I'll start off this next session. My name is Adam Wood. I look after European tech research for Morgan Stanley here in London. It's a great pleasure to welcome Ravi Kumar, the CEO of Cognizant, to London with us. Ravi, thank you so much for taking the time and joining us. L et me kick things off. I think you took over as CEO of Cognizant in January 2023, so just coming up to two years ago. Could you maybe just remind the audience of some of the changes you've made within Cognizant over that time frame and how those initiatives and those changes have progressed?
T hank you for the opportunity. W hen I was coming in, one of the things I reflected in January 2023, and I saw Cognizant from outside as a peer and a company which was for a long period of time very successful. In the 30-year history of Cognizant, I would say we were in the winner's circle for more than 25 years. T he strength of Cognizant was it sensed what was coming on the way on technology waves over the last few, every technology wave. It incubated those technologies for downstream innovation. It built the tooling. It built the capability, and it created significant value for clients. C lients were excited about the fact that we could do this ahead and keep them differentiated. Investors loved us because we created a flywheel of opportunities.
O f course, our employees loved us because we did cutting-edge work ahead of others. T he changes I made were to pivot back to that heritage, which I believe was the key for differentiation. W e've now put up a leadership team which kind of cuts across the whole spectrum of capabilities needed for a tech services company, all the way from tech services to BPO services to infrastructure services, which is a new capability set. N ow engineering services, because over the last 35, 40 years, tech services companies focused on enabling businesses. Now there is a unique opportunity to embed technology into core products and services. T hat's one change I've made. I took the focus back to India. W e now have a new Chief Operating Officer in India. We have established a leadership team in India. T wo-thirds of my workforce is in India.
We now have a new CFO. We have a new CHRO. We have had a huge presence in healthcare and financial services. We progressively moved into communications and technology. I have a new leader for communications and technology. We have bought a company called Belcan, which gave us the engineering capability. W hat I've actually done is I'm using that as a beachhead to move into industrial and manufacturing. I've got a new manufacturing head coming on board. A whole bunch of changes. In fact, a large portion of our business comes from the United States. I think we are understated in Europe and Asia-Pacific. W e established a structure for capturing opportunities in Europe and Asia-Pacific. W e have two new leaders there. E ffectively, I've gone back to the roots of who we are.
I've put a structure which will allow us to capture those opportunities and seize those opportunities for the market. Very early on, I have called out an investment on AI. We are investing $1 billion in AI. I think there's a unique opportunity for the tech services industry. I've kind of created AI labs which can sense what is coming. Then I've created a structure to industrialize it inside. I think we are excited about the progress we've made in the last two years. I would say it's a phase of stabilization. We are just at that inflection point of starting to be in the phase of growth.
I f you look back at what you've achieved, what's gone well, what's gone badly, i n terms of getting Cognizant back to its roots, do you feel as if that work's done now, or is there still work to do?
Q uite a bit done. That's why I said we are at the end of that phase of stabilization. my attrition rates are at an all-time low. I have 13,000 returners coming back into Cognizant in the last two years. My client attrition rate has gone down significantly. I t's non-existent. Our large deals bookings are on a historic high. We've built a muscle on large deals. Last year, we did 17 large deals, more than $100 million. This year, in the first three quarters, we've done 19 large deals. I have a quarter to go. I'm excited about the fact that we've made all this progress. There is always work in progress. W e have still not capitalized on the opportunities outside of financial services and healthcare. We have still not capitalized on the opportunities outside of Americas.
We have set a structure up, and we're hoping that in the next wave, we're going to see the growth coming in these new areas where we've invested. We have invested in ER&D, which is a new space for us. We now have a $2 billion portfolio. We are probably the top five players in the market with the addition of Belcan. It's an extraordinary opportunity because engineering services is not about building systems for enterprises, enabling businesses. It's actually embedding technology into products in an era where every industry is going to be a tech industry, only $150 billion, actually $250 billion, has got outsourced so far in engineering services, but the total spend is almost $2 trillion, we do think there is headroom there, there are things to do, but we are set for a phase of growth.
That sounds, W e've written a lot on that whole idea of the operations side, the factory software coming together in products. C ould we dig in a little bit on that as an opportunity area for you? You've made an acquisition there. Could you talk a little bit more about that business and how you see it?
E ngineering services historically was not a big space for tech services companies because tech services companies, as enterprises were globalizing, focused on helping enterprises enable it. Engineering now is embedding technology into products. W e are living in the age, in a golden age of technology, where software is the new alchemy. L et me give you an example. Let's take automotive. Tech services companies in the past were enabling the business of automotive, like building HR systems, building finance systems, building sales systems. Now we have uniquely two additional opportunities. You could put technology into the car for the infotainment inside the car. You could actually write code, which is the core which takes out the internal commission business to a software business. That is an unlimited opportunity. Embedding technology into medical devices. Belcan, which we bought, is embedding technology into the aerospace industry.
T hey do wing design. They embed technology into it. Embedding technology into industrial products. I think there is so much to do. T hat's almost an unlimited runway. It's also countercyclical to tech services. Tech services has a different spend pattern. E ngineering services has a different spend pattern. N ow, for a $20 billion company Cognizant is, we are in all the four vectors of the business: technology services, engineering services, BPO services, and infrastructure services.
That sounds a great way to both get into new industries and to smooth the cycle.
Absolutely. It also gives resilience to the platform because you just have a much bigger spread.
That makes perfect sense. Now, you alluded to the big investment, billion investment in AI, and I guess this is a topic for a lot of investors that we see these huge opportunities that we can automate much more with GenAI than we could with RPA and traditional coding, but there's potentially a risk that at the low end of the pyramid, those jobs either get automated or certainly maybe don't recruit as much as we did, but for you, maybe it's an opportunity to break the link between headcount and growth that's been there in the past. How do you see GenAI as a pro or a con for Cognizant?
W e made very early investments, pretty much like how Cognizant did in every technology wave, invest ahead of the curve and monetize on the opportunities coming on the wave. I would actually put this in three broad segments. The first segment is today and now. One of the biggest use cases for generative AI is tech for tech, applying technology on technology. In fact, this is such a mainstream use case that many analysts feel, is this sector going to shrink because you want to cannibalize your own self? Remember, every technology wave, this question came up. When offshoring happened, we thought this technology sector, the tech services sector, would shrink. It didn't. It actually grew. When cloud happened, we flipped the model. We flipped the plumbing to building ratios. Plumbing was 70%, building was 30%. The cloud flipped it on its head. We do 30% plumbing, 70% building.
I nterestingly, the elasticity of spend on technology just made technology create more spend. Right now, as machines write code, and in our earnings last quarter, I spoke about two million lines of annualized code written by machines. I t is roughly around 13%-14% of the code which we write. I n fact, many other companies have also started to quote, not in our group, but in the software sector, like Alphabet said, 20% of the code is written by machines. Your first impression it creates is it takes away your business. The reality is you could use that as a tool to create more opportunities. You could unlock modernization spend. T here is $1 trillion of technology debt sitting on balance sheets. $500 billion is actually spent on addressing that spend every year. You could unlock it because you don't need the tribal knowledge.
You have the human capital. Y ou don't need the financial capital needed to unlock it. You can take the backlog off. S ome CIOs I've met have actually said, "I want to do more for less." T ake the backlog off. W e have used this as a unique opportunity to win more business and increase our wallet share by actually consolidating and taking over portfolios of our peers in the same client base. I f you're ahead of the curve and if you have the tooling and the techniques to do that, you could commit that productivity and get it. But it is going to give you a runway of cannibalizing, but equally creating opportunities to consolidate and do more. I t is the first vector of growth. It's going to have marginal growth because there is cannibalization.
T his is an improving technology, which means this number of what the machine is writing code is only going to go up as a percentage. Now, let's come to the second vector. We do 1,000 projects on generative AI. Half of them are tech for tech and in the category of writing code. The other half are business-related. They're revenue-related. They're innovation-related. The categories are content aggregation, customer experience, employee experience, content creation, net new content creation. These are the categories we are seeing projects. Now, this, you could argue and say, wait a minute. Aren't the foundation models generating the output? W hat does a tech services company do? The foundation models are generating the output. But that is not business grade. That is not enterprise grade. That is consumer grade. To make it enterprise grade, you have to run the last mile.
You have to put the checks on the hallucinations. You have to make it responsible enough. You have to create traceability of it. You have to modernize the data and modernize the cloud underneath it, and more importantly, in recent times, the output coming out of these models is not ready because you need to put the reasoning layers and the cognitive layers to make it move from what they call System 1 technology to System 2 technology, which is you don't take pre-trained output, but you take pre-trained output, and then you take it through a scenario building and a decision support process and the reasoning process, and already, a lot of the generative AI models are starting to put that reasoning template, but it needs the heavy lift from a system integrator to agentify it, t hat opportunity is unique. It is net new.
That spend didn't exist. Then the third segment of this, which is the most interesting segment, is how do I take this agentification of all the work we do and take the labor pools which we didn't address and start to actually agentify it? Let me give you an example. If you are, say, implementing Workday, the spend of $1 on Workday software has $2 to 3 of services for us. That was our total addressable spend. Now, imagine if I start to take the spend around it. Because once I agentify it, I can take onboarding of employees, recruiting, engaging employees. All those labor pools are ready for agentification. If you take HR software to be set at a $20 billion market, services was $40 to 50 billion. But the size of HR organizations and the compensation and the salary, the labor pools attached is $250 billion.
T he labor market, just the labor pools which I can address is much more. If I take another example, like Salesforce, right? S alesforce is a $35 billion company. The services underneath is probably $70 to 100 billion. But imagine the number of things salespeople do around it. They punch in opportunities. They punch in leads. They do forecasts. They do sales compensation. They do reporting. All of that is $1 trillion. That labor pool is my new addressable spend. T he point I'm making is tech spend of enterprises is no longer the budget I'm looking for. That was the traditional budget I was looking for. The new budget is operations spend of enterprises. Anything below the cost of goods sold is available for agentification. I n my third category from the first, I'm writing software. Machines are writing software. The second is I'm agentifying it.
The third is I'm taking the labor pools which wasn't my total addressable spend. I'm actually going to make that an addressable spend for us. T his progression as it goes makes tech services and at least companies like ours who have invested ahead of the curve to be the beneficiaries of the process. I'm keeping engineering services aside. Engineering services is a completely new runway. It's a runway where you are embedding technology into products. I t is going to have a different flywheel. I f you put all of this together, AI is going to be a huge beneficiary for IT services companies like ours. It will actually differentiate the losers and the winners. But with the framework I just established and the stuff we are doing, we do believe we'll be in the winner's circle.
If we think about those three areas, t he one that sounds to me the biggest is that area of you being able to attack the labor pools around Workday and Salesforce?
That takes the longest gestation period as well.
That makes sense. A part from you being ahead of the curve, is there any other specific advantages for Cognizant to be able to attack those labor pools and other people?
L et me go back to, again, the history of the company, heritage of the company. We always had two categories of firms. There were companies which were built with the industry domain expertise. T here were companies which were built with technology. Cognizant's birth happened way later than the first wave of outsourcing. The first wave of outsourcing was the Y2K boom. Cognizant was born literally after the Y2K boom. It was born in 1998, just when the Y2K boom was over. C ognizant was built for the application outsourcing space, where the domain expertise and the technology expertise and the confluence of these two was the strength of the company. I would believe that strength is now more relevant than ever before.
T he ability to tap onto the third pool, which I spoke about, the third category, needs operations strength and industry domain coming with deep engineering. I think we sit at that intersection. I n fact, in places like health care, we have manifested that into a platform play because Cognizant has a platform in health care where we manage health care operations for companies. Two-thirds of the insured population in the United States goes through our platforms. W e think that operational knowledge will help us to tap into that third pool because it is a confluence of those three things I spoke about.
We could talk about GenAI for hours. But I'm going to get told off if I don't move on and talk a little bit about macro demand and so on. I f we could just talk a little bit about, I guess, clients are starting to prepare and think about budgets for 2025. In terms of your conversations with customers, what are you hearing from them in terms of how their planning is going?
I would say at the same time in 2023, we had less visibility into 2024 as much as we have now into 2025. W e have improved significantly since last time. There is an unlock in some areas. F inancial services have started to come back. Discretionary spend has started to come back. W e were doing large deals. Discretionary was the challenge. Discretionary has started to come back in financial services with the two rate cuts which have happened. S ome of that uncertainty out, L ook at it. Last year, by this time, we had uncertainty of elections, geopolitical situation, interest rates, inflation. Some of it is gone. The election is out of, It 's out of our way. Inflation is much more tamed down. Interest rates, we've made good progress.
I would believe the discretionary will start to build up from here. It'll be interesting to see in January and February how the budgets shape up. But I'm also excited about the fact that budgets is only one piece in the puzzle now. There are more things which are happening beyond the traditional IT budgets. I'm excited about the fact that it is starting to come back. I t's much more stable. But there's more work to do. B y January or February, we'll start to see more visibility into the year. But it is looking much better than what it was last year.
That improvement in financial services and discretionary, is that quite U.S.-centric? Or are you seeing that in other areas?
I t's more U.S. I think it will kind of have a ripple effect to other parts of the world. But most of the business exposure is also to the U.S. in a much larger way.
O n that financial services side, that's been an area of strength for you for a couple of quarters, even if we haven't had those green shoots. I h ave there been things that you've been doing specifically at Cognizant to address that market, to improve things that are in your hands?
I n 2023, when I started, we had both market was soft. C ognizant by itself did not take off on financial services. I think we have now established a strong leadership team. We have put in measures to seize those opportunities. Now that the market has started to come back, in fact, spend on generative AI is one of the highest in financial services. They are the ones who are doing the most experimentation. The regulatory, there's a hope that the regulatory spend will start to go down. T hat will unlock dollars for innovation, unlock dollars for discretionary spend for future work. I think it's a combination of the two, which has helped us. W e have structured better offerings. We have created a bundle with software-led offerings. We have also expanded into regional banks, which are doing end-to-end IT outsourcing.
Fintechs have unlocked more outsourcing than before. F intechs had a free runway on EBITDA. Now that the cost of capital has gone up, they are starting to look for ways to scale at a more economically viable cost structure, which means they are starting to outsource. W e're seeing a lot more traction in financial services because of this.
O ur own CIO polls, a ctually, we had a couple of people from the Morgan Stanley tech department in our Barcelona conference a few weeks ago. They were saying that to date, the GenAI spend has been cannibalistic of other areas of budgets. Are you starting to see that?
We're starting to see that get diverted.
Interesting.
In fact, in many places, banks are also starting to think about, can I shrink the run and underwrite the savings to the build? T here is a re-look at how much you want to spend on the run side to the build side. Can I underwrite those savings so I can actually generate that momentum? Some banks have started to consolidate as well, the provider base, which gives us unique opportunities. In fact, financial services is also the largest sector for mainframes. I f you really look at it, I can show you live a mainframe code, which I can take, put it into an AI algorithm. W ith no human intervention, I could translate that to Python code or an Angular code. L ook at the possibility of taking all of that technology debt and translating that to real value.
T here is so much exciting action which is up there. T he thing is, there is still some uncertainty around the geopolitical situation around the world. But if the trajectory of interest rates and the trajectory of inflation continues to go the way it is going, we hope to see more.
Perfect. I'm aware the time is sort of opening it up to the audience. Are there any questions we can take from the floor? Maybe I'll carry on then. Just maybe turning to margins, could you talk a little bit about the levers that you have to manage margins and profitability as you look into 2025 and beyond?
Early on in 2023, we started this process of a program called NextGen, where I wanted to take cost out. T hat was structural cost below the gross margin, which is primarily SG&A. Now, structural cost takeout is always a once-in-a-lifetime opportunity. I t doesn't happen every time. T he idea of doing that was to go back to the roots of who we are, take the money out, invest back into the future. T hat's how we could invest into AI. I also invested $1 billion into ServiceNow. We bought this company called Belcan. There was margin dilution because of Belcan. I had to compensate for that. T hat was a structural program to do that. W e put this in three vectors. First, how much do you put that back into the margins? How much do you invest into the future?
How much do you share the productivity? Because large deals, when you do, they have an investment window because the money comes later because there is a productivity improvement plan. Y ou invest upfront on the transition costs and everything else. W e put all three into the mix: how much we enable the business, how much we put into the margins, and how much we invest into the future. I f you have noticed, since I've come on board, we improved our margin to 15.1% last year. W hen I started, it was much lower. W e gave a 20 to 40 basis points increase for 2024. W e bought Belcan. It diluted us. But I've got back to the same number for 2024 as a forecast.
We think we can keep the expansive plan of margins for the future, keep the investments going, and continue to stay focused on sensing what is coming. Now, on gross margins, we did three or four things. In the last three quarters, you would have noticed, in fact, in the last two quarters, we sequentially grew. But our headcount has gone down. W e grew by 2% sequentially. But our headcount went down. Now, that is reflective of how operationally we are running, which contributes back to the gross margin, and how much of AI-led productivity we are saving and how much we are sharing with clients. S haring with clients gives us the right to win. Keeping it back is about how much you can do ahead of your peers that benefit remains with you.
W e are able to actually do more for less and improve the margin. W e have now, in 2025, we have a tailwind from the NextGen program we have done in 2024, as well as the operational efficiencies and the AI-led productivity and sizing the pyramid. We think we have a runway on it. I 'm aspirationally wanting to keep the expansion on an ongoing basis but keep an eye on investments of the future, growth is the driver for the change and for the growth in EPS.
T op line is the most important thing. T he investments need to happen to keep that?
I t's a combination of the three we're doing.
Perfect. Appreciate it. Maybe time for one last one, just on pricing. I guess we've seen cycles before where some competitors have been very aggressive on pricing to win volume. I n terms of what you've seen competitively, and are we starting to see pricing stabilize, improve as discretionary comes back and things improve?
R ight now, the pricing is. There is very little to look at. In pockets, there is pricing. Like if you look for deep engineering, there is a pricing. You will get a pricing premium. That's also because the talent pool is much smaller. But if you look at pricing in context to traditional work, there isn't so much. In fact, on the contrary, a lot of our clients are looking for productivity to be shared with them because they know that AI is nonlinear. It is not labor-linked. It is nonlinear. I t's hard to say when pricing we will get a pricing uplift. M ainstream pricing uplift. We should potentially wait for more discretionary to come back till you see that.
But I'm a believer that it will kind of move into an equilibrium when the demand and the supply start to fall in place.
Perfect. Well, we're bumping up against time. I'll close it there again. Thank you very much for joining us.
Thank you.
Very much enjoyed the conversation. Thank you.