Fireside chat · 12:10 to 12:35 PM · Thursday, May 14, 2026
All right. Happy midday. Stock market's doing great. Depending on who you're asking. Fantastic. Anyway, before lunch, I'm honored.
Stock market's, you know, voting machines in the short term and weighing machines in the long term.
Yeah, it's completely unfathomable right now what's happening, and now they're all getting excited about Trump and China having some dialogue. I think Cisco just went up 17 points and laid off 4,000 staff at the same time. Go figure. So anyway, I'm joined by Ravi. A couple of things about Ravi, which I'm learning more about, is he started life as a nuclear scientist. And he was recently elected to the Times 100 AI list. There you go. Two things over there. He's also the CEO of one of the largest IT service providers right at the coalface of, I think, everything that's happening in our industry right now, having a lot of dialogue with folks like Ravi, and you heard from Frank D'Souza yesterday.
I have two toddlers at home, and every time I call up my wife and say, can I go and pick up my kids, she says, oh, yo, has the stock price gone down today? Because then she figures out that there's nothing you can do. You can shut it down and then go and do things which you can do normally.
Exactly. And, you know, let's talk a bit about, you know, we've been talking about outcomes and who's underwriting them, but, you know, where is Cognizant today on taking commercial risks with client outcomes and what has to change internally for you to scale that model, Ravi?
Yeah, so here is what we did as an industry for the last 30 to 40 years. Way back, we were systems builders. We built systems for companies. Expertise was the asymmetry. Companies saw this as a non-core thing for themselves, so they basically outsourced it to specialists, if I may. Then came the offshoring wave where we kind of built capability at scale. We delivered more agile, more predictably at a lower cost. We built a pyramid model which our clients could never replicate. So they said, you know, let's hand this over to an outsourcer who will do it better for us. Why break our heads on this? Then came enterprise software. It was repeatable in nature. We got changed in our first principles. We got to being system integrators. Then came SaaS software, the plumbing turned on the head. You know, we were all builders, so as plumbing turned on the head, as builders we got more to build, more software got consumed, system integration was a roaring business.
Here we are now, where we have a technology which is not just deterministic, it can actually mimic human labor, it is action oriented, it is multi-modal, it is having advanced reasoning. So the first principles of companies like ours have to change. We go from system integrators, we were system builders, system integrators, to AI builders. Work is gonna be more bespoke. Work is bespoke because it's not deterministic, it's more probabilistic. So what was the underpinning of this industry was writing software development cycles, writing code. That's gonna change to managing the context layers to get this probabilistic technology grounded in the heterogeneity of companies. So that's a change in the first principles. So it's gonna actually be more bespoke than what we think it is.
Second, historically we kind of had this pyramid as the asymmetry. We built engineering capability at scale significantly cheaper than others, and therefore we were this, that was the defensible moat. Now interdisciplinary skills will be the defensible moat. That's the second change, reforging of the first principles. A colleague of mine and I wrote something with the Fortune magazine, Andrea and I, and that's the reforging of the second principles, which is primarily you take a capability set and you create an interdisciplinary capability set at the intersection of technology, which is supposed to be AI now, domain and operations.
The third, we kind of ran this whole engine on services. We're gonna see platforms and services forged together. You know, SaaS companies would be pushed to doing, not selling per seat, not selling ARR, but selling outcomes. And companies like ours will be forged to sell outcomes. The fourth change, reforge of our first principles, is to underwrite those outcomes. What I mean by underwriting outcomes is, we need to either own the AI outcomes or business operational outcomes. So I should be able to go and tell a client, you know what, I will take care of your customer service. So as a company, we do a ton of work in healthcare. We'll take care of your healthcare operations. We'll manage lives. Focus on underwriting, focus on acquiring clients. Don't worry about healthcare operations. That's not your defensible moat. We will underwrite it. We will run those operations for you, which means you need digital labor, you need human labor, and you need to intertwine it, and you need to own those outcomes.
If you own the outcomes, what that means, what Phil asked for, you will have to, it's a higher risk, higher return business. Because when you underwrite the outcomes, you will be paid on those outcomes. You will have to predict it early on how these outcomes are gonna be. And if you miss it, you're not gonna make money, the penalties are gonna be huge. So that reforging, these four first principles have to reforge. And I'm absolutely sure some players will be winners on that side, some players will be losers. I'm pretty hopeful we will be in the winning lot. But that is the shift the industry has to make to stay relevant. There is a ton of work to be done, but you can't do it by who you are. You can do it by reforging yourself. There's a lot of heavy lift needed. I don't believe that a bunch of forward deployed engineers for whatever we call it now is enough to make it work. I mean, forward deployed engineers were built at a time, I mean, it came from Palantir, it was built at a time when the worldview was they are broken systems in companies. We will come and stitch it together. We'll give you an integrated platform. You take it and you run with it. The agency's with you. That worked. I mean, that really worked because the world view was right. The world view now is manage AI outcomes, which means you need to be an operator who is managing those outcomes. You're running operations for the companies and you're intertwining technology and operations with different domains to get there. That is the secret sauce which will keep us relevant in the middle of this transformation.
Yeah, and do you see, I think we're just at the beginning of this AI operations phase and you've seen, you know, Anthropic and OpenAI launched our own initiatives. I even got some noises from Gemini today about, we need to talk to you as well, sort of thing. These guys are panicking because they've sold the dream and only 10% of enterprises have even got close to getting there. Surely this is going to be the golden age of services now in terms of helping. This must be the golden age of services coming. But, you know, Wall Street isn't seeing it. Yeah, so look.
So first of all, I forgot to let you know, Phil is one of my favorite people, so every time he calls me, I show up. Yeah, so look, the drift to enterprise value, and the velocity at which the drift has to happen has not happened yet. There's a trillion dollars invested in AI infrastructure. Consumer AI is happening fast. I think that drift of value will happen because consumer AI is casual, it is not heterogeneous, it's not mission critical. Enterprise AI has to be grounded in the heterogeneity of enterprises. That drift will happen on the last mile. I mean, both these companies have fascinating technologies but they have not grounded the technology in the hustle of companies to get real productivity, to build new products, to build new services.
You know, I was telling somebody a few weeks ago that I wish this technology was kept in the can and we shut the lid and we kind of, you know, brooded a little more and then we opened it. Then we would have not been going behind which jobs to eliminate, but we would have been going behind what the new things we would build. I mean, go to the Sloan Kettering, you will still find cancer treatment being done like the same old way. They brutally put radiation on the thing which is cancer infected, like a Stone Age treatment. I wish we used this technology for newer purposes, human endeavor couldn't capture. But here we are where we are.
So we've opened the lid, so now we're going, and the economic uncertainty around us also puts us in a spot where instead of using it to amplify, create more throughput, if growth was the imperative, you would say, how can I get more throughput with the same set of people? How can I break constraints? And therefore, let's use AI to do that. Now what we're doing, growth is uncertain. Let's eliminate and create value for earnings. This is one of the most fascinating platform shifts in the history of mankind. So we should be able to pivot it in that direction, hopefully with growth imperatives coming on. So right now, the technology is not grounded in the enterprises, it has to be. Services are very important to ground it.
The metrics on how we measure this will change completely. Today we measure it on token maxing. I mean, token consumption. I mean, how can token consumption be a metric? It's ridiculous. Uber stated in their earnings that they exhausted their tokens on the fourth month. Who's paying for it? The venture capitalists are paying for it. The private equity is paying for it. So, if I'm a services company, what am I supposed to do? What was our craft? Our craft was we would use human labor to deliver things which more efficiently, more effectively, more predictably and cheaper. I should be able to create a harness on the token flows into my company. That harness will have the rules, it will have the repeatability, it will create the effectiveness to do it, and I have community knowledge. I've done it for a thousand clients. Wrong example, if I go to a plumber, I don't do it myself because that's an expert. So tokenization has to be integrated in your workflows in the company, and the right metric will then flip to what is it that I could do more efficiently with the same tokens, equivalent to human effort.
What is it that I could do integrating it with human effort? And therefore, my clients will come to me saying, you know what, I'm wasting my money on this. I've anyway outsourced my human labor to drive agentic to a company. Let me hand over the token spent to them as well. They will effectively do it. I have now AI-infused rate cards in the company. A0, absolutely manual labor. A1, human labor verified by machines. A2, machine labor verified by humans. A3, completely autonomous labor. We should be able to integrate this into the hustle of the company, we should be integrated this into the flows of the company, and we should be able to make this a defensible moat for our clients to say they will be able to, provided tokens are no longer commoditized. If tokens are commoditized, then cost is not the issue. We don't go to a consultant to look at our electricity bills because it's a free-flowing utility. However, it is gonna be expensive, but even if it's not expensive, for a moment, let's assume it's commoditized. We will see open-rate models coming in faster. How you predictably do it, because it is a probabilistic science, it is contextual. How you effectively do it, how you increase velocity, all of those are gonna be the craft we have to build for which our clients will actually give us services to do which are integrated between digital and human labor.
That is the thesis I have on the strength of companies like, I mean, why does the autonomous driving of Tesla better than you as a human labor? Not because it's a machine, because it has the learning of millions of drivers versus the learning you yourself have, and therefore it should be better than you. It's the same. You should come to us because we have built this harness. And by the way, when we built this harness with people, our clients would come and ask the same question, how many places have you done this before? Can you show me resumes? Can you give me a client reference? That knowledge was mobile. Right now that knowledge would be, it would be actually buried in the company's ecosystem because it is permanent. And once you build that, the flow of this tokenization through companies like Cognizant will be much superior to you doing it yourself. So that's gonna happen for sure. And I think we're building that craft.
And as we build that craft, I mean, I'm not saying the science is unreal. I'm saying the way you deliver it, the way you get productivity, I think is being underestimated. It's also being undervalued today. And that's why the markets are kind of, they don't understand this. And when they don't understand, I mean, the thing about stock markets is modern finance, modern capitalism has been built with three underlying principles, two or three underlying principles. One, the future should be legible. Two, tomorrow will mimic today. Three, we should be able to take long, slow bets. That's what modern finance is built on. If you take that foresight out, modern finance has no answers. So the first thing they do when there is a fog is they take the money out. So I hope that fog clears faster. But I'm absolutely convinced that the bridge to enterprise value will come from services companies. How you forge it is different. I mean, it could be a forward deployed engineer. It could be what I said, which is like a frontier operator. It could be, for a moment, let's assume the two frontier companies build hundreds of thousands of people to deliver it. Or it could be companies like us who do it. You could argue on that point. But what you cannot argue on is you don't need anything to do it.
Yeah, it's just fascinating thinking about where this is going. And I know you mentioned something about this, but there's an emerging view that the way we tax companies by taxing labor more than capital is unintentionally accelerating the shift to AI agents. So do you agree the system is structurally pushing firms towards agents over people? And do we need a temporary rethink on taxation to avoid a shock on employment?
You know, I wrote this article along with my colleague, Andrea in the Newsweek, you should read it, which is kind of, if, what was the headline? If capital can think, who pays? And look, capital, fiscal infrastructure in developing and developed nations, 70% of the revenue governments collect is from labor. Labor taxes, payroll taxes. So you don't tax the capital because you expect capital to create more labor. You expect capital to increase wages. So give a freeway to labor, catch the money at the end. That's what governments do all over the world. So we wrote this article together, it details this out. And every time this has worked.
But if capital is competing with human labor, which is what it is doing today, because capital is generating this digital labor, and digital labor is competing with human labor, and human labor is getting taxed, and digital labor is not getting taxed. So if capital is thinking, who pays for it? That's the topic. So clearly, look, I'm also not saying you should tax the capital, that means capital, there'll be flight of capital. It's gonna penalize capital deployment, which means you're not gonna create more jobs and more shared prosperity. I'm saying it's a temporary bridge. A temporary bridge so that you use capital to amplify and augment human capital versus eliminating human capital. And as you get to that point where you think that augmentation is happening, you take the bridge out. You take this temporary relief out. So it kind of gives you a little bit of a smoother ride on the transition to the other side of the world.
And remember one thing. There's one more Time article we wrote which is about AI drifting to the worker at the bottom. It's not necessary that if capability drifts to the worker, value drifts to the worker. Value doesn't actually follow capability. Value follows controls. So what is happening today, the value of AI and the capability of AI, it is drifting to the worker, but it is captured as margins in enterprises. It's captured with the capital, the entity which is owning the capital versus drifting back. So if you have to design this in a safe, smooth landing, you have to revisit fiscal infrastructure. And that'll help you to land safe on the other side and help you to use this to augment human capital versus eliminating human capital in the short run. So I do believe we need a revisit. If capital is not creating more jobs, if capital is competing with human capital, we should have a fair game. Then it should have the same weight, it should carry the same burden as human capital.
Good. Okay, so if we look at three years from now, what do you think will separate the firms that scaled AI from those that talked about it? And what will they have done differently that others didn't?
And what was the last one?
And what will they have done that others didn't?
Look, you know, I'm a big believer that there were first principles on which this industry was built, the IT services industry. Those first principles have to be reforged. We need to become these bespoke AI builders. We need to, I'm absolutely sure 80% of the input factor was human labor. Now human labor is only one input factor into the mix. You have digital labor, you have interdisciplinary skills. Asymmetry is not from expertise, it is from how you apply intelligence and how you apply it in an interdisciplinary way. I have two toddlers at home. They don't ask me this question yet, but if they ever asked me what should I do in the future, I would have said build capability at the intersection of technology and the domain.
The third reforging of first principles is the business is gonna be platforms and services. Clients always wanted this, enterprises always wanted this, but this is a moment for us to shift and do that. The agency of outcomes still lies with customers. We should own that agency, and it lied with customers because we had a distributed value chain. Somebody sold the infrastructure, somebody sold software, we sold services. You couldn't have aligned the accountability. Now you're integrated, so you can align the accountability. And finally, I mean, it is a platforms plus people business. And finally, underwriting, having the courage to underwrite outcomes, AI outcomes or business operations outcomes. And that is the new company, which is, you know, if you want to build that new company, a category one company, that is the new company you ought to build. Every time there is a disruption, there's a need for a new kind of companies, and there is a need for a new kind of people. And that new kind of people are this frontier operators or people who can use AI technology to run operations for companies, run mission critical operations for companies. These are the reforging of the first principles. And whoever actually reforges them fast will actually be winners. They'll be big winners and big losers.
Thank you, Ravi. I don't think anyone could have articulated the change that our industry is about to go through better than what Ravi's just done. Do we have any questions? Maybe gentlemen at the front.
Thanks for that perspective, Ravi. I just wanted to ask that, like, in the last couple of weeks, we have seen a lot of these frontier technology providers starting their own services arms, right? Or acquiring companies. But then you have talked about AI velocity. You know, all these investments not leading to outcome, 10 to 15% pilots really getting ROI for the companies. So where do you think is the gap for these, you know, enterprises to get outcome from the AI investments? Is it because of the AI velocity gap? Is it because of orchestration? Is it because of governance?
It's a combination of things. First of all, look, you know, the good news is they've all launched, so they're endorsing that services are needed. Right. Okay. Now, every time something like this happens, we always look at a successful set of companies who have done it, and we'll say, oh, we can do this. We don't see a set of companies who have fallen off. We always see the ones who have been, so it's always an opportunity. Now, when the cloud revolution happened, hyperscalers wanted to be in services companies, build services arms. They haven't. And they're not making an attempt again. Why would you actually break your head on services? It's a lower margin business. The hustle of services business is very different. It's very different to the innovation hustle. It's applying innovation, it's not building innovation. Can you straddle the two worlds? Maybe. The probability of straddling the two worlds is, I mean, the chance of winning is much lesser. You could actually lose the anchor on your innovation if you're actually focused on this.
Will there be arms of this? Of course, I mean, there's professional, I mean, Oracle wanted to do consulting way back in the 90s. Do any of you hear Oracle Consulting? So do software companies have professional services? Yes. But the universe of services is not that small subset. It's a much bigger subset. So I'm a strong believer that that is needed. What is needed is different. And that's why I don't think what we are is enough to capture that opportunity. We have to reforge these first principles I spoke about. The capability set is different. What is needed to bridge that velocity gap? I think this is a contextual science. If it's a contextual science, it's not gonna behave the way you want it to behave. It's designed, it's actually designed to be contextual, which means you have to ground it into the heterogeneity of an enterprise. We have believed on a science called context engineering, we have been working on it. That's an important reason.
The second important reason, the economics. I mean, there's already quite a bit on economics. Economics of doing this. The third is integrating this into the work, into the flows of a company. Autonomous technologies were ready in 2006, we are in 2026, only 0.25% of the car rides in the United States are on autonomous cars. Why? Because it is riding on human infrastructure. To add to the complexity, humans and machines are riding on the same infrastructure. If you draw that analogy to enterprises, it's the same. Enterprises have been built in a way, so you'll have to reorganize them, you'll have to reinvent them, you'll have to reimagine them to integrate these flows in the business. There is a whole lot of interplay between deterministic layers in the company and the new contextual layers being built. That is another piece of work. I mean, there is so much to be done.
You have to apply this to operations of companies. We were systems builders. So our universe was tech spend of enterprises. Our universe now is ops kind of enterprises. You have to integrate that to operations of companies. Operations are gonna be done in a different way. So all of this means, I mean, if everybody has access to the same models, where will the heterogeneity come from? The heterogeneity has to come from how you apply it. 70 to 80% of the Fortune 500 have SAP. The performances are different for all these companies. And how they've deployed them are very different. So I think we are underestimating the heterogeneity of these enterprises and how to integrate them. How to integrate them in a newer way. How to push this into areas where, look, this is going to cover things which classical software didn't. Classical software was deterministic. It only covered portions of the enterprise. This is going to cover the rest. It's going to take some portions of the deterministic out.
I mean, there is a cloud around the SaaS software. I mean, there's a fog around the SaaS software era because SaaS software was built for higher terminal value. You know, the assumption was it is perpetual. The assumption was you are capturing future value on top of it. Potentially you may not capture future value on top of it. You might capture future value on the side. You can do it on Cloud Co-Work or somewhere else. And if you do that, that's bespoke. Irrespective if it's captured on SaaS or it's captured outside, we still have a role to play. So, I mean, I feel more optimistic than before that every company will actually embrace this in everything they do, and they would need help.
We have one more question, I think Saurabh.
Yeah, hey Ravi, thanks for being here. I always get something out of every time that you show up at our summit. I thought you got a lot, but something. So Ravi, in the morning we had a OneCouncil breakfast. OneCouncil is our enterprise sounding board. And we had 20 enterprise leaders. And the one thing that came out very clearly out of that conversation is what should be our operating model? We've talked a lot about how should IT services, the providers be organized and operate. But how should the enterprise start to get organized to take advantage of this? Because change has to happen on both sides for this too.
That's a phenomenal question. Thank you for that. So look, you know, enterprises have to flip and look at us not as a capacity arm. Go to all tech services companies across the planet, almost 25 to 30% of their business's time and material. So it's capacity-based. 30 to 40% or maybe 50 to 60% is fixed price. And the small portion is outcome-based. And outcome-based is also not outcome-based. It's transaction-based. Like my TriZetto systems are transaction-based. We sell it BPaaS. We have a platform and a service. Now, how do I sell that? I sell that as number of transactions, number of claims. I should be able to sell them as how many lives I'm managing. If I'm auto-adjudicating claims, I can't actually price my model on claims. I should be pricing my model on the number of lives I'm managing.
So enterprises have to start to look at us as companies which can own things for them, own outcomes for them, so that they could focus on their own mission. I mean, that is a big shift. And that shift means you need more trusted partners. You need partners who are in it with you on your journey. So that's a big, big shift in the mindset for enterprises to engage with us. It's not going to happen overnight. I see this as a continuum. There are times when I've found it's not easy to tie it to outcomes. Why we have agile parts where our client teams and us work together. How do you tie it to outcomes? Who do you hold the outcomes? So we actually launched something called AI-infused rate cards in the company where we said, okay, we'll push this in. Now, in the process, we graded human labor and then we suddenly started to feel like, wait a minute, digital labor can be a craft. So embed the digital labor, tell them you don't open the tap, we'll open the tap. That's a new model.
So I think our clients have to start to see us as companies which can take higher risk, higher appetite to be embedded into your future and therefore we are more strategic than ever before. I mean, that's how we would see it. And in fact, if I was an investor, I would actually say, this is great because this is going to be more terminal value because you're embedded into the mission of the company. If you are embedded into the flows of the company, your engagement model is more outcome driven. So if it's outcome driven, it's actually more sticky. It's significantly more sticky than before. So that's how I would believe enterprises should view us and they should not see us as a capacity-based thing because you are not going to get the most value in this new model. We can take one more if you have time.
We've got to move on, Phil.
We're under orders, apparently. All right.
I have to pull the hook. I'm sorry. Ravi, if you do like to stay for lunch, you know, you're more than welcome to, but it is our lunch break right now. And Saurabh, you and Nigel will be up here in about 10 minutes to welcome everybody back that is interested in learning more about the HFS AI-First Deal Lab. So with that, please visit with our Hot Tech vendors and our other Tech Showcase folks. Please feel free to take your food back and learn more about our data insights tools that we're making available to the customers here. And enjoy a good conversation amongst yourselves. Thank you.
If you don't have an account, Register here |
With the exception of our Horizons reports, most of our research is available for free on our website. Sign up for a free account and start realizing the power of insights now.
Our premium subscription gives enterprise clients access to our complete library of proprietary research, direct access to our industry analysts, and other benefits.
Contact us at [email protected] for more information on premium access.
If you are looking for help getting in touch with someone from HFS, please click the chat button to the bottom right of your screen to start a conversation with a member of our team.
