Still good morning, everybody. The clock is running, so we are going to go ahead and get started. Good to see everybody. We are thrilled to have Cognizant here today. It is a company I have covered I think that is you, not me.
Yes.
I have covered Cognizant, I think, for 15+ years, so happy to have the company with us today. It is certainly a dynamic time. Let us just jump into a question, if that is okay.
Please.
If you think about, I want to ask a Tyler made sure I was not going to ask any curveballs, but I want to just talk about calendar year 2027. I will ask it in the industry context, and if you think about the industry at large, Infosys, Accenture, IBM, Cognizant, what are sort of the puts and takes for industry growth next year?
Sure. Let me start with the caveat that 2027 is still a little far off, and the priority for every firm or customers we work with is really to get through 2026 next couple of months. But having said that, I will share with you what I see on the horizon right now for 2027. One, the industry, as you know, has been battling with some sort of deflation related with AI on our applications business, and that has been probably the most intense in the later part of 2025 and whole of 2026. And various companies have spoken about where that is in its journey, but it seems to me that a substantive portion of that, depending upon which company you are, has been absorbed in 2025 and 2026, and still industry has remained flattish to slight positive in terms of growth.
And that should become in the later half of 2027, if not in the first half of 2027, some amount of relief to the industry that they are not countering accelerated deflation. That is one. The second, and we spoke about it in our earnings call, that the first-degree impact of AI was the deflation in services that I spoke about, but now we are beginning to see what we call the second degree or second impact of AI to our business, which is some amount of tailwind in some of the service lines. This is the kind of work that we are getting because of AI. It would not be present if AI was not on the horizon. For example, cybersecurity related with AI is an absolutely new opportunity TAM that has opened up because of AI and with AI.
Similarly, modernization of work, of legacy applications, which earlier was considered too cumbersome and too difficult to get through, even in a five-year time zone, is possible now in 24 months, 30 months. We are beginning to see some of those projects. Yes, there is, without quantifying it, how does it play out, but qualitatively speaking, the burden should probably come down of deflation in second half, and energy of new services, which have come into its existence because of AI, should begin to take off. Yeah.
Let me just push on that a little bit before we transition, in that it assumes some static nature associated with the AI deflationary forces. What I mean by that is, if you think about AI models, number one, are going to get better, and so what you can do today will be less than what you could do in the middle of this same time next year on AI models. And two is, the amount of work that you can apply to those AI models or those AI models to work would also grow. Will that deflationary force be the same in 2027 so that the goodness can actually show up for the industry, or will the AI force sort of be incremental, therefore, even if you get some new source of demand, you sort of end up in the same place?
It is difficult to call right now because you do not know what is in the store for you. Our own experience that I can speak about, since we have seen the first big wave around, let us say, coding, I think the incremental impact, even with more advanced model, has been very little. I am not unduly worried about on the same piece, a better model can do significantly better sort of deflation to the revenue. To your other question saying, can it expand to a different type of work, which it is not applicable today, yes, that remains a risk in our business as any other business where AI could play a role, and that we will have to watch out for.
Okay. Let's transition to a couple of things and what is a proxy for growth. As we think about trailing 12-month booking grew 4%, is that a good proxy for growth as we think about next year versus, and/or where your guide is for Q4, or just think about, help folks think about proxies for 2027. What do you guys think about, or what would you point investors to?
I would start by saying that the more growth is better by design, so I would not say 4% is adequate or great. Having said that, we had three mega deals in 2025 which were sitting as a base for 2026. So if you take into that consideration, you also take into consideration that while revenue has a differing magnitude of how much of revenue is getting impacted by AI deflation, almost 100% of bookings that comes through goes through an AI cycle. So bookings is in an earlier cycle of AI deflation compared to revenue cycle, which still has tailwinds to go. So to that extent, I feel good that we are still showing a positive trajectory despite a much larger base, with three mega deals in the base, as well as a full-blown impact of AI deflation that we have taken in 2026.
We feel we have what we need to deliver to the guidance for this year, and we will come back to you with what we think would be our revenue trajectory for 2027 in the beginning of January.
Okay. Maybe talk a little bit about what the conclusion should be on new and expansion bookings have become a larger portion of the mix. How do you think about that in terms of, when you try to construct the growth of the business, how do you think about the changing nature of mix?
Absolutely. That is an excellent positive statistics that we observed and we shared with investors in the Q3 earnings. Essentially, we have seen a 50/50 mix of renewal and 50% mix of new business. It has shifted by about 10 percentage points in favor of new business. Why? That comment was made in some form to support the fact that we still see growth despite a 4% TTM bookings growth. The reason is, when your renewal, because of deflation or even traditional productivity levers, your revenue of $100 becomes $90.
Right.
You have a deflation of $10. When you have net new, you have entire new of $90. It is, so to say, almost 1:6 or 1:7 times superior revenue accretion that takes place when you have renewal versus new business. Since we saw that first time in that shift, we wanted to share that as a comfort metric to the headline metric of 4% TTM, which obviously looks lower than what we would like it to be.
You think about, just focus on that gross retention number, as we say in the software world. Historically on renewals, you started, if I go back two or three years, it started $0.97, $0.98 on the dollar. Now you are starting at a lower number. You mentioned $0.90 on the dollar for that renewal. Is that representative, and has it been stable, like over the last quarter or so?
That has been stable over last couple of quarters, for sure.
Okay.
It has, as you know, accelerated vis-a-vis where we were in 2022 or 2023 when there was no AI deflation. But from that step down, that has remained broadly stable since beginning of this year.
That, leading question here, but that's probably one of the reasons why you're more encouraged about 2027. If that stays stable at 90%, that's a key assumption.
That's correct, yes.
Okay. Financial services has been really good for you guys. What investors look at is the rest, other industries going to migrate to financial services, or is financial services going to migrate to something that looks like some of your weaker industries? But how are you thinking about the proxy that financial services may represent as you look at calendar year 2027?
Traditionally, we have always seen financial services as a lead indicator for demand, for three reasons. One is, they represent the largest percentage of IT spend as percentage of their revenue. They have It's not one additional procurement. They are very purposeful about that spend. 10%, 11% of revenue comes through various form of IT and related spend. Second is that they read the trend early, precisely for the first reason. They are first to get onto something which is meaningful faster than others. And third, they have a view of the world, meaning, they know where the softness in the world is, and they would be cautious if they saw a certain segment of wholesale banking or certain sector which is going to come and hit them as a non-performing asset, et cetera. They will be cautious in the spend.
The fact that they are leading with the spend is a good indication for IT services industry that they see this AI, which is probabilistic system and deterministic, classic deterministic system at the interplay, something they want to invest in. And there is clearly discretionary spend that they are spending on. It also tells us one important thing, that when the discretionary spend comes back, the whole conversation of AI deflation is little bit muted because BFSI is growing 12% for us even after the deflation. It is not that there is no deflation in BFSI.
Deflation is sort of uniform. But the moment the willingness to spend comes in, they use the deflated dollars to put back into IT, and therefore we are growing. So we like it. We believe that hopefully the others will follow BFSI.
Okay. I want to jump ahead to one thing on modernization. I'm always intrigued by the comments around modernization, and Infosys talks about it, TCS talks about it. What is modernization? Because I always think about mainframe programs. But maybe give us a little bit of characterization about modernization. What are clients trying to accomplish, and how much is it driving growth?
Yeah. Modernization is essentially any application which has been built on protocol, code, architecture, which was 10, 15 years back. It has multiple issues. The traditional issues which have surfaced is inflexibility to do a lot with that code, so people have built code on top of that code. People have built sidecars of databases to use this data and do something on top of it, and so on and so forth. The reason that has happened is because the whole application is going to take a long time to be reconstructed in a new language or a new architecture, and really no CXO would want to undertake a bet like that for five years, which is starting from here, the project becomes successful in five years. People have used in past modular approach, but that has been on periphery, not in the core.
To your point, it is of course mainframe, but it is also industry-specific. Banking financial services is very well-known for having core technologies which have remained unchanged for many years. But if you look at airline industry, if you look at retail industry, if you look at travel and hospitality industry, even telcos, where the core operation of connection and text exchange is taking place, that's really old technology. This is, I mean, one of the big projects that we have won is in telco space to modernize the heart of their technology. It's no periphery, it's really the heart. There we are working with Anthropic jointly to deliver that. It is a large opportunity by any magnitude.
Just, I'll go back up to macro, but I want to go micro for one second, because I've had the conversation with a number of your competitors about mainframe modernization. As you think about that as a business, are you helping customers get off of mainframes and/or basically migrate from COBOL to Java? What's a bigger part of that mainframe modernization?
It is every technology, and I don't want to name a particular technology, but I would say it is every technology where the owner of the technology has made the customer feel entrapped in a technology where the price continues to rise of that technology, the additional features are very limited, and it's really to maintain your operation, you need to pay a 10% premium to us. There are host of software companies of past which are now being run like that. That's an opportunity. It's not just a code of 1995 that we are upgrading. We are even going out for new technologies for something that was done in 2005.
Okay. All right, I'm going to ask two more questions and open it up to the audience. If I want to back up, I'm just going to go macro. It's November 12th today, and just, are your sales reps giving you any feedback on what budget cycles look like next year, and/or what's sort of bubbling up to the top or the bottom? What investors are really concerned about, and I've had this conversation with our internal teams at BMO, about how are we going to fund AI and how are we going to fund security? There's not enough money. But what are the characterizations of budget cycles that your sales reps are giving you about next year, if at all?
I would say they're broadly neutral right now. I don't see a negative tone. A negative tone is reflected not really as a next year cycle, but saying we had originally predicted $7 million-$8 million on furloughs, but now we have to account for $12 million because customer wants additional two days at furlough. That's beginning to tell you which mindset customer is and how we are going to meet customer in first week of January. The good sign is sometimes people pull back on furloughs saying, "We need to deliver this project because customer really wants to go live in early January." So I'm not seeing either, is where I am. I see, as we predicted at the time of our earnings, more or less same trend continuing.
Okay. Frontier models are building deeper enterprise deployment capabilities. Where's sort of the natural boundaries that you think between Cognizant and some of the frontier models, and where is there perhaps growing, where is there some incremental competition?
We actually see it as a great proof point of the fact that it is not a switch on, switch off technology, and you will need a trusted partner to convert the technological prowess into business value that can get delivered to the shareholders eventually. You can't rely on just buying something and getting deployed. In fact, a lot of customers have done that since beginning of the year, and now pulling back and scratching their head to see how are they really getting value. So it's a great proof point to IT services industry that you need to handhold this technology to create value for customer. Now, question is, who's going to handhold the customer? Is it going to be the organizations which are partners or which are stood up by frontier companies, or it's going to be a third-party providers? We have always seen that.
SAP also has a consulting arm. Almost every large software company has a consulting arm, which they use for the sort of very difficult case studies. But rest of the mass value generation takes place to third party because we are seen as an advocate of customer whose focus is to minimize cost and therefore token consumption and maximize the value for customer. Whereas if I am an implementer for an LLM provider, my interest would be to maximize the token for my parent company. I think both of us will coexist, and they will attack maybe the tip of the iceberg and tip of the spear, and then we will follow through with the body.
Okay. A micro question, going back to micro for a second. There's a little bit of change in terms of the third-party resales in Q4. Just remind us of what that is. I think that's reselling computer equipment and things along those lines. Importantly, how should investors think about that as we're tuning our models for 2027, the potential upside or downside from that particular line item?
I would say there are four components which go into what we call as integrated offerings. One is infrastructure, second is SaaS, third is security software, and fourth is inference. The reason we are now trying to bundle and deliver is we want to go towards getting to an outcome-based contract or an outcome turnkey implementation.
This is not something that we are doing first time when every new element has got added, including when cloud was added first time. Cloud was going through our books, and we were implementing it for customers. Specifically for quarter four, I would say maybe the contribution would be similar for what it has been for first nine months, and it can go slightly up or down depending upon the type of work that we are doing, and the type of work is how much of the projects where we are doing a turnkey engagement, and those have gone up beginning of this year. Now, 2027, I think the other components may not materially move, but inference is key. If inference goes through our books, you would see that component going up.
Okay.
It would be good because then as a service provider, we are undertaking human effort and virtual effort and delivering an outcome for customers. If you really ask me, that's the future I see of the industry, and that is what we are working towards to make sure that we are channelizing inference to create an outcome for our customers. It would be a great thing for industry if the inference starts flowing from IT services industry.
Okay. Let me open it up to the audience if there's questions. Okay. Brad, go ahead.
Yeah. You mentioned-
Okay.
Hi. Thanks. You mentioned more outcome-based pricing. I want to ask about that. I think you mentioned 58% of your revenues are now fixed price, and that continues to rise. How are you building that into your delivery and into your pricing? What should be the expectation for that mix over the next 12-24 months? Then ultimately, how does that tie into your employee productivity as you think about some of your per employee metrics?
Sure. I think there are two parts to it. One, the current rise in fixed price contract, as we call it, I would still call them fixed price contracts versus classic outcome-based contracts. I will come broadly to what that outcome-based could look like. Outcome-based contract is, I will underwrite a mortgage application for you at the standard that you prescribe. That's an outcome-based project. A fixed price project is, I will finish this implementation for X dollars in this milestone. So that's the difference.
The 58% that we spoke about is really fixed price and not outcome-based. Outcome-based is very tiny percentage for the industry and for Cognizant today. But we aspire to grow there. Fixed price, the fact that customer has moved from time and material to fixed price is half a step already. That could give us comfort that we will eventually continue to grow there. My own sense is that 58% should continue to inch up every year as we progress more.
I think about the risk profile of that. I used to work in the construction business a long time ago, and fixed price contracts can make money and lose money. When they lose money, they can go bad. So how do you think about the risk profile of growing that fixed price contract? You have to get the cost curves right.
Yeah. It's definitely riskier. It's definitely something that you need to have a close watch on a monthly basis. You need to have a tight governance around that in terms of deployment of resources and squeezing out resources when you don't need them. So it's a slightly different or more mature way of delivering services, but there are a lot of industries that like it. Manufacturing loves it. Consumer products love it. Health loves it. BFSI has remained traditionally time and material. They continue to remain in time and material business.
Okay. Let's transition a little bit to some of the AI and the lines of business. You've talked about AI tailwinds represent 40% of the portfolio and is growing. How does this transition unfold? Meaning, when you look at calendar year 2027, I assume that that 40% migrates up, delivers growth, but what happens to the balance of the 60% portfolio? Is there compression there? Maybe just talk about how that mix contributes to growth next year.
Yeah, I think this is again the numbers that we shared first time in quarter three, because we saw that now we are seeing the second order impact of AI, which is the AI tailwind new services that we can sell because of AI. We are beginning to show that sliver separately so that the investors are able to see that the average performance doesn't hide the early success around AI. We think that growth will continue to take place. We have shared four service lines around that.
BPO, infrastructure, cybersecurity, and industry solutions. Industry solution falls in that modernization category, where we work with a particular industry to modernize whatever they are doing. We recently took over a very large insurance platform, which we will modernize and deliver services on. These four services we definitely see are beginning to see tailwind around it. Now, what happens to the 60% of the business is something that we see closely. My own sense is that the risk of AI deflation is going to moderate more and more as we move forward on that 60%. You will see some part of that 60% beginning to come into the second-degree effect, and we'll see new demand that is coming because of AI. That's why we started calling both out separately.
Okay. As you talk with the customers, how does sharing happen? What I mean by that is, AI benefits accrue. I know at BMO, we look through our contracts in detail with TCS, with yourself, Infosys. How do you think about governing the behavior of your sales reps and the clients to how much sharing goes on because of the AI benefits?
Essentially the way industry is operated is that what is visible, a large chunk of that visibility at the time of contracting will go to customer. A smaller chunk will get retained with you as a contingency against any risks that you are undertaking. But where the IT services makes money or has comfort about this model is that visibility around what you can do new continues to improve as you execute a contract. For example, the contracts that we signed in 2024, I underwrote certain productivity goals. They can be more than comfortably met in 2026 because the technology has really progressed. So, what is visible at the time of contracting, really a large chunk goes to customer in terms of gain share or their productivity lever to sign, underwrite a large contract. But as we progress in fixed price bids, that stays with us.
If it is a large T&M contract, even that flows to customer, and that is what happens for, let us say, large BFSI customer.
Okay. Do you think the nature of that sharing arrangement changes in 2027, or does it remain fairly static?
I think it will change when we have inference baked in, and it will not remain what it is today because today's arrangement is based on the current human resources model where if customer wants to cancel a particular contract, all I have to do is to ramp down those teams and find another project I can deploy. Worst case, almost any part of the world, I can reduce that cost with whatever restructuring effort I need to take in 90 days, 180 days at max.
Right.
There is no residual cost associated with an earlier decision beyond six months. It may not be so with inference embedded. I may have bought something, customer have bought something with a commitment to an LLM provider for three months, three years, four years, five years. So, that setup cost is going to be higher, committed cost is going to be higher, and therefore that risk-reward model has to be different, but it will evolve as we go.
Okay. Let's go through a couple things. I met with Ravi a few times and Tyler in the last 12 months, consolidation was one of the key themes, i.e., a firm like Bank of Montreal, we are in the process of doing this. We are buying IT services from fewer vendors. How do you think about this trend in calendar year 2027, and how does it impact Cognizant?
I think the trend is a cycle for every customer, where customer consolidates and then over a period of three to four years, various departments, a new acquisition, a new initiative which starts in some part of the world proliferates the vendors. Then, a senior leader comes from either IT, procurement, operations side and says, "No, this is not right. We need to work with fewer vendors." So, it is not a one-time phenomena, but there is a certain amount of opportunities which present itself every year. So, I don't see that changing in 2027 compared to 2026 or 2025. How does it play out for Cognizant? I think it plays to our strength. I think we, almost every time there is a consolidation, we have been able to win share. Probably that's the reason why our BFSI growth is 12%.
It's partially the discretionary of a customer base, but it's also because we have won share from others and it's helping us grow. So I quite like that phase when customer comes and says, "I only want to work with two or three vendors," because almost invariably I know Cognizant should be able to stand on the other side comfortably.
Yeah. Yeah. How do you think on pricing? Because consolidation sometimes brings pricing, like any other company, you get better economics if you buy more from that company. With the advent of maybe consolidation remaining steady, AI growing, how do you think about pricing trends today? How are they different from, say, last year, and how do you think about pricing trends as we look at calendar year 2027?
I think 2026 has been probably the most intense year on pricing in over last three years. That is probably because the industry saw very little, at least in the first half, the impact of AI was very meaningful.
Yeah.
Therefore, the growth was scarce, and therefore, firms were competing to retain share or gain share. That's the phase we are in. We'll have to wait and watch what happens in 2027.
Yeah. It'll be interesting if you're right that retention stays the same, so there's no incremental deflationary force. If that works, if that plays out rather, you could even see a less pricing pressure, right? So you get a combined benefit. Now, the inverse is true. If you're wrong on that deflationary force, that 90 goes to 80, then pricing will compound that as it has this year.
That's correct. That's correct.
Okay. BPO has been an area that's surprised us a little bit. If you had asked me three years ago, I would have said AI would destroy BPO. But in fact, it's been very strong for everybody but IBM, actually. How do you think about the durability of BPO growth as you look out over the next couple years?
Look, we see continued momentum. And the momentum is because there are far more opportunity around agentic that organizations were not leveraging before. I'll give you a simple example. If in a traditional business in 2024, you had 500 people in Romania or Philippines and continuing to do work on some amount of transaction processing, no one in that company will think about reducing cost further there. People will move on to other opportunities. Now, it's reverse. If you have 500 people who are working in Romania, I think it's a great opportunity to deploy AI and take out substantive amount of cost and maybe give it to a third-party player if you do not have prowess to go about doing it. So, in effect, we see a lot more TAM being open because of AI versus the compression. Of course, there's a compression because of
Yeah
the impact of AI to our existing revenue stream. But the opportunity that we see on the horizon is higher, and therefore, you see almost everybody in industry growing at almost pre-COVID numbers or
Yeah
We see BPO growth quite sustainable, to your question.
In other words, the TAM opening up because of AI is bigger than the deflationary force in BPO.
100%, yes.
That continues in 2027.
Yeah.
Okay. Okay. Let me go to next set of questions, but let me pause here. Does Shital or anybody else, not cold calling, but have any questions? Brad, did you want to jump in on anything? I will keep going then, unless there's volunteers. I ask every IT services the same question. I usually end with it, but I'm going to jump in here. When you think about Cognizant in three years, does Cognizant have more employees, fewer employees, or the same employees?
Let me start with what I see as servicing, and let me then come to number of employees. I think because it's very intuitive once you understand what is my vision of what I do. Large companies like BMO, like JP Morgan, Bank of America, will have 70% still deterministic systems because it's a core arbitrage that you can't get away from. If you run an SQL query and in a nanosecond get an answer, you will still leverage it to answer that question versus asking an agent to go through 350,000 records and answer that question for you, right? So there is a clear value that we see for deterministic system that will sustain.
So you would have a scenario where 70% of the system is still deterministic, 30% of the system would be probabilistic, where you are replacing now with human experience, judgment, and whether a transaction should proceed or not proceed with probabilistic systems around agentic. It will be a single workflow where, depending upon what is optimal, a deterministic system will play a role or an agentic system will play a role. Now, we see us providing both to our customer.
Because we come with a I mean, the feeder for AI is the context, and context is sitting, of course, in people's mind, but it is largely sitting in workflows, in databases, in application logic that has been built today, even beyond the SaaS software. So if we are providing that, we will naturally have 30%, 40% of work being provided by virtual agents, which is inference and tokens, and 70% of that work being provided by humans. Now, if you believe that we become 1.5x today's size, then the 70% of 1.5x is going to be greater than today's 70%. Today's 100%. So that's fine. I mean, you will still grow as humans because you have expanded what you do far larger, and you will still need 70% of that to come from employees, human employees.
If you believe we will grow only 15% in five years, in aggregate, in that case, that human workforce will shrink.
Yeah.
There is no choice around that.
Yeah. I mean, I think that 70/30 analogy too, you need fewer people to deliver that 70.
Of course. That's right. Why not? Absolutely.
Okay. Maybe as we are coming down the home stretch, I am going to go a little bit to margins. On one hand, price pressure has created some tension on margins. On the other, I think AI is actually a help to margins because you are delivering with greater efficiency, and the customers are allowing you to keep some of that, giving a lot of it back. But how do you think about the puts and takes of margins, particularly gross margins, as we look at calendar year 2027 and beyond?
It is going to be a combination of business mix. I think if everything remains more of the same, then there should be a slightly positive tailwind to gross margin because of greater offshoring, because of improved pyramid, because of AI productivity not passed on to customer, because pricing will take away some productivity, let us say 80% of it, but 20% you may be able to retain. So there could be. But if the business mix changes and I am passing inference through my books then the growth could be very good, but gross margin will look very different.
I do not know what they would be, whether they would be higher, lower, I do not know, because today we do not have that scenario. So I would see that is something that will evolve. More of the same, there is propensity for us to improve on the gross margin.
Okay. We are at 4:00 PM, so with that, I am going to thank Cognizant for coming to Bank of Montreal's Tech Conference. We appreciate your time, and we wish you all the best of luck. Thank you.
Thank you very much for having me.