My name is Brian Essex. I'm J.P. Morgan's large-cap, mid-cap software analyst. With me today, I'm very happy to have IBM. Specifically, we have Mohamad Ali, who's the Senior Vice President of IBM Consulting. Mohamed, thank you very much for joining us. Thanks for having me, Brian. For those in the audience, I'll try to leave 10 minutes or so at the end of this session for Q&A from the audience. Just to give you guys a heads up, we'll flag that again before we hit that mark. Maybe for some background and perspective, Mohamed, you spent a lot of your career building enterprise software- and AI platforms. Right. What brought you back to IBM? Why did leading a consulting business inside a technology company feel like Right the best place where you can make an impact? Yeah. It doesn't feel obvious, does it? Just to give everybody a sense, I've been in the software business for about 30 years, and now I lead the consulting business at IBM. Just a little context, so when I got my graduate degree, this is like in the early '90s, I started a company called Neural Applications Corporation. It's exactly what you think it is. It's a neural network company in the early '90s. We built it, and we sold it. Then I ended up at IBM for 14 years, and I mostly did software. Arvind and I were peers. We worked for the same person for 4 years while we built the software business. Then I left for about 14 years, and Brian, you know this, I was CEO of a data protection cybersecurity company and then another company. After we sold the last company to Blackstone, I got a call from Arvind, our CEO, and he said, "Hey, listen, I'd like you to come back to IBM." I said, "Well, what do you want me to do?" He said, "Consulting." I said, "You know that I'm not really a consultant." He said, "You don't understand." I said, "Okay, well explain it to me." He said, "Listen, the consulting business is going to transition to a human plus digital labor business. What is digital labor? It's a bunch of software. I need somebody who has a software perspective to go build the environment that deploy hundreds of thousands or millions of digital workers and transform this business." That was two and a half years ago. We've been doing that. In Q1, 40% of our signings, which is actually higher than that, the $22 billion, 40% of the signings was solutioned with human plus digital labor. We're running it at scale. We have over 400 projects like this. I could look into the factory, and I could see the human plus digital labor manufacturing these solutions for our clients. We've been able to apply this technique at IBM. We call it Client Zero. We applied it to a $25 billion spend. That spend is now about $20 billion, so about $four and a half billion reduction in productivity. We've been able to apply it to AI first companies like Riyadh Air, which we started working with three years ago. Now they're flying. They have aircrafts, 200 aircrafts in the sky flying. This human plus digital approach to this particular business is what I've been focused on. I think it's why Arvind thought that somebody with my background would be the right person to transform a consulting business. Mm-hmm. I think there's been some talk too about Well, I know there's been some talk about leveraging things like AI for better productivity across the organization. Yeah Client Zero, right? Right. How do you think about how much of that savings has been within the consulting group versus the rest of the organization? Yeah. Have you productized that across your consulting organization? Yes, we have productized it, and that's really important here because that, I think, really differentiates what we bring to the market, right? You're here because you track IBM, right? You have seen the increase in profitability and investment in R&D and so forth, right? When you add it up, hopefully, you could get to the $4.5 billion that we talk about. In terms of within consulting, if you look at consulting, we essentially have two businesses in consulting, a business that is growing dramatically and a business that's under compression, right? When you mix those two together, if you compare 2024 to 2025, our pre-tax income is actually up 20%, which suggests that this business that's growing, this GenAI powered business that was 40% of our signings in Q1, is generating a lot more profit than the other business, which sort of makes sense, right? That was about 200 basis points, right? We do report our consulting numbers separately, so you can actually see that in the financials. I do think, as we continue to deploy this new model that we've created, which we've actually productized, right? We've built a platform called IBM Consulting Advantage, and on top of that, we have 4,000 high-quality digital workers that are deployed every day. These are 4,000 unique ones, so there are tens of thousands that are running right now. If I pick up my phone or my computer. I look into the factory, I actually see them, right? I have one account right now where this morning I just looked, and there are 800 human beings, and there are 300 unique digital workers. These are unique ones, right? When they're spun up, right now there are probably thousands of digital workers with those 800 human beings doing the work. It's almost like looking into a car factory. How many of you here have a Ford truck, by any chance? Raise your hand. Okay, one person. How about a BMW? BMW? No BMWs? What do you guys drive? All right, I went to this Ford F-150 truck plant. I don't know if any of you have been on this tour, but it is an amazing tour. You look down, and what you see is a bunch of robots and a bunch of human beings, and they're building a truck. When I turn on my computer, that's what I can see now. I can look into Bangalore and into Romania and into Louisiana and New York City and see the human beings working side by side with these digital workers building these solutions for our clients, whether it's an SAP solution or a data platform or a migration to the cloud. It's a whole new way of running a consulting business. That's why that part of our business is growing so well. With that construct, the number of humans and then the number of digital workers, what is the senior to digital construct of that- organization, how has that changed over the past year? Yeah. How do we think about, you have a senior person that has ownership of- Yeah a number of these digital workers? How has that progressed? Yeah. We just announced this thing called FDUs for deployed units. I was at one of your competitors. Actually, you worked at this company last week. The guy said to me, "I have 5 FDEs from," I won't say what company, "and they're these 24-year-old kids, and they can build really great stuff out of this LLM, but how do I deploy it into my processes? I can't do that." These four deploy units, we used to have groups of 30 people doing projects. Now they're six. Yeah. They're a combination of those kinds of FDEs, plus people who understand the processes. We can take something that would take a year now do it in eight weeks, which is a little scary in itself. There's a bank right here in this city where we took 1,400 processes. You would never actually look at 1,400 processes in the old days. because it's just too much. We are re-engineering those 1,400 processes using this FDU model. This the question you're asking about the ratio of people. Right. I'd have to look and see what the seniority level of these six people are. It's a combination of quite senior people. quite junior people and then a whole bunch of digital workers to go do this work. If you think about 1,400 processes and re-engineering them with human plus digital, that just could not almost ever have happened. There's this question about is AI just replacing traditional work? There's a big part of that. There's a whole new kind of work that's happening, like this 1,400 processes that we would never have re-engineered before, but now you're doing it. That's net new. Yeah. Very cool. Where are you seeing From a vertical exposure and a geographical exposure, I guess, can you generalize where you're seeing the greatest adoption of that? Is that primarily U.S.-based? Yeah. how to- This is sort of funny. Yeah. We run these contests now. Our platform is called IBM Consulting Advantage. The idea is every digital worker that we have has got to run on this thing, otherwise we starve them. We don't give them tokens or whatever. Now we have a common platform for all of our digital worker capability. We run these contests, and now every quarter, I give out these three awards for the top three users of ICA, is what we call it. The first quarter we did it, all three were in Europe. It was a little bit accidental because at this one account in Europe that manufactures beer, they did it extremely well. It's like the 800 people, the example I gave you. The account in the next country over came and copied what they did, and the one over there came and copied what they did. That's why they were so good. Just this last quarter when we did it again, they were actually all in the Americas. Almost every region now is adopting this. In some ways, this is sort of how we track our evolution, not by how many AI classes people pass. How many projects? We probably have about 2,000 projects at any given time. We have about 400 of some of our largest projects that I can go on this dashboard, and I could see the human plus digital workers. We reward people and teams for that transition. How much of this, though is what I would classify as old economy versus new economy? One example I'll give you is, I had a conversation with Abnormal AI. Yeah. They started a whole new division with one person. Yeah who has an army of agents working for him. Yeah. I would classify that as maybe a cutting-edge new technology. Yeah. What are you seeing on your side? Yeah new versus old? We're seeing, I would say, two big parts, two big trends in our business. One is that there's a lot of old stuff that is just compressing. If you had a $100 million BPO business, the client comes to you and says, "Hey, now it's got to be $70 million." You have no choice. You have to do that. That is just compressing. This is all this traditional BPO stuff. Luckily, we're not as exposed to that, but some of our competitors are really exposed to that. Then the other side, what you've got is, I want you to re-engineer a bunch of my stuff with GenAI. That is broken into two parts, I would say. One is kind of a shift from old to new. There's that shift happening. There's this whole net new business that wouldn't have existed before, but now we have GenAI. Let me give you an example of that. Does anybody know this company called Pearson Learning? Raise your hand if you do. Okay. Lots of you. All right. More than have BMWs apparently. Pearson learning, some of you might know, they credentialize human skills. Right? I don't know if in your company you get badges for doing your business conduct guidelines or your security training or that or the other thing. Do you guys get badges in your companies? Raise your hand. Okay. Certificates. Nobody gets badges. Certificates, right? Right. Pearson Learning has this system called Credly, and we use it at IBM. A lot of companies use it to do those kinds of certifications of human skill. They came to us and they said, "Hey, listen, there are going to be millions and billions of digital workers, and we want to get into the business of certifying these digital workers the same way we would do human workers," right? We said, "Great. We're going to help you build a whole product there." We're helping them build that product. At Think, which was actually in this building two weeks ago, right? We actually demoed how this thing works, right? Now, as myself, right, I have a security badge because at IBM, I have a business conduct guideline badge. I happen to have a generative AI badge as well. If you're a cloud architect, you might have a cloud essential badge or something, right? These are people. Now we have these digital workers that do similar things. How do you credentialize those? The Pearson system will actually test these digital workers, and you can't test it the old-fashioned way, where you just ask it a bunch of multiple choice questions because it'll just memorize everything, right? You actually have to give it complicated, novel workflows that it has to solve. When you have two digital workers testing each other, you don't have to do multiple choice, right? You could ask complicated things. That's how they're credentializing these things. Then our digital worker gets a badge, and it actually shows up in the Credly system. That is a whole new business for Pearson, right? Pearson's human credentialization is business A, and this business B is completely net new. That's why in this kind of new part of our business that's growing, that's 40% in Q1 signings, there's a part of that that is net new that just would not have existed before. How does that change the composition of the business from a, I guess, strategic consulting perspective to a runtime consulting perspective? Yeah. Does that shift in who maintains that business longer term for like a Pearson, for example? Yeah. Therein lies the opportunity, right? Yeah. I think traditional consulting is changing dramatically. We've adopted this concept of services-as-software. Has anybody heard that term here yet? 100%. Okay. When I first joined IBM, I put up this chart in one of our meetings, I called it Consulting as Code, I was told not to use that term. Then this analyst firm came up with services-as-software, and we're like, "Whoa, that's great." Right? Forbes had a big article on it last year. Yeah. Yeah. We started down this services-as-software path because if you think about it, traditional service is going to get compressed, traditional software is going to get compressed, but there's this whole world of digital workers, billions and billions of them, that are going to come into existence. Software companies are going to create them, services companies are going to create them, but there's this whole new market there. That's a big part of where we're going to play, right? We're using our 4,000 unique digital workers to perform services today, and we charge for that, right? That's why our margins are going up. We're also providing those digital workers to a company like Pearson, we're helping them build it and run them on a platform like IBM Consulting Advantage, and we're charging for that, right? You could see there's this whole runtime business that you had created, and in some ways, I think this is why Arvind wanted somebody with a software background to go figure out how to do this, right? Got it. Super helpful. Maybe if we could take a step back a little bit and look at the profile of the business for the rest of the year. Yeah. Right. I think this past quarter, consulting was relatively flat. Jim's guidance was for low single digits for the year. What are you seeing inside your business that gives you confidence in more of a growth ramp in the back half of the year, particularly in an environment where it sounds like some of the consulting peers are a little bit more cautious about the macro? Yeah. I'm very confident as you could probably guess here, right? Why am I so confident? A big part of it is, I started in 2023, right? By the beginning of 2024, we launched IBM Consulting Advantage. It took us about 12 months to kind of get it up and running at scale. Then in 2025, it really took off, right? You see the delta, right? You see 20 points improvement, 20% improvement in profitability, right? For the last five quarters, the gross margins have increased every single quarter, right? Then in Q1, we took some of that, and we invested it in our sales, right? That's happening, and we are also seeing this expansion of our GenAI powered business, as I said, to 40%. It's now 30% of our $31 billion backlog, right? Yeah. It's faster than any of our competitors. If you do an apples to apples compare to when our biggest competitor was reporting, you would've seen when we were at 30%, they were at 10%, right? We're definitely, by those metrics, moving this part of it faster. You guys know this, right? You're all smart mathematicians. If you have two businesses, one that's going like that and the other that's going like this, and this one is more profitable than this one. It might average out to flat growth for a while, but at some point, this becomes much smaller, this becomes much bigger. Right. I think that is going to happen. I'm confident it's going to happen. Jim can opine as to when that's going to happen, right? My job is to push as hard as possible for that to happen as soon as possible. From a profitability point of view, the trend is already there, and I think that trend is going to continue. Even if the thing stays flat for a long time and the profitability increases, your enterprise value should increase, right? Because your profitability is growing. I think what's going to happen is that that faster growing business is going to start outpacing, then you'll see growth there as well. All right. Does that logic sort of make sense? 100%. Okay. Yeah. Then maybe from a macro perspective, what are you hearing from your customers, particularly in terms of budget sensitivity and. Yeah Discretionary spend? How are they thinking about spending on these projects? Yeah. Has that changed in the recent month or two? Yeah. I think everybody knows discretionary spending is people are cautious, right? Very cautious. How are they spending on this 40% of signings that we have? A fair amount of that is that we're not going to spend over here, but we're going to spend over there. Hey, you know that thing that was $100 million that now I'm asking you to do for $70 million? What if I gave you somebody else's piece of the pie, and for that $200 million, you could do it for $140 million? Now I have $60 million that I can actually invest in this new area. We're seeing that quite a bit. That model is happening quite a bit. As they invest in sort of AI-related services, they're starting to realize that, hey, we could do a lot more with this, right? Pearson, for example, when we did that transition, that transition was focused on productivity within Pearson, right? Doing traditional IT more efficient. They quickly realized, look, we could build this whole new product with it. All of a sudden, budget came from the business units to go build that new product. Not from IT, from the business units to go build that. We're starting to see new pockets of spend, especially on the net new part of it. Mm-hmm. Got it. Then from that budget perspective, we've heard a number of cases where business leaders are getting incremental budget by slapping AI on top of their- Yeah the name of their budget, right? You've been in this process where there was a lot of experimentation over the past year or so. We're starting to hear more things moving to proof of concept. I mean, where are your customers in that journey? Yeah. How many of them need verifiable evidence of ROI before they pull the trigger on a larger spending project? Yeah, it's sort of interesting because people spent a lot very early and didn't get much results, right? I think some of you have probably seen that study from MIT that says 95% of all these GenAI projects fail, right? How many of you here, raise your hand, have seen that study or heard of that study? Okay, a fair number. I think it's absolutely true. About a year ago, I was in London and I was having dinner with the CFO of a very, very large company. Halfway into the dinner, he leans over to me and he goes, "Hey, listen, I have 1,500 POCs and not a dime of benefit to my bottom line. What is this AI stuff about?" I said, "Well, how'd you do that?" He goes, "Well, we have 50,000 employees. We bought certain licenses for everybody, and we said, 'Hey, go be productive, go do POCs and report it back.' I'm not seeing anything." I think there was a lot of spend very early on. People are starting to realize that that's not how you do it. When you step back and you say, well if 95% of people are failing, there are 5% that are successful. Whether it's by accident or not, like when we started this at IBM two and a half years ago, what we call the Client Zero project, we didn't start with the AI, we actually started with the process re-engineering. We took the IBM company, we decomposed it into 490 workflows, and I still have that spreadsheet. We took 70 of these workflows, and they were the low-hanging one that represented the $25 billion of spend, and we said, "Hey, can we re-engineer that with human plus digital labor?" In the beginning, we built 700, I'm sorry, 220 digital workers, so about four digital workers per workflow, like HR, finance. That's how we got those productivity gains that we're able to now reinvest in the business for growth. It's that model that we're going out to all of our clients with. Your question about, yeah, people are slapping AI onto things. Are they getting benefits or not? With this Client Zero example, we have some credibility, right, when we walk in. We don't just walk in saying that we did it. We walk in with IBM Consulting Advantage, and I bring it up and I show them all the tools to re-engineer their HR workflow, their finance workflow, their FP&A workflow. Like I said, I was at one of your prior employers just last week, and this is a smart company, right? I bring that up, and the guy goes, "Oh my God, you can not only just map out this whole process that I've got, but you've got the digital workers to actually insert in those processes," because we did it in Client Zero. That's where I think a bunch of our acceleration's coming from. Got it. If you think about these projects, these AI-driven products or projects. How is your workforce composition aligned to deliver on those, and how are these contracts constructed versus maybe the rest of the business? Shorter duration, longer duration? Yeah. Outcome-based versus. Yeah. Great question. Let me start at the end there. Sure. These tend to be shorter duration. We're actually seeing our contract length come down. The other thing that we're seeing is that erosion is stabilized, is flat, right? That gives you confidence in the backlog, that you have a stable backlog, and it looks like it's going to be yielding revenues faster because they're shorter like that. That's one thing. I think on the workforce, the workforce I think is the biggest challenge for every company. Your company, my company, everybody, right? How do you get people to shift to this new way of working? I was at two accounts last week, two different ones than the bank that I went to, right? The first one is a $250 million a year account for us, and in Q1, it grew 26%. You say, "What is going on there?" It's because we have teams of people there who are bringing all of the digital workers, our IBM Consulting Advantage, the new way of thinking about that client's problem, and they have credibility, and it's growing. I go to another client with our team, and it's a $50 million a year account, and it's not growing. Part of the reason it's not growing is that you could sit there and watch that team try to articulate how to help this client, but they're using the old way, right? They haven't learned the new way yet, right? How do we get all of our people over to where that first team was? That is the big challenge, and I'm sure everybody's having the same challenge in your own company. How do you get people to adopt these new tools? That's an area that we're pushing on extremely hard. This is why we're running the ICA contest. We're recognizing people. We're pointing out where people are not doing it right, et cetera. I think the skills composition is the big question that I think we're all dealing with. Great. I think we have about seven and a half minutes left, and I promise to leave some time at the end for questions. If there are any questions from the audience, please raise your hand, and we'll pass the mic around. We have one in the back there. Mic right behind you over there as well. There you go. Yeah. Hi, thanks. Chris Zaharias with Multiverse. I wanted to ask if you are facing any of the compute limitations that the hyperscalers are seeing. We've heard a lot about throttling that folks like AWS and Azure are being forced to do. Are you seeing that in your business at all? Thanks, Chris. Not yet, but we're going to at some point. We're using hundreds of billions of tokens per month now, right? It's probably one of the largest. We're becoming a very large consumer of this. Two weeks ago, I think, right here, the head of sales for AWS was on stage with me, and a big part of what we were talking about is how these 4,000 digital workers run on AWS in multiple geographies, right? In India, it runs on AWS. In certain other geographies, it runs on AWS. One of the more interesting one is now IBM Consulting Advantage with the 4,000 digital workers is FedRAMP certified, and it runs on AWS GovCloud, and we're actually provided these kinds of services into regulated industries using that stack. It's a great partnership with AWS. We have similar partnerships with some of the other cloud vendors. We have not hit the point where throttling is an issue. One of the strengths of IBM Consulting Advantage, though, is that it sort of helps direct which LLMs we use, right? We have a whole suite of LLMs. Certain LLMs, by policy, we can't use in certain industries, and it determines that. Also, there's certain LLMs that you can use unnecessarily, right? If you're summarizing a document, do you really need the latest Anthropic version, right? You don't need Mythos to summarize a document. ICA sort of has a router. We are managing ourselves, we haven't hit the wall where somebody else has to throttle us. Yeah. Does that answer the question, Chris? Okay. Maybe to follow up on that, what about with your customers? I think at the recent IBM Think, you announced a platform where you can help your customers- Yes throttle between different platforms. Are you seeing any traction? It's still early, right? Yeah. Are you seeing any traction from that because customers are sitting there saying, "Well, wow, I just blew through all my tokens for the year, and we're only in April. IBM Consulting Advantage, just as naturally, it is the most advanced thing of its kind that we have in IBM because we started building it almost three years ago, right? We're using it at scale. How does that get to clients? It gets to client two ways. One is through our Orchestrate product in our software group. We have some technology that came out of Consulting Advantage that's in there. For clients like Pearson, we actually offer something called Enterprise Advantage, which is basically a Consulting Advantage at the bottom. Access, we call it Advantage Marketplace, I'm sorry, Advantage Platform is the thing at the bottom. Once they have that installed, it automatically does this for them. The middle place is Advantage Marketplace, where they can actually buy digital workers from us, and Advantage Service, where they can buy FDUs, right? These forward deployed units to help them build their own digital workers, right? Those are the two ways you could get this throttling, this sort of token management capability, either through Orchestrate or Enterprise Advantage. Very cool. I apologize, I think there's a question in the back there. Hi. Thank you. Nick Blasi from Eagle Advisors. Could you share a little bit about IBM's quantum computing business? Is there any update there? Sorry, can you repeat that? Sure. Sorry. Can you share an update on IBM's quantum computing business? Oh, quantum. Thanks. That's going great. What in particular are you thinking about? Just whatever's on your mind. Thank you. Yeah. I tell you, this is sort of the second reason I came back to IBM. I'm an Electrical Engineering background, but I started out in physics. We're sitting on, I feel like a gold mine here at IBM. The technology's come a long way. Basically, the machine works, right? Two weeks ago, I think we showed that not that long ago, we were only able to simulate six atoms in a protein, and now we can do 12,000. I think by the time we get to 30,000? Is that the number? Once we hit 30,000, it becomes applicable to cancer therapy and so forth, right? The machine works, and you say, "Well, okay, if the machine works, then what's the issue?" Part of the issue, I think you guys might know this, but I'll explain it anyway, is a quantum machine is like a cheetah. It runs extraordinarily fast, faster than anything on the planet, and then it pauses, right? Why does it pause? Because unfortunately, you can't fully isolate these things, and there are all these quantum particles floating around in the air, and they penetrate whatever isolation you've got, and it hits the qubit, and it goes out of state. You have a lot of errors, right? You can solve this with really advanced error correction that's never been done before, and that's one of the paths that IBM's working on. You could try to isolate it better, and then there's sort of a third way to do it, right? As you make that part where the cheetah has to take a break smaller and smaller, this thing just takes off, right? That's what we have today. We have a quantum cloud, where there are at least 250 clients on the quantum cloud today doing real work, like HSBC, like Cleveland Clinic. It's super exciting. That's a different part of the business. In consulting, we focus on quantum-safe. We provide all the quantum-safe services. Soon we'll be doing quantum computing services. Yeah, for those interested, they spend a lot of time. I think Jay, at IBM Think, just a couple of weeks ago, gave an update on the timeline, and went into a good amount of detail on the quantum business. I think we might have time for one more if there's any from the audience. One right there in the back. Is there one more here? Yeah. Yeah. Okay. I'm not sure if you'd agree with this, but it seems like the cost of tokens is going to be an important gating factor in adoption. I'm just wondering if you have any forecasts on where we're going to see the lowest cost tokens going forward, whether it be hardware, software, IBM's management of all the above. Yeah. This is a great question. Two and a half years ago, when I started working on this, somebody said, "Well, if we worked really hard, we could get the cost of these digital workers to be a dollar per hour." I was like, "No, no, they need to be like a penny per hour," right? A lot of this comes down to how you manage the tokens and how you use the tokens. That's a big part of why we built Consulting Advantage to begin with, to manage the use of the tokens. Now, we get a lot of benefits out of this because we get automatic FinOps. When I turn on my dashboard, I could look right this second as to how many tokens from Anthropic, OpenAI, Gemini, Granite, et cetera, are being used. Managing the tokens is a big part of how we address this. The other part of it is I do think token prices are going to come down, right? Who's going to pay $25 for 1 million token going forward, especially when there are other providers that are coming online with quite competitive LLMs. I do think that over time, the LLM costs themselves are going to go down. People are going to continue to use more and more of them. I think ultimately, the challenge to adopting AI is going to be what we're focused on, which is how do you re-engineer the workflows with human plus digital labor? Those are the tools we've been building, both for our application within IBM and Client Zero, and now at that bank where we're doing 1,400 processes. I think if I was going to forecast token usage, it's through the roof, and I think token costs are going to come down. Yeah. Models will become more efficient as well. Yeah. Yeah. Exactly. Great. With that, I think we're out of time. Mohamad, thank you so much for joining us. I really appreciate it. That was awesome. Thank you. Thank you, everyone.