Panel · 9:45 to 10:30 AM · Wednesday, May 13, 2026
I'd like to invite up a great panel of people across the industry. Frank D'Souza, we have CP Gurnani, Yusuf Tayob, Nan Li, Puneet Mehta, and Lisa Stump. Come on up. Yeah, great to see you. Come on through. Okay, that was quick. All right. So let's maybe start from left to right. Very quick, name ranks here on other.
Good morning, everyone. Frank D'Souza, co-founder and former CEO of Cognizant Technology Solutions, now an investor at private equity firm Recognize.
Hi, I'm Puneet Mehta. I'm the founder CEO of an agentic AI company called Netomi. We do the autonomous front office for the world's largest enterprises.
Hi, I'm Nan Li. I'm head of global transformation at Conde Nast. You may know us as Devil Wears Prada, but we publish a variety of magazines.
Hi, morning, everyone. This is CP Gurnani. I am a founder and CEO of a company called AIONOS, which is AI for enterprises.
Yusuf Tayob. I'm from Perficient, where I've been CEO for the last year. Perficient is about 7,000 AI strategists, designers, and engineers, and we're really building a native AI services business 24 years at Accenture prior to Perficient.
Good morning. Lisa Stump. I'm the Chief Digital Information Officer at the Mount Sinai Health System here in New York City and the Dean for Information Technology at the Icahn School of Medicine.
Thank you, Lisa. So we have people who've built some of the biggest tech firms in the world and people who are running some very ambitious, fast-growing tech firms and then people who are consuming and buying the services. So this should be a great conversation. So let's get started. Going the wrong way there. Personal accountability. So be honest with the room. Are you personally using AI every day to do your job, or are you still delegating that to a task force and waiting for a report? So Nan, would you like to kick us off?
Sure, absolutely. I feel that the usage of AI has to start with personal accountability. If we expect others on our team and stakeholders throughout the business to embrace and adopt AI, we have to lead by example. And so I personally am a huge fan of leveraging generative AI for everyday work purposes. A lot of what we do tends to be complex program management, strategic redesigns, change management support. All of that has been made so much more fluid with the usage of AI. I think personally as well, I'm sure like many of you here, I'm dabbling in things like vibe coding, agentic exploration. We have got to be doing these things as leaders within our businesses.
And I think with the same question, it would be great to hear about personal accountability. Maybe Frank, I'd love to hear from you a bit about how you're as an investor now as well.
Yeah, look, I think for us, I can look at it at two levels. At the individual level, I think I've been playing with the technology almost from the beginning, doing a lot of the things that was spoken about. Particularly, for me, it's helpful on writing and research. These are the two personal use cases that I find most helpful, and we can talk about that. But at the enterprise level, we're a relatively small investment firm, and we have systematically gone through, my partner Charles Phillips is leading this and looked at every single part of the investing process and how we can apply AI. It's remarkable how good AI is in sourcing, helping us find opportunities. We can source at scale, which we were doing with humans before, and now we're able to canvass, you know, we're very focused on investing just in services business, digital services businesses. That's a universe we think of, you know, for our space, about a 100,000 companies in the world, maybe a little bit more than that, and now we can source and stay in touch with these companies at scale. But then if you look at the whole investing process from sourcing all the way through making an investment decision, we have figured out how to use agents at every step along the way. At two different degrees of maturity, some much more mature than others. And then the last thing I'll say is that after we've made an investment, we have a portfolio, we've made 15 investments, and we have 13 companies active in the portfolio right now. We used to hire operating partners, and we still do, who help us figure out how to help drive change and how to keep track of the performance of our portfolio companies. We're now building a digital operating partner that sits inside of our portfolio companies and gathers automatically key metrics, surfaces insights, tells us when there's an improvement opportunity, when the metrics are telling us that a portfolio company is not quite on the right trajectory. So I think through the whole investing lifecycle and then through the value creation lifecycle, we're thinking very actively about how you use AI.
Thanks, Frank. So where has AI fundamentally changed how work gets done? And what do you have to dismantle to make that possible? Yusuf, would you like to?
We dismantled everything. You know, I think probably what we've learned over the last year, first to ourselves or for ourselves and then we do for our clients, is that it's very difficult to adapt legacy processes, you know, fix legacy processes and get more than something slightly better than legacy answers. It's difficult to sort of try to change a legacy operating model or keep a legacy operating model, adapt it a little bit and get a different culture. You know, it's different in our business. If you're a services company, it's difficult to use legacy SIs and get better than legacy results. Our view is you have to sort of change everything. And you have to start, you know, from scratch. We call it zero-basing the company, you know, and we look at every process. We look at everything we do with talent. We look at everything we do in how we serve our clients, and we do all of it differently. In fact, we believe different is so important that we've rebranded our entire company around this concept that you have to think about it as different. You can't just think about it as doing it incrementally better. You know, in our organization, I've been there a year. We spent the first sort of three months talking to our clients, doing a lot of research in terms of what do our clients expect from service providers like us in the world we live in now. And then we spent the next two quarters literally ripping everything apart and rebuilding it. And we did in two quarters, you know, what I think it would have taken us probably two years to do in sort of a more legacy approach. And every day I tell the leadership team, we're going to change every single week. And we're going to continue to change depending on what we hear from many of the headlines that you shared with us earlier today.
Right. How much resistance did you find you were hitting in terms of people who are used to working in an older way of redirecting information up and down a hierarchy?
I mean, I think human behavior naturally resists change. The first thing we did is we told everybody that everyone's got an opportunity. And, you know, I don't think people are inherently opposed to change. I think they naturally resist change because they're not sure what it means for them. And yet I think, you know, what we've seen, the creativity and the entrepreneurship and the intelligence that our people have when you give them that platform to create that change, has been remarkable. And so I often say, Phil, we don't actually have a change management problem in our company, which is surprising to me, right, given the work I've done most of my career. But I think giving people that platform that says, look, everyone's got an opportunity, and the more you change, the more we're going to reward you, and then standing behind that has really helped us.
Thanks. So I'd like to get into, we were talking a bit about debt earlier, but one debt we didn't mention is something we like to call leadership debt. So what are you actually fixing in your process, data, and talent foundation before you can actually scale AI on top of it? Maybe Lisa at Mount Sinai, share some of the specifics.
Yeah, you know, and I talk about that leadership debt often and very clearly with my executive colleagues in the health system. You know, healthcare in general is traditionally slow to change in the operational sense, right? It's a bit of a binary environment, right? We've got bleeding edge, leading edge exploration and experimentation and research around new clinical treatments and cures. But we keep working the way that we've always worked. And many of you probably experienced that. It's difficult to get an appointment. It's difficult to understand your bill. Sometimes you don't even know where to send your payment for your bill, right? Healthcare has not evolved in many of those ways. And so there's this urgency right now to deploy artificial intelligence to help solve those problems. And the leadership and decision debt is very deep around we've made historical decisions about how we organize our people and our functions and how we operate. that if we layered AI on top of or even in the middle of, you're just going to accelerate and accentuate a problem and miss the real opportunity. So I'm working intensively over the next month, actually, on a series of executive meetings and working with our human resources leaders to say, what are all of those jobs that today are highly exposed to artificial intelligence? Where are their heavily task-driven functions? How are we going to think about those in terms of the workforce? We need to make leadership decisions then. One, there's technical advice I and my teams will provide. These are simply the things we don't recommend leveraging AI for today. Here are all the things we think we can as leaders in healthcare with humans at the helm. We will say there are still some things, even though we could automate, we shouldn't yet. And we'll move those in the human bucket. But we're going to end up with a group of functions and tasks that need to be performed by people. And then figure out how to organize our people around that work. And that's a very different org structure. You showed it, right? It's a very different operating model. We're going to have people then performing a different set of tasks than any single role has ever performed. And so we talk about upskilling and retraining. We haven't even sort of fully identified what our people need to do in partnership with the AI agents that are now part of our team. And that is exactly how we are thinking about it. So I reiterate over and over, it's people, it's process, it's tech. The tech is work, and it's hard, but it's work that's often the easiest part of any of this equation. It's getting back to how do we organize our people and redefine our processes in a very sort of legacy, hierarchically driven industry so that we can achieve those highest aims.
Do you feel you need to get to the point where you almost have to put the people thing aside for a minute and you just look at how do we want to achieve outcomes that we need to be successful, right? So you figure out the outcomes you need to achieve and then you almost work back from there to think how do we then organize around that? And then it's how do we retrain our resources and figure out if they're capable?
Yeah, to a large extent, I agree with you. I think, you know, many of us are in highly regulated industries. Healthcare is absolutely one of those and we are still in a place where certain tasks and functions by regulation will require a human, but absolutely agree with you. And we think about those outcomes in the terms of the experiences that we want to deliver. What's the experience that we want our patients to have? What's the experience that we want our doctors and care teams to have? What's the experience we want our workforce to have, right? You want to come to work in a place where you have all the information that you need about your own personal benefits or policies in the organization. And then I support an academic institution. And so what's the experience of our teachers, learners, and researchers? And so we are defining those outcomes and then working back, as you said, and through that thought process of what could we automate and what should we as two sides of that coin.
Maybe, so CP, you've led Tech Mahindra for many years and built it to a certain point. And as you look at the industry now, how would you approach addressing these leadership issues and this leadership debt we talk about?
So, I would look at it as two different phases. You know, at Tech Mahindra, it was more about standing up every quarter and justifying, and the usual question from the analyst was how many people you added last quarter. Usual questions were productivity per employee and so on. And Frank did the same for many years and I did the same. When you step back, you knew that the world around us has changed. I mean, that was December of '23, early '24. And we knew that we have a chance to do things in a different way. You know, like Yusuf said about listening to the clients, we also went and spoke to the sectors we wanted to be in. It became very, very clear is that the market is dramatically shifting towards outcome-based pricing. It's no longer about, as you put it in your presentation, dollar per hour. The second part is that I would say is that while the clients were talking about AI, but most clients had one challenge, that they want to do AI probably to put a tick on their checklist. They were discussing AI as a discrete project, not as a way to involve the whole company. And I think my biggest job, or my team's biggest job, has been to engage with the clients and tell them that this communication has to be open. You cannot say that I'll rip off the quality assurance systems and then one fine day some 20% people will lose their jobs, or 30% will lose their jobs. I think the moment you bring all the stakeholders together, you define the objectives, that project succeeds, the others are always work in progress.
Right. Maybe, Puneet, in response to some of that, you're driving like an AI-driven, aggressive CX startup, right? How are you taking advantage of this mindset in the market? Is it something you feel you can exploit, or do you also suffer from some of the same issues of people being too discreet and not collaborative enough with AI?
No, we are heavily exploiting it in a good way, if I could sugarcoat that word. So just about Netomi, we run the autonomous front office at companies like Delta Airlines, United, MetLife, the NBA, the list goes on, many of the Fortune 500. And one of the key things we noticed in that mindset shift, the CP that you mentioned, model companies are selling tokens. And all the professional services companies, system integrators were mostly selling time and materials up until now, but the companies were wanting to buy outcomes. So we come in the middle where we are like, how do you convert a token economy into an outcome economy? And we also, if you look at the round we announced last week, we raised another $110 million, and this time Accenture led the round. So we are partnering with companies like Accenture, companies like Adobe on one side, and the earliest investors in Netomi were the founders of OpenAI and the founders of DeepMind. So it's almost like to bridge that gap for the outcome, because you no longer have the luxury to deliver outcomes over 18 months. That used to be a typical deployment cycle. When we deployed at Paramount, the entire deployment was three weeks. On the other side of that deployment, they had autonomously handling over 80% of the traffic in the front office. So it completely changes the unit economics of that business. So long story short, I think we are riding three waves there. One is this conversion of token economy to outcomes. Two is that the outcomes need to arrive really quickly, that you no longer have the luxury of six quarters because the world changes in six quarters. And the third one is that literally enterprise AI needs a win because if you look at the world's largest enterprises, like how many of these wins are publicly talked about today and they're quantified and measured. But also, every stakeholder around the room needs that win, right? So we are kind of catalyzing that and we are delivering that.
Yeah, I mean, it's an interesting concept that everyone wants to move way faster. We're used to everything at our fingertips. I mean, I see you nodding a lot, Yusuf. You're saying how you turned around a company in a few months. What works with faster change?
I mean, first of all, it was a successful company, and we had the benefit or luxury of operating from a position of strength with a lot of client trust and everything else, but what our clients were telling us is they wanted to see it different going forward. And what we found was if we could get everybody to think different versus think incrementally better, we'd get a very different outcome. For example, we looked at all the work that was being done to connect dots internally, all the sort of operational activities. And, you know, last year I'd say we were still in the AI era of productivity. This year I think we're moving to the era of intelligence. And so everybody was thinking about how do I put agents side by side with my operational people to make my operational people more productive. And we said we're not going to do that. So we eliminated this chief operating officer role entirely. We just cut the head off. And we said there is no COO in the company. And therefore we don't need any of these operational functions because there's no COO to report to. So let's figure out now how to do the processes and the functions completely differently. We rewrite them on behalf of our clients. And we did it to ourselves, too. Our CMO's here, as we did our rebrand, we looked at how do we build now our entire image, including our website. And we do this for a lot of our clients. We build all of their end-to-end digital experiences, their websites, everything else for some of the biggest companies in the world. So we said, we're going to use our own team to do this. And the initial estimate that came from the team was we're going to do it in 10 months. Actually, it was 12 months. And then they used AI and it became 10 months. And we said, well, that's not going to work. We ended up doing it in 10 weeks. And we didn't make the existing website better. We redid the entire thing, including data. We did it in 10 weeks. Is it perfect? No. Is it excellent? Yes. And we're learning from it every single day. To answer your question, that website is now built to optimize for AEO and it's actually built to serve the agents that are going to be crawling it more than the humans that are going to be crawling it. And so every single day, because of how fast we moved, we're learning more, we're adapting, we're changing it, we're changing it, we're changing it. And we're, I don't know, six weeks post-go live now and I think we've achieved as much in how we've changed it in six weeks as the 10 weeks that we put it in. And we'd be in the design phase right now if we took the original estimate from our very own AI first team.
Yeah, optimizing for AI. I mean, we realized very quickly that there's way more people trying to find out about services and software by going to ChatGPT than going to our HFS website. You've got to admit you're no longer a destination site. No one here is. So let's take it to the next conversation around control layers. So who actually owns the AI decision-making layer in your company today? And do they have the authority to stop something that's already running? Frank, would you like to kick that off?
Yeah, look, I mean, I think I'll speak more about the experience we have with clients because at Recognize we're a relatively small partnership, and so it's relatively easier to make decisions across the firm and drive change, drive process. You know, that's probably one of the, I think as I listen to this conversation, if I back up for a minute, you know, there's that old saying, the more things change, the more they stay the same. And so some of these things that we've talked about have, in some sense, always been the case. Technology has always driven change. Technology has always required forward-thinking leadership to adopt. Technology has always required that you lead from the front. Those things, I would argue, haven't changed. Now, this is, in my view, AI is, at least in my career, my lifetime, the most profound technology shift we've seen. And so I don't want to minimize the impact of AI. And that characteristic of AI, if you will, is what makes this more difficult for adoption because the adoption of AI plus the fact that AI is inherently probabilistic, or LLMs and generative AI are probabilistic technologies, makes it such that you need to get a lot of, particularly in regulated industries, but even in unregulated industries, you really need to get consensus across the organization on how you're going to deploy, why you're going to deploy, how you're going to check the outcomes and outputs, what are the guardrails, that requires a significant collaboration across the enterprise. The places that I've seen that don't do it well are the ones that don't have a function that can integrate across all of those multiple stakeholders in an organization. The places that I've seen it work the best with clients are where there is some individual. It could be the chief AI officer, which is an increasingly common role that we're seeing out there. It could be a chief transformation officer. Sometimes it's the chief operating officer who has been empowered, usually by the CEO, to say, your job is two things. One is to prioritize value. Where is the value highest from AI in the organization. And then the second is to prioritize velocity. And you've got to do both those things simultaneously in order for this to work. Because if you don't prioritize value, you wind up in this pilot hell that we all talk about. Everybody's doing stuff and nobody's figuring out why this is going to work, if it's going to work, what are the measures of success. And if you have value but you don't have velocity, you wind up in the people who say, well, it's going to take 10 months or 12 months to get something done when it could be done in 10 weeks. So the best companies that I've seen are the ones that have figured out how to do both value and velocity simultaneously.
Wow. So let's think about build, buy, or partner here. Where are you building your own capability versus partnering? And what happens to your differentiation when your partner becomes your competitor? Maybe Lisa, can you think about that one?
Yeah, you know, we've built a strong talent pool internally of, and we're fortunate, right? We have an academic institution that is part of our enterprise. So we've got a strong team of clinical data scientists. We were the first academic medical center in the nation to stand up a department of artificial intelligence and human health. And so we have the capability and the capacity to build. I don't believe we always should. And for every one of us that thinks about in the moment that you're sick and that you're most vulnerable, the role of AI needs to have a human at the helm, and we firmly believe that still at this point. But we are building where we've got that capability and capacity to marry our subject matter expertise, in this case around clinical care, medicine, and scientific research. And we are building the models on top of an architected data that's rich with a very diverse population of patients and sized adequately to develop those models. That said, we are buying and partnering where we believe we've got capability in systems and solutions that we already own. And so again, if it's not perfect but good enough, let's accelerate to get that velocity through some of that thinking. And then we are partnering where a company brings sufficient tooling and platform that would take us an inordinate amount of time or money investment, but we bring subject matter expertise to layer on top of that platform, those are where we're seeing the partnerships really pay off.
Right. That's really interesting. And maybe from a growth company perspective, Puneet, build, buy, or partner and how do you stop your partner becoming your next competitor?
Yeah, the way we look at this, I'll answer the question from two lenses. One is companies that are adopting AI and then the partner ecosystem, because like I said, we partner with everybody from traditional software companies. So Adobe is a partner, but then professional services companies, model companies. So we are surrounded by partners. So on the companies adopting AI side, I think the question you've got to ask is, if you're using a technology, is this company willing to deliver on outcomes? Price it on outcomes. Like, don't even pay for POCs. Challenge them and say, hey, I'm not going to pay for the POC, but I'll pay, you know, higher on the outcome side, per unit outcome. And second is, you know, see if they can deliver something in under a quarter, even if it's kind of the first launch. So I would say that should be the litmus test for any large company adopting a technology, if you want to partner with somebody or not. Because at the end of the day, there's so much transformation to be done. You have to pick your battles. As a large company adopting AI, if you're stuck in creating an entire application tier, then it's just going to slow down the process. And then on the other side, the way we look at it as a growth stage company is there are companies out there that already own a relationship, that have process knowledge, or are an integral part of the workflow. It's wiser to not try to rip that out because it's just too frictiony. Because again, if you're sitting here two years from now, the world is gonna look very different. And the ones that would have really brought in that change would not try to rip and replace entire systems, would try to see if they can, in the first phase, bring in AI into existing processes, existing systems, and also just looking at the next 24 months. So we are a big proponent of that partner architecture.
Thank you, Puneet. CP, what's your perspective? Build, buy, partner? What should the industry players do?
You know, again, let's put it into two different buckets. One is when you're responding to a client. Now, when you're responding to a client, many times you don't have all the components, but you do know the architecture. You do know what you're trying to deliver, and sometimes the client is involved in helping you partner. But fast forward once you have delivered, you always end up also building a few components. Many times you could own that IP of those components. So one is, one bucket, I see it as a natural progression, instead of applying any creative strategy, this has become more of a reaction to a particular client situation. The second as a principle. Now, as a principle, we all work for value creation, value creation for our clients, value creation for the firms that we serve. And more often than not, we do realize that the investments some of the larger players make, whether it is the cloud infrastructure providers or whether it is the OpenAI or Anthropic. I mean, it is better that those foundation layers you are not building. But what you build as a policy is your moat, your differentiators. And I think it depends on the industries that you serve. Like, for example, today we serve telecom, we serve the travel industry. And in those industries, revenue management system or better customer experience or being able to serve a data exchange for the data monetization for the client, we are building it because that's our differentiator. Now, on the third part, are partners becoming a competitor? You just scared me. I mean, you said Anthropic, OpenAI putting $4 billion to become a deployment company. Yeah, sure. I mean, they will become competitors, and we will have to be agile, and we will have to figure out what is the best way to dance.
Best way to dance. Yusuf, I'd love your perspective on this. You know, how do you view the whole partner-competitor situation when you're looking to grow a business?
I tell our teams every day that our entire industry exists because of what ecosystem partners and the platform companies create. Not the other way around. We exist because we help our clients solve problems. We exist because we have partners who make great technology. We exist because we have tremendously talented people who want to grow their careers. That is our entire purpose for being in the tech services industry. So we are decidedly and proudly partner first. I also think that when I speak to most of my clients, they're trying to make sense of all the chaos. They're basically taking the slides that you showed and asking us to decode them. And whatever we tell them today, we show up a month later and everything's changed. We've got to help them decode it again. And so where they tell us they get a lot of value is we bring the industry depth and expertise. We've got deep functional expertise. We understand their businesses. We know how to drive behavioral change amongst humans. We don't make the technology, but we certainly help make the technology deliver outcomes and deliver value. And so it's our view that we lean forward into partnerships, but then we help our clients sort of sort through how do we take those partnerships and actually help create real value. And I think in the age that we're living in right now, where everything changes every couple of weeks, that role that we play is critically more important. And in terms of them becoming our competitors, I've been avoiding this question since last week. And it keeps coming up everywhere we go. Carrie's back there laughing because I was like, I don't want to talk about this. We had multiple reach-outs from media and stuff, and I'm like, I don't want to talk about it. But you look at the OpenAI thing, you look at the Anthropic thing. Again, I was with OpenAI last week. I'm with Anthropic next week. They were great partners of both of them. And again, the conversations we're having with them is, yeah, they can bring forward deployed engineers to help move consumption onto their platforms. But they're still asking us to bring the industry expertise. They're still asking us to bring the functional depth. They're still asking us to help them design the thing. They're still asking us to stay around and make sure that it captures value. And so, frankly, we're leaning into it entirely with them, and we're benefiting from it.
Hey, Phil, I want to jump in here. Two things just following up. On one point, the first point, maybe a little bit of a divergent view from Yusuf. I think that going forward, services firms that don't have some degree of proprietary IP don't succeed. Because if you're a services firm, you've got great talented people, but you're using the tools that everybody else has access to, I don't know where your moat comes from. So I think that every services business needs to think about proprietary IP going forward. Now, so this build, buy, or partner question, the answer is yes to all of them. You build, you buy, and you partner. It's not like one or the other. But if build is not part of your plan going forward as a services business, build being building proprietary IP, I think you have a difficult road ahead. And then, maybe again a little bit more controversial, but on this issue of the deploy coders from the model companies, look, I mean, I think that's a natural evolution, but I'd love to hear from Lisa and Nan here. In my experience, customers want services businesses that are agnostic and objective, not services firms that show up and say, "Look, I'm a forward-deployed engineer, but whatever your problem is, the answer is OpenAI or the answer is Anthropic." Look, there's a role for that, frankly. Oracle has a great services business, but that Oracle services doesn't negate the need for objective third party services companies that look across the ecosystem. So I think there's a role for both, but I don't think it's like a deploy coder from OpenAI when the answer is always gonna be OpenAI is I think gonna be somewhat constrained in what they do in the market.
If I could just comment, you know I think you're absolutely right, right? When you named a number of the companies, right? Whether it's OpenAI, whether it's Palantir, whoever, that model of we're going to ingest all your data, we're going to run it through our proprietary platform, we're gonna put our forward deployed engineers, I almost get an allergic reaction now to the forward deployed engineer model, because they embed with the operators, right, the frontline teams, and everything they build is then within their platform. And you get this lock-in of dependency, because then if you've built 12 workflows in an individual functional area on that platform, and I say, "I think we can do it better over here," then I'm disrupting that workflow, right? And it's well, but it's all in this single pane of glass and this single environment and workflow. It keeps scaling in a way that becomes financially unsustainable. And so the services firms, I do rely on being agnostic in terms of, and persistent, no pun intended to another firm, driving value and letting us define the outcomes that we're seeking to achieve and helping us do that in a way that's also sustainable over time. Otherwise, we end up locked in in a way that is frightening, frankly.
Can I add to, I completely agree, Lisa, because obviously in our case, we are still evaluating in a lot of instances how we want to create that moat from a capabilities perspective. And there is certainly a fear of, again, every model comes out, every new competitor comes out, everything is moving up in lockstep, but where is the future, what is the future gonna look like 24 months from now? So that's something we're very, that's very top of mind for us in terms of how do we make sure we're making the right choices for the company, not only for the next 12 months, but for the next, call it three to five years. The second thing I will say that is really interesting, I think I might be one of the few in the room who come from this publishing media world, is as you can imagine, we have quite a different relationship with a lot of these major LLM providers. Our moat is content, and it is largely written content. And so from an evolution of 2024 to today, we've had a pretty contentious relationship, turned into quite a nice outcome with the OpenAIs, the Anthropics, the Perplexities of the world, where again, they were crawling our data. But by turning that into a positive monetizable relationship, so now we license our premium, researched, high quality journalistic content to these providers, we've actually created a revenue stream for ourselves. So it's been quite a shift in the mindset the entire publishing industry and the media industry interacts with these players.
Interesting. So they've actually created a whole new channel that you're actually grasping for, because I think the publishing industry was in a lot of trouble before.
Yeah, exactly. Because in the era of AI content slop, and we all see it online when you see a link to an article, we are finding that people actually flock to a voice of authority, a tone of reason, researched, fact-checked articles, so on and so forth. And so we look to the OpenAIs of the world to surface our content organically when folks search for, you know, I want to see what to wear in Germany in the summer, or I want to understand what is going on with ICE in Minneapolis, right?
Right. And then in that regard, what is the specific line you will not cross, even if a competitor crosses it first and wins business because of it?
Yes, that is absolutely, it's very pertinent for us, and the ethical line we will not cross, and there's no regulatory hurdle or anything, unlike some of these other industries, but we will never use AI to generate content for us. We will always have a human-led, human-driven voice for every single piece of content we publish, so a lot of our competitors are leveraging AI to do first drafts of articles, first drafts of news, first drafts of summaries, newsletters, so on and so forth, we will always have not only the human in the loop, but the human voice lead the loop, if you will.
We have a big AI slop debate later.
Yeah.
Okay. So let's get to the final question. So five years from now, the leaders who got AI right will have made one decision the others refused to make. What is that decision, and have you made it yet? Puneet, would you like to kick this off?
Yeah, so as we have deployed across some of the world's largest companies, we've had a real view into what has worked and what has not. And what has really worked well, especially with a problem as complex as an autonomous front office, is when they looked at the sanctioned architecture that Netomi offers, that governance piece has a main part of the puzzle, where it was not an afterthought. It was not a bolted-on thing. It was not like, hey, I'll automate a handful of tasks and hope that it works. The governance was a key part. And the second moving part was, you know, how do you identify where to apply AI? So, you know, the formula that we usually offer to companies that use Netomi is, if I was drawing a Venn diagram, it's like, what's high repeatability, what's low to medium exception management, and what's low to medium business risk? So the intersection of those. So it's almost like we are building this the way self-driving cars were built, because a lot of the organizations would be self-driving or autonomous. So I would say those two things, where the governance and the sanction piece was not bolted on, it was a main part, then identifying the surface area.
I see you nodding there.
You know, all these years, one thing I'm convinced is that for AIONOS, it is not AI, it is applied intelligence. And to me, whether it is five years or 10 years from now, I mean, I still strongly believe that the company that I will build or the customers that I interact with, I mean, they all not only believe in values and ethics, but also more importantly, building trust and relationship with them. So Phil, all of us getting together out here five years from now also, hopefully, we will be talking about our friendships also, not just the technology.
Right. So, focusing on human relationships. Okay. Maybe, Frank, you'd like to finish this off. What do you think leaders who got AI right, what's the decision the brave ones made, which the others didn't?
Phil, I'll speak. Since I think much of the audience here is from the services community, I'll speak directly to those leaders. I think you have to be, the ones that will succeed will be the ones that have the courage to drop revenue before you grow revenue. There's a J-curve coming at you, whether you admit it or not. Whatever you've been doing today, you'll be able to do with a fraction of the human beings. Your pricing models today are largely time materials pricing models. You have a J-curve coming at you, and have the courage to lean into that. Those that do, I think, will get the respect of clients. Those that do, I think, will get the respect of investors. And then you'll come out stronger the other end. But if you fight the trend right now and don't admit that you have this J-curve coming at you, I think it's a recipe for losing the trust that I think CP pointed out, losing the trust of your clients, and eventually losing the trust of your investors.
So you think the investors, Wall Street, they look fairly negatively at services right now. Are they waiting to hear what is the plan? How are you guys going to get away from a linear model? Is that?
Yeah, I think that's absolutely right, Phil. I think that we now know that for large parts of the services ecosystem that there's a tremendous productivity benefit from generative AI. That should not be a debate any longer. And the only question is how much and how much faster and better do the models get over time? Given that, for any given unit of work that you are doing today, you can do it better, faster, and arguably cheaper using AI. And I think investors understand that. When we look at investing in businesses, we talk to management teams. And we're investing in smaller businesses, so the J-curve is not quite as deep for smaller businesses. So the issue is a little less pronounced. But if our management teams don't have an acknowledgement of the J-curve and then a plan to say, yes, I'm going to manage through this and come out the other end, then we think they're somewhat tone deaf.
And it feels like the biggest issue isn't the technology as to how the hell a company is going to pull this all together. And I feel like we're in a world that's dominated by technology dreams, but we're the people who have to make it happen for them, right, Yusuf?
Yeah, I mean, first of all, we're taking advantage of exactly what Frank said, which is companies that aren't leading into the J curve are the very companies we're looking to disrupt. If you can do the SI at 30% to 40% less effort, we're saying let's do that much faster. And as a result, we're betting that companies are still spending lots of money on this. Phil showed us this morning, right? So we're betting that we're going to get more volume in a, take a declining TAM of legacy SI, you can get more volume while you build native AI services. And so I completely agree, lean into the J curve. I think the one thing we haven't explicitly said here on the panel, but everyone has implicitly said, is the talent piece. I think if we're sitting here five years from now, the companies that have gotten it right have enabled their people to be different. That's truly the key, right? I mean, Nan's talking about the fact that AI is never going to create content. That's not going to happen. Lisa's talking about the fact that in their company, we're not going to put life and death medical decisions in the hands of code. We're ultimately going to have a medical professional make those decisions. In the services industry, you know, we are building outcomes and value and even IP from the people who power these systems. And we're giving them opportunities to have, you know, great careers like we all got to have in the last 30 years in this industry. So I think the people that are empowering their people not to just do it better, to learn, to adjust, to adapt, but to do it differently and to reward them for doing it differently, those are going to be the winners.
Yeah. Terrific. Well, I think we probably have to wind this up and hit some coffee, but Frank, Puneet, Nan, CP, Yusuf, and Lisa, thank you very much. What a great conversation.
Thank you, thank you. Thank you. You too. Awesome.
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.
