
We still believe in that long-term growth rate in analytics that we talked about, you know, about two years ago now.
Okay.
Right? And so we still are a big believer that the analytics piece of our business, you know, when you're coupling all four of those pieces, should be able to grow, you know, in the teens, double-digit teens going forward, right? Particularly anchored on payment integrity, analytical services, and data management. And as we do more AI services within analytics, that should be a big enhancer for us going forward because now we're increasing the value to clients, so that should really help us continue to grow within that data AI kind of, you know, piece of our business, which is more than 50% now.
Yeah, for sure. Data is a very important building block of AI. Be able to get their data estates in the right spot before they can get the full benefits of gen AI, so that makes sense to me. I guess kind of shifting gears to digital operations and solutions, I think that's been growing kind of faster than you would've expected in recent years. And how much of that would you attribute to the macro clients being cost focus, versus investments or just improved execution there?
I think that's largely been a factor of our ability to embed digital and AI into the workflow and be able to transform our clients' operations, and be able to show them the pathway to kind of making significant improvements to their operating processes. So, that that's resulted in a higher growth rate. We've also seen larger deals come into the pipeline because clients have greater conviction in our ability to execute. And so that's you know driving a faster growth rate for us. And you know and finally our ability to be able to access more C-level executives now that our you know presence has been built up, that's giving us a greater opportunity.
So is there a good opportunity? You mentioned digital and AI. Is there an opportunity to cross-sell between analytics and digital operations and solutions? I guess, what percentage of clients are using services in both segments?
So, today, we have close to about 50% of our analytics revenue coming in from clients, who also use us for digital operations. And the balance is, you know, they only use us for data analytics. And I would say almost, you know, the penetration of analytics services with our digital operations clients, that's pretty high, and that's something which we can continue to kind of build forward on. I think the real opportunity for us is, can we take data management to all of our clients in digital operations and kind of expand that? Can we take data management to all of our clients in analytics and expand that? And so that is a big opportunity area for us.
Got it. And you touched, Maurizio, you touched on analytics, getting back towards that medium-term growth rate. Digital operation solutions, you've been far outpacing the medium-term growth rate that you put out there. So do you think over time it may revert back to slower growth levels, or do you think what we've been seeing, you're able to potentially sustain, maybe not this level, but maybe a little bit below going forward?
... I would look. If you look at our growth rate over the last couple of years in digital operations solutions, it's been, you know, in the low teens. There still continues to be very good momentum within that side of the business. If you look at our pipeline within digital operations, it's still extremely healthy. Clients are really looking to transform their businesses, and they're coming to us for their efforts to be able to do that through technology, and particularly within AI. For us, continuing at that growth rate, at least going into 2025 , we still have strong growth to be able to continue to maintain that growth rate.
In that pipeline, are you taking share from the market or moving stuff that was maybe in-house with existing clients, so that you're taking more share from those work processes? Or is it potentially just TAM expansion in general?
There's still a lot of business that we win from clients outsourcing their operations, right? We still will win certain deals against competitors, but that big piece that we win is really from clients outsourcing their internal operations, and getting more... They're getting more comfortable with that and putting more of that out to bid. And so we're able to use our value, you know, proposition to really win those bids.
Are you seeing potentially first-time outsourcers getting more in the pipeline?
To a certain extent, yes. Yeah, we're seeing more of that, you know, come to the market. And that's, you know, that gives us great opportunity, both in digital operations and within analytics.
Got it. Maybe shifting gears a little bit towards talent and delivery. Could you walk through some of the trends you've seen in terms of attrition, utilization, what you view are kind of manageable attrition levels for you?
I think the attrition for us on an overall level has been fairly stable and, no real, you know, change in that. We have seen the attrition rates in our analytics business be slightly lower than what it's been previously. I think, you know, for us, managing attrition is something which, you know, at these levels is something that we plan for and we are able to, you know, implement and execute upon quite easily. I think the big thing is gonna be about being able to skill our employees and train our employees on some of these new AI technologies.
And that is an investment that we are making so that we can have people who are ready and be able to deploy this at scale, you know, with our clients. So that's something which we are investing in right now, and hopefully, we'll be, you know, ready so that when we do have that opportunity set, we can deploy this at scale.
Got it. And earlier this year, you announced a workforce rationalization that was maybe over oversized in terms of the amount of workforce that was being impacted on it, and the impact on the stock. But could you go through, I guess, what was the rationale from doing that workforce rationalization? Like, what job roles were impacted the most? Was it more of a focus on shifting skill sets to where the market's going versus improving margins?
Yeah. So first off, in the magnitude of that impact or that change was very small-
Yeah
... and it was like 1% of the workforce. Two, it was basically a mismatch of skill sets, and you know, we ended up having a surplus of skill sets in areas that you know, we weren't seeing adequate amount of demand for, and we wanted to kind of change that and reposition that, and I think you know, we went through that exercise in the early part of the year. That's now stabilized, and you know, you can see that on a net employee headcount basis. We've continued to add headcount every single quarter, and we've been adding on more and more headcount you know, as such.
These kinds of changes are something which, you know, we can, you know, plan for better and make sure that we can minimize this kind of a action, but it was something which we felt was necessary, and we had to make a one-time adjustment, and so we followed through on that.
Got it. And just in terms of go-forward delivery model, is there an opportunity to continue to have employees working remotely, working from home? And just in general, as we work through kind of our virtual workforce, there's an opportunity to have more of a distributed workforce model, where maybe you have smaller offices in other cities, and employees can kind of commute to that.
Yeah. Look, I think, the hybrid, you know, business model continues to work well. So we do have, you know, a portion of, you know, our employees coming to work, in the office, and we have a portion of our employees working from home. The ideal state for us would be a third, you know, coming and working from the office, a third being hybrid, and a third being, remote, and I think we are basically gravitating, you know, in that direction, but it's been a slow kind of progress and a slow kind of transition.
I think there's also opportunity for us to be able to build out and go into more tier two cities and more closer to where our talent is rather than having all of that talent you know in large centers in the metropolitan cities. So that's something which will happen over a period of time.
So potentially increase the potential talent pool where you go after it, in terms-- as you move into, some additional cities?
That's right.
Got it. I guess in terms of shifting gears more, as you know, in terms of recent margin performance, what have been some of the puts and takes, drivers of performance, and going forward, what's the opportunity to maintain or expand, continue to expand margins in the near term and long term as you deploy more AI, as you deploy selling licenses potentially for IP? Is there an opportunity to expand margins going forward?
Yeah, so we've done a lot on margins over the years. We closed 2023 at 19.3% adjusted margins. The first half of this year, we had a lower margin in the first quarter 'cause we wanted to make a certain level of investments to start the year, and in the second quarter, you saw a much, much higher margin. The average margin for the first half of the year was 19.4%. Right in line with what we were guiding. Like, we've always talked about have this year being an investment year, of having fairly flat margins with 2023. You know, for us, going forward, increasing margins 10 - 30 basis points a year should be achievable, but it's gonna be a little bit different than how we've done it in the past.
You know, now that we're getting much more into data and AI, driving more value to clients, getting more into those higher value opportunities and services, we should be driving a higher price and, which should result in a higher gross margin going forward. And you should be able to see that over the years to come. What's gonna be the offset there is what we talked about, is we're gonna need a higher level of investment. That'll be the offset. The net of all that should be an incremental increase to our overall margin now going forward on an annual basis. So we've come a long way from having 14%-15% adjusted margins about four years ago, to where we are today, over 19%, and then you should start to see, you know, us progress on an annual basis.
But much less than the big gains that we've had over the last two to three years. Increasing our margins 10-30 basis points a year is absolutely an opportunity for us, and that should help us continue to grow EPS faster than revenues.
On investments, where are you making those incremental investments? Is it more just broadening out your data and AI capabilities? Is it recruiting and hiring? I guess, where are we making investments currently?
It's really. It's, it's both, right? And so we are recruiting and then building, you know, software, AI solutions to really. It's capability development now going forward within AI is gonna be the big area for us now going forward.
Got it. And we're running up on time, so I guess just last question. What? Any last final thoughts for investors? What should we expect for EXL over the next couple of quarters?
No, look, I think we are well-positioned with the capability set that we have. We think that the demand environment is, you know, stable and, you know, strong. It really boils down to our ability to execute and our ability to innovate. And, we're gonna continue to focus on both of that and continue to drive our business forward.
All right. Great. Thank you, guys. Appreciate it.
Thank you.
Thank you, Ryan.