Phil Fersht — CEO, HFS Research[00:11]
It’s great to be back with an old industry friend, Ravi Kumar, who’s obviously now the CEO of Cognizant, and he’s had a little time to get bedded into his new job over the course of this year. It’d be good to get Ravi’s views a little bit about how things are shaping up for Cognizant and also his views on Gen AI, the flavor of the year, which is really driving some big shifts in thinking and ideas within the industry. So over to you, Ravi — could you tell us a bit about the state of play and maybe what’s new with Cognizant?
Ravi Kumar — CEO, Cognizant[00:51]
You know, Phil, you’re absolutely right. I think everybody is wanting to drive out cost and equally innovate in a hurry. In enterprise landscapes in the last 12 months, there has been accelerated change and higher clock speed, I would say. Of course, there was higher clock speed since the pandemic, but that clock speed continued. However, the change ahead of us is higher than what we see or what we saw in the past, because every business is wanting to become a tech business, and AI is getting deeply embedded into enterprise landscapes. The difference is customers now want to take out cost, underwrite the savings to innovation and change. They actually believe and they want to do more for less. In fact, the belief is reinforced by AI that they can do more for less. So they’re equally paranoid about the pace at which they want to make change, but they want to underwrite it from the cost savings. They know that the pace at which AI is going to come and disrupt them, transform them, will mean that they have to find the ways to fund it. So that’s what is happening with our clients. Clients are also focused on the foundational elements of an AI-powered organization — data readiness, privacy and security, making responsible AI and function-grade AI. And I would say in the last 25 years or so since I’ve been associated with this industry, clients are wanting to build their own tech muscle, because technology is so core — software is the new alchemy for every business. Tech is deeply embedded into products and services, so they want to build their own muscle, which is counterintuitive, but I believe that as strategic partners we can help build the muscle. So this is what is going on with our clients.
Phil Fersht — CEO, HFS Research[02:57]
In terms of the customers that you have and the atmosphere that is in the industry right now, do you think we should be enthusiastic or scared about AI, and what’s your thinking behind this?
Ravi Kumar — CEO, Cognizant[03:00]
Phil, you and I have talked about it on this topic pretty often in the last 6 months. You’ve written quite a bit about it. I think any big disruptive discontinuity is going to be an opportunity and a threat. I see it in two swim lanes. We want to apply generative AI to our businesses and be more productive for decision support, and the biggest opportunity is for developer productivity — that’s my first swim lane. And I’ll come to the second swim lane, which is about applying it to client landscapes. Developer productivity is a jagged technological frontier. What I mean by a jagged technological frontier is that it has an uneven edge, it is irregular, and the quality is very disparate. What I mean by that is some tasks are easily done by AI while other tasks, seemingly similar in difficulty, are outside the current capability of AI, all within the knowledge workflow of tasks. So I therefore call it the jagged technological frontier. The utility of AI can fluctuate over the course of a developer’s workflow, with some tasks falling inside while others fall outside the frontier. So you always need a human in the loop, at least with the AI we have today. There is a report from a bunch of researchers on developer productivity, and in the knowledge industry, the speed of task completion — how quickly you complete a task — has increased by 25%. The quality of output, performance, which is evaluated by humans, human evaluation, is around 40%, and the total number of tasks completed has actually grown by 20%. So effectively all users have benefited, but the bottom half of the users have benefited more than the upper half of users. The high-performing users have benefited less, and the bottom half of users have benefited much more. It is an interesting paradigm where the tasks outside the AI frontier have to be strategically switched between humans and AI. You have to integrate AI and human capabilities at a granular subtask level to create something meaningful, or I would call it amplifying human potential. So that’s broadly how we are taming the whole AI benefit into the developer productivity life cycle, if I may. The second flip is, of course, how do you actually take this to your clients and apply it to businesses. And the most number of use cases I see today, Phil, is in productivity, efficiency, having a human in the loop, and the most immediate ones I’m seeing disrupted are customer service and operations and the potentially straight-through processing opportunities. These are the ones we are seeing most disruption on the client side. So we have a unique opportunity to apply it to ourselves, but also apply it to client landscapes and disrupt ourselves and of course disrupt client landscapes.
Phil Fersht — CEO, HFS Research[07:02]
Yeah, so it’s very interesting that we’re looking at honestly the mid-level going up like this and then the top level going up a little bit. So we’re seeing a real equalization of capability in our industry. We’ve looked over the last two to three decades of technology innovations and they’ve always created jobs. Despite fears that people would be automated out of existence, do you feel — is it different with Gen AI that we’re going to start to see jobs under threat, or do you think this is just going to create a whole new level of employment and focus elsewhere?
Ravi Kumar — CEO, Cognizant[07:24]
Phil, generally AI is a big discontinuity. If you go back to the discontinuities of the past, every discontinuity has created jobs of the future and taken away jobs of the past. It has actually created more jobs of the future and taken away jobs of the past. So I’m very optimistic that it is going to create significantly more jobs for the future. When enterprise software came in, in the 90s, everybody said the tech services space is going to be cannibalized. No, it didn’t — we added more jobs. It went to the cloud, more jobs got added. With machine learning algorithms, when productivity got improved for developers, you could actually consumerize technology. We saw more digitization, but more jobs got created. Generative AI is shifting the agency of technology to the end user. That’s a big shift. Machine learning algorithms shifted the agency to the developer; generative AI is going to shift that agency to an end user, which means you could use technology to empower yourself, to amplify your own potential. So I do believe you’re going to see more jobs in the future. What is certain though is the kind of jobs will change. In the past, the tech services industry hired procedural and algorithmic skills. We started to evolve ourselves to hire creative and heuristic skills. I think that equilibrium is going to be much more, and that’s because you need a human in the loop for using the power of AI, even if you want to productionize AI to enterprise grade. You need human factor intervention at the end of that cycle. To give you an example, an AI program is going to tell you the probability of rain; it’s not going to tell you whether you should carry an umbrella. It depends on the individual. You could carry an umbrella at 20% probability of rain; I could not carry an umbrella at 100% probability of rain. That human factor intervention at the end of the cycle is to train the algorithm with human factors and human behaviors and personalize the algorithms for end use. That whole piece will need more problem finders versus problem solvers. It would need more heuristic and creative skills, which will make this industry much more diverse in what you see and introduce domain-specific use cases like content writing to teach the model. For non-factual questions, content rating, creating regular expressions, annotation of content — remember, generative AI is not just about related, labeled data, it’s also about unrelated, unlabeled data, and that’s why it’s more creative. That’s why it has hallucinations. Hallucinations are by design. And to curate that model, you need newer capability and a different cognitive. So I’m excited about the kind of things you could potentially do and how much you could amplify the end user by, as I said, shifting that agency of technology and handing it over to them.
Phil Fersht — CEO, HFS Research[11:45]
Yeah, exactly. So let’s think about the services industry now. You know, we see 6 out of 10 of the G2K are now implementing Gen AI across at least one business function. Most of the rest are still in the pilot phase. We’re also seeing, while clients are figuring out what the uses are for this and where some potential is, there’s a lot of resistance to change still, and this seems to be a common theme that we’ve been dealing with for a very long time. How would Cognizant help its clients over this path of maybe changing their underlying data infrastructure, thinking about reconnecting your organization so it can take exponential advantage of these technologies, versus just layering on some tech to get more out of what you’ve got? How do you see the services industry taking enterprises over this change dimension, maybe getting them to do things they hadn’t done before? Do you see this as a major challenge and, you know, who do you think is going to win and lose as we come out of the next 18-24 months of this?
Ravi Kumar — CEO, Cognizant[13:17]
So, you know, you’ve seen this before. Tech services companies were leveraged as partners for scale and capability, and in the early days for labor arbitrage. Tech services companies of today are going to transition from labor arbitrage to technology arbitrage. That technology arbitrage will be the tooling, the instrumentation, the platforms which are powered by AI, which you could actually get your customers to embrace and take them through the journey. They have to build their own muscle and you have to handhold them to build their own muscle. And therefore I say it’s not just about lending the human capital, but also lending that value chain of human capital. So it’s a shift in our approach — that you can’t just lend your human capital for the journey, but you have to lend the technology infrastructure you built for that change you’re talking about. So I call it the tech arbitrage versus the labor arbitrage. In every enterprise, wherever you have actually seen preliminary initial embrace of AI use cases, we’ve always seen the change to be accelerated, provided you can create upward mobility in those jobs. Every job can be upward mobile if you can use the power of technology, power of AI, and make people do more valuable jobs and more prosperous jobs so that they see the value in embracing technology. So while technology related to AI is going to eliminate jobs, it’s going to create upward social jobs. Look at the United States — the United States has 8 million open jobs. It adds 300,000 jobs every month incrementally and it loses 4 million jobs every month. Who are those 4 million people who are leaving jobs? These are the ones who don’t see that upward mobility in their jobs. They’re stuck on a repetitive task all their life, and they feel so bored about it. They feel so dejected about it. They don’t see the American dream. They just walk away and they keep rotating and rotating and rotating, and that’s why we see this flux. Now imagine the power of AI, which can create that upward mobility so that these people who are so upset with the workplaces start to see that upward mobility. Using the power of technology, in every enterprise the challenge is to actually tell workers that this is going to benefit them with higher productivity, higher efficiency, higher upward mobility, so that they embrace it, and as they embrace it, they do more value-added jobs. Yesterday, Jamie Dimon, the chair of JPMorgan, commented that workers are going to see 3.5-day weeks in the future because of AI. What do you infer by that? You would infer that workers are more productive, they’re going to make more money, they’re going to be more prosperous, they’re going to be more happy on one side. On the other side, you would infer that there is a lesser number of things to do, and therefore workers could get paid less. The reality is machines are going to take over your repetitive tasks. You will have more productive tasks to do. You will have more upward mobility, and therefore you should be more satisfied, and you should be able to get more wages for lesser time and hence be more prosperous. So I think every industry will go through this. I kind of split the industries into two buckets, one which is bits-related and ones which are atoms-related. The ones which are bits-related will become more bits, and the ones which are atoms-related, which are traditional industries like manufacturing, will immerse themselves with more technology. All of them will emerge in some way on technology which is powered by AI. So I see this as a good problem to deal with. I see this as a way forward for that upward mobility I spoke about. I see this as an opportunity to bridge the divide which we created in the past. I see this as a way to create more productivity in our jobs and therefore better prosperity for workers.
Phil Fersht — CEO, HFS Research[17:52]
What happens beyond this ceiling, Ravi, do you think? Right now this is a lot of getting the most out of what’s out there. Gen AI as it is can improve our ability to code faster, to do customer operations more effectively, more efficiently, be more creative, more collaborative. What do you think happens beyond this? I mean, we’re just at the start of a new curve, right? What do you think happens when we talk in two to three years’ time?
Ravi Kumar — CEO, Cognizant[18:25]
You know, we all want to be in an AI-first, machine-plus-people endeavor. The global corporate world has only 28 to 30 million developers, and the world needs more technology professionals. So if AI can do 40% of the code, you will see more tech intensity. You will see more allocation of budgets because of higher productivity. You’ll see more for less. So I see this as a much better virtue and a much better opportunity for technology to be in the middle of enterprise landscapes. I’m excited about the fact that the shift towards creative and heuristic skills, and the balance to algorithmic and procedural skills, will be rapid. You’re going to see more STEM, non-STEM disciplines as much as STEM disciplines were in the tech industry. We’re going to see more problem solving, and we’re going to see more problem finding as much as problem solving. And finally, we’re going to see technology arbitrage versus labor arbitrage and therefore a more strategic opportunity for tech services companies to partner. So I see this as a wonderful opportunity to be a more strategic partner than before.
Phil Fersht — CEO, HFS Research[19:46]
Fantastic. Well, this has been very good, hearing your advancing thoughts. I know we’ve spoken a lot about talent and the emergence of tech in the past, but this last 6 months has been most invigorating, and it’s great to see your thinking advance as well and really get a view of what the future is going to look like. So great getting some time with you today, Ravi, and look forward to seeing you again very soon.
Ravi Kumar — CEO, Cognizant[20:10]
Always a pleasure talking to you. Wonderful to know that you’re kind of deep diving into a topic which is of deep, deep importance to our clients and to the industry. So thank you for doing this.
Phil Fersht — CEO, HFS Research[20:46]
Thank you very much, and we’ll talk soon.
