International Business Machines Corporation (IBM) Earnings Call Transcript

Good afternoon, everyone. Welcome to our annual Think conference, which is our client and user base conference overall. So we have a mixture here today of both all of our sell-side investment community and also our buy-side community. So you get a little perspective of IBM. But before I turn it over to the panel, I want to give you just a little bit of a taste about what you're going to see for the next 1 hour, 1.5 hours, and we'll definitely keep this informal. We've got a little presentation style, but we want to make sure it's open for questions, comments, thoughts. But you're going to see, hopefully here today, the vision of what we've been trying to build here over the last handful of years at IBM to basically reinvent a technology-focused, innovation-focused company. And Arvind kicked off his keynote this morning. He talked about the vision we had 4 or 5 years ago about the two most informative technological shifts that we've seen a long time, that being Hybrid Cloud and AI, but more importantly, the synergistic effect of both of them. And when you think about that, we've been repositioning this company, our portfolio, our operating model, our set of offerings and the integrated value of how we bring a platform-centric model together. What we're going to share here today is something we take to pride, and that is called Client Zero. Arvind spoke about it a little in the opening. Rob just talked about it in the Automation side. In fact, that concert discussion downstairs started out on a whiteboard in my conference room about how we were actually leveraging technology inside IBM. But why are we having this conversation? Number one, yes, from an investor perspective, we laid out at Investor Day our next vision of our model, durable, higher inflected revenue growth, higher operating margin company and a stronger free cash flow generation company. How do we do that? Productivity is at the core. How we transform the way we run our company is at the core of that. But more important than just those financials, and you're going to hear from the team here, is this team actually defines what best is in running technology in a multinational, multidimensional enterprise company, whether it's supply chain, finance, HR, operations, you name it. We leverage technology inside IBM to drive that productivity, but we leverage it to tell our technology teams and our consulting teams what best is. So with that as a quick opening, I'm going to turn it over to team, and they're going to tell you about the story and the journey.

All right. Great. Thank you, Jim. Thanks, everyone, for coming here today. I have the opportunity to not only kick us off, but I've also got the really, really important pressure of keeping the clicker. And the pressure kicked up a notch when I realized it's a one-way-only clicker. So hopefully, we'll be able to navigate this well. As Jim framed, we've got a really ambitious agenda for you today by witnessed by the number of speakers that we've got. And what we really want to do is advance the ball -- hopefully advance the ball for all of you on three primary questions. The first is, what is this enterprise AI challenge? Why is it so difficult? As Arvind said earlier today, we're starting from an industry perspective to move out of the experimentation stage into real scale and deployment. But I think by just about any measure, while everyone is experimenting something like less than 30% of those experiments are getting into production. So why is it so hard? Why is it so difficult? What differentiates the enterprise space versus the consumer space as it pertains to AI? And we'll explore this question. But if I just kind of plant a couple of seeds for all of you, the answer is really because it's so much more than about the technology itself. It's so much more than about what model you use with maybe a little bit of UI slapped on top for document summarization, e-mail drafting, meeting minutes, et cetera. To get into the enterprise space to get up into the workflow, it takes having control of your data. So this problem statement of the 99% of enterprise data that sits behind companies' firewalls that's unstructured, that's hard to get access to and bring together. So there's a huge data dimension to it. It's about how you organize your company by the workflows, how do you align resources and skills against defragmented workflows in order to drive simplification, eliminate low value-add work and yes, bring automation. I think the theme of automation you'll see play out in our conversation. Cost. When you're dealing with experimentation and you're in an R&D world, you allocate a certain amount of your budget, you prioritize experiments against that. When you run out of money, you run out of money and you move on. But if you want to take some of those experiments and scale them up into production, it's got to make sense. It's got to have an ROI. And even more to the point, we see a lot of our clients struggling with just not just the cost, but even understanding what the cost looks like. There's a lot of concern about when you scale into production, are we going to get surprised by just how costly these models are and these deployments are. And the last dimension is governance. I think every C-suite, every Board-level conversation around AI and generative AI grounds itself in how do we responsibly deploy this technology. From a data perspective, data privacy, copyright considerations, providence of where that data is coming to train a model to how do you actually measure the performance of a model? Is it introducing bias? Is it robust enough to navigate all the changes that go on within the environment? Is it transparent to the data? Is it explainable, right? Within the finance domain, and myself is controlled within the finance domain, it's just not good enough to have a model that can predict in sort of a black box fashion. Even if it's good at predicting, it needs to be prescriptive because behavior change within the enterprise is so critical here, and you'll see that need to explain why what the levers are to drive an outcome. So there's so much that comes beyond just the technology to scale in the enterprise. The good news is IBM, we have an answer for all of these elements, which leads to the second question. Second question is how are we doing it? I think we'll demonstrate for you, we have some success within this space. How are we achieving that? How are we achieving the $3.5 billion that Arvind talked about earlier, and we've talked about -- Jim has talked about in our earnings. which then leads to really the best question, the third question, if you believe this is really hard to do and if you accept and see that IBM is getting some traction and seeing some success with a long runway to go, what does it mean? What do we stand to benefit in terms of our own productivity within the enterprise, but also very importantly, in what Jim said before, Client Zero. How do we differentiate ourselves and demonstrate to our clients there's differentiated value that allows us to support our clients, allows us to support our business overall. So it wouldn't be an investor discussion if we didn't start with a flywheel. We tend to cover flywheel. It's a very good illustration of what we're talking about. I think Arvind said earlier that productivity is the heart of the enterprise. And I think we definitely embody that, right? We drive productivity for the purpose of freeing up resources to reinvest back in the business to drive innovation, innovation that matters, that creates customer value, that drives demand and revenue, operating leverage and productivity and so on and so forth, the virtuous cycle of the financial model. I think the other thing I'd point to is that this is very much the epitome of who we are as a company. This is cultural. For 100 and almost 115-year-old technology company, you need to be committed to reinventing yourself for the purpose of serving our clients, creating innovation and doing it in a trusted way. Tom Watson Jr. said back in the 1960s, something that I think sort of resonates the ethos of an IBM. In order for a company to stay great in a world that's constantly changing and evolving, it needs to be willing to change everything about itself, except for those core beliefs. And we, I think, embody that in how we approach all of our efforts, and hopefully, you'll see some of that on display. So if I take the flywheel and kind of unravel it, if you will, across the page and fit it any other way, and walk through the components of what we're talking about here. The first is the productivity element. And for those of you that have been following IBM for the last couple of years, you'll note that when we started this conversation about driving productivity, we started and we flagged a commitment of around $2 billion that we felt within a couple of years in terms of run rate spending, we could drive. I think for all of us with maybe the exception of Arvind, our confidence level increased as we went. And Arvind is the exception because he was setting objectives out from the very beginning in terms of what we could go drive here. Our confidence increased, our view of what the yield to value increased, time scales pulled in. We went from $2 billion to $3 billion to $3.5 billion, all within the same horizon, and we sit here today with $3.5 billion delivered through a series of projects. I think we're now at 45, 46, I lose count a little bit. We've been meeting weekly with Arvind on this since the very beginning. And if I break this down into a couple of categories, and we're going to delve into some of these as we go, but $3.5 billion, as Arvind said earlier, it's hard tangible savings. That's part of my role as controller and the team to really make this something tangible that we could see in the ledger. Over $1 billion, $1 billion to $1.5 billion from reduction in vendor spending. Think of this about as our supply chain. 350,000 SKUs, 10 million shipments a year, 2,000-plus vendors around the world, an extremely complicated supply chain and manufacturing process within our hardware business. How do we simplify that? How do we augment those workflows with AI in order for us to ensure that we meet demand in a timely fashion, that we supply and source the parts for our programs and that we manufacture in the right place to optimize the supply chain overall, drive lower cost from a freight perspective, from a parts consumption standpoint, from an inventory management, lower carrying costs, if you're more efficient with this. And then very importantly, being agile in the world that we live in with a lot of exogenous variables from a geopolitical standpoint, from a tariff standpoint in order to accommodate and be adaptive to a changing environment. Supply chain, subcontractor spending within the Consulting business that you can start to see play through as an element of the productivity and margin performance we're seeing within Consulting. And our IT environment, which Matt and Ed are going to talk more on. So the intentionality of Hybrid Cloud by design, one environment from infrastructure to network application data and the workflows that line on top that's driving considerable savings with our vendor stack within this space. $1 billion to $1.5 billion from vendor spend and then $2 billion from what we would call our shared services. And think of this as G&A, all the teams up here on the stage, how we work to defragment our organizations around the world, drive work elimination, work simplification and then now bring AI against those workflows to drive automation overall. So what do we do with that? So $3.5 billion exit run rate coming out of 2024. So a little bit less than that within the I&E in '24. But in a couple of years, that's close to 5 points of margin opportunity for the business. Where has it gone? We've reinvested and grown our investments nearly $1 billion within R&D, and very intentionally moving up our E to R from an R&D investment from -- this chart shows from 9% in 2020. But if you go back a little before that, we, for the longest time, been 7%, 8% E to R within R&D, moving up the stack to 12% with a design to continue that up to mid-teens over the horizon. Significant investment in Quantum, our generative AI stack, Hybrid Cloud with Red Hat. Z, we've got our Z leader sitting here with z17, launching investments, considerable investments in innovation. Our go-to-market, we've invested heavily in our ecosystem, as we've talked about, technical sales capacity, net of all of that, we've expanded our margins 200 basis points over that 2-year horizon, which allows us to drive our cash flow, growing faster than our revenue, allows us to invest more. We have more financial flexibility, more M&A activity in this virtuous cycle, you can see play out. Revenue growth overall, we feel accelerating in a sustainable way from minus 3% to plus 3% to 5% plus and a glide path that we move forward on. So you can see how this plays out within the company, how it's played out over the last couple of years. We're going to now shift. I'm going to hand off to Joanne here, and we're going to shift into the how, right? The how question of how have we been able to accomplish this. So Joanne?

Great. Thanks, Nick. Yes. So I think you've heard a lot about, obviously, the vision we had it was an extreme version of productivity. So we've seen about what the size of the opportunity is. And what we wanted to do is ensure that we were really driving the right ROI, that experiments were not what we wanted to do. We actually wanted, as Nick indicated, to actually fundamentally reimagine and change the way that we actually run every element of IBM's operating model. So the real vision here was that at this particular moment in time with a more simpler IBM, how could we take a step back and ask ourselves what actually can we eliminate? How can we look at our complexity and see a vision of clearly taking out steps, standards, removing application fragmentation and really aligning on a vision that actually to run a digital enterprise, you actually do need to completely radically reimagine your end-to-end workflows. We focus on how do we simplify. So you think about every element of how we are. So we obviously run procure to pay in our procurement processes, as Nick indicated, running our supply chain. We obviously have every element under Jim of record to report. And for me, obviously, a big element of how do we run quote to cash? How do we touch everything from our clients and our partners right the way through to invoicing and into the ledger. And so we wanted to look at what does that take today to run every business unit in every country that we operate in the world. We also then envision that we could actually use IBM's technology to really automate everything that is transactional in nature. It gives us a great opportunity. So Client Zero, as Arvind mentioned this morning, is actually a great exciting vision to be drinking our own champagne. But sometimes it's also our opportunity to eat a little bit of our own dog food, because what we do and actually do is lean in with our product teams at the alpha and beta level of capability and functionality, with a vision that actually we can run IBM with 280,000 IBMers in 170 countries across 4 or 5 lines of business, and we're a really super use case for how you actually can run and drive a digital enterprise. And we give that insight and that feedback to our product teams on an ongoing basis. And then Arvind's vision and it's AI first, and it's AI in everything we do. How do we truly bring to life our own transformation story. And so today, we're going to share with you some of those use cases that are more mature and perhaps have driven the best outcomes and the biggest return on ROI, but we have over 70 use cases that today are generating incredible value. How do we did this on an ongoing basis is both tops down and bottoms up. So the tops down is Arvind's vision that we could drive IBM to be the most productive company in the world. And that really leans in on us looking at how do we want to create a small steerco, not heavy in terms of cadence, but very focused on boldness and impact. And also because we truly need to actually transform every aspect of who we are, we want to make sure that we're making the right trade-off decisions on risk versus reward. Clearly, obviously, as well the size of the prize and the value and the disruption that it might create and might deliver. So the steerco was Arvind; Jim Kavanaugh, our CFO; Nick Fehring and myself. So small and mighty, every Friday afternoon for 30 minutes of showtime. Showtime being a very good insight about what we've dreamed and envisioned, that we could go after. It allowed the project office to run 2-week sprints, driving insights on what it took to run an operation today, how many steps in the process, what does best-in-class look like? When you talk about being the most productive company in the world, you need to understand what our competition are doing, what's industry leadership doing, what does best-in-class look like. And right the way through, honestly, as well, looking at how many applications today do we have in this place. Nick mentioned a reduction in vendor spend. One of the big ones that we leaned in on quickly was because we have grown up with very different business lines, hardware, software, cloud, and obviously, our TLS maintenance business. We've also done a lot of acquisitions. And so we looked at what does our IT landscape look like today and how much of it is really integrated end-to-end, which makes it touchless, makes it seamless, and is a really great user experience. And that obviously was eye-opening. So it allowed us to call out what could we remove, how can we reduce spend. And then the final one, I think that's probably the most exciting part of our Client Zero experience as a team is we get to be the incubation hub for all of our technology, of course, bringing the integrated value of our consulting team with their deep domain expertise. So when we redesigned the end-to-end workflow, we truly brought people who deeply understand finance, who deeply understand procurement, partnering with our technical teams who deeply understand the product capabilities and functionality that it brings. So this has been a 2-year journey. We're going to take you through and share with you some of the use cases that we've achieved. But it would be wrong of me and remiss of me not to tell you that it clearly starts with data. Data is the foundation. Transformation of end-to-end workflow is the enablement. Technology then becomes an easy deployment. And then, of course, as we talked about, this is technology on one side and culture and people and behavior on the other. So as we talk you through the next kind of use cases, we're going to give you some insights about what that did for IBM and obviously, what we're generating in the road ahead. So first of all, I'd like to have Ed Lovely share with you something on our data.

Yes. So next slide. And one of the things we've learned in our journey at IBM is complexity is relatively simple. Simplifying is really hard. And when you have a BHAG to be the most productive company in the world, you have to change the way you operate. And you have to get down to the atomic level of how do we operate transactions, how do we run our company. And what we realized pretty quickly was that data was the center of that universe together with workflows. And so you would think IBM being a tech company, we would start with technology. We started with data and workflow, and that's how we transformed our company. We became obsessive with data a few years ago. We created an enterprise data architecture. We have one for IBM, a set of data standards, master data management, governance and privacy so that we manage our data as a competitive asset. It's become the foundation of how we scale AI for the enterprise. Workflow, Joanne mentioned, workflow is a sticky, tricky thing when you've grown up over 114 years. Processes can sort of get stuck in the mud a little bit, and you really have to reassess and we had the luxury of sort of working with IBM Consulting to understand how do we break down a workflow, how do we focus on the business outcome and then radically simplify how do we get to that outcome. We've done that workflow by workflow, by workflow across IBM to the point now where we have mostly horizontal workflows that follow the flow of our transactions. We've come out of our silos. We embed technology throughout. We don't start with technology, but we use technology to enable our speed and our outcomes. And at IBM, Matt will share with you a little bit about how we've moved most of our internal workflows to our cloud. We've AI-enabled most of our workflows across the company. We have many more to go, but we made tremendous progress on taking the human glue out of how do we operate the transactions of our company. And it's often very difficult for large enterprise to get empirical fact-based data on your cost and your value. And we've really leveraged Apptio and Turbonomic to really give us empirical data on the cost of our workflows, the cost of our infrastructure, the cost of transactions and then how do we optimize that over time. We've made tremendous progress. It's really what's driven a lot of that $3.5 billion of productivity savings across the company. If you flip the slide, Nick. This is just a sort of a representative list of the IBM technologies that some of these we've co-developed with our product teams. We test, as Joanne mentioned, as alpha and beta, we help them scale products for large enterprise using the size of our data models. The watsonx capabilities, .data, .governance, Db2 Warehouse on cloud, Watson Orchestrate, Red Hat, OpenShift and Linux, IBM Z, our Granite models tuned for business, Apptio and Turbonomic. And we're also really starting to leverage our consulting advantage AI assets internally to streamline our workload. So this has been our formula for how do we transform the company, data, workflow, tech-enable, measure. And we've done that play over and over, and we're going to keep doing it over and over as we continue to eke out productivity improvements.

Great. Next chart. So as I mentioned to you, workflow by workflow, over 70 workflow use cases that we've actually deployed over the past 18 months to 2 years. What's exciting is we've infused AI into nearly every domain across the IBM company. Now as Ed mentioned, some of it we've gone end-to-end, and we've completely integrated and we're completely an AI enabled. And you could say we're done, you're never done in this transformation and AI-enabled story, but it is a level of capability. And so as we kind of go into this with you today, we thought we'd use four mainstream use cases and take you layer deeper. So HR support, IT support, financial planning and analysis, and then the whole opportunity and pricing analytics. And then one that we also think is one that which touches everything we do across the company, and that's what the work we've been doing on contract analysis. So first of all, Nickle, I'd like to hand over to you.

Sure. The HR use case was one of our first Client Zero use cases at IBM. And it is not where most companies think about starting, not only because of what Ed talked about is the data is very fragmented, but it's also a pretty high bar. It touches every single person in your organization and has very high privacy requirements on the data. But when we think about the value that we were going to drive, what we found was the environment we operate in, and this is industry agnostic, is the same for all HR and talent departments around the globe. And that is the complexity of compliance is increasing. Cities, states, countries, jurisdictions are getting more fragmented. No two sets of jurisdictions have the same laws that you need to comply with from a benefits employee perspective. Typically, that's something an organization would throw money at. You would put more people against that problem. But technology offers an interesting way to solve that without additional people. You also have to keep in mind that your employees' expectations about the experience that you're giving them internally is increasing every day. They are having amazing consumer-grade one-click experiences in their personal lives. As they come into the enterprise, they want those same experiences. Again, this is typically solved with people and money. But in an environment now where organizations are looking at extreme productivity so that they can fuel growth for our client-facing missions, you've got to think differently about how you do things internally. We consider ourselves to be an AI-first HR function. I didn't say AI only, but AI first. And it is enabled through a front door that we call AskIBM and AskHR. When we first envisioned this, we thought that this was simply going to be using our AI assistant, our traditional AI technology, and it was essentially going to be a high-powered chatbot, answering Q&A for employees. But fast forward to where we are today, much of what you saw on stage during the keynote, we internally developed the watsonx Orchestrate, everything around Agentic AI was actually developed with the HR use case. It will answer Q&As using generative AI. It will do transactions for you. You want to update your address, you want to transfer an employee, you want to give somebody a salary increase. It is all done in this chat interface. And finally, it does bring in Agentic AI. This idea of, you never have to log into a system. Last year, 50% of IBM managers and employees never logged into an HR system. They simply went into AskHR, the Agentic AI then used the transactions of all the platforms we use underneath. We did 11.5 million interactions with our employees around the globe in 52 languages last year. But the thing as you think about measuring value is that 94% containment rate. That means it didn't get escalated, a ticket wasn't created, 94% of the time, we solved this in-house. Yes, this did drive extreme productivity for us, contributing to kind of the $3.5 billion, lots of cost savings on outsourced vendors, capacity tools and systems. But it also drove innovation, because we do this all on the watsonx platform, I add every 6 days new functionality to the AskHR assistant, every 6 days. It is just running because it's all on that one platform. We can also partner with a lot of other HR talent technologies out there and surface it through the Agentic AI. The last point I would make, Nick talked about growth in our flywheel. And if you think about that is the way organizations are now thinking about talent experiences in their organizations is every minute somebody spends in an administrative process, sometimes stuck in an administrative process, is a minute they're not spending with clients or they're not building product. So AskHR was an opportunity for us to take that friction out of the system. And our Net Promoter Score with our employees right now went from plus 19 to we are sitting at plus 74. Try to find an HR department in any organization that is sitting at plus 74. And you'll see that, again, it creates this internal value that frees them up to go do other things.

Excellent, Nickle. So listen, Matt, I know that we've learned so much, right, in our IT team. So tell me what we took from this HR experience and they replicated...

Pretty much everything that Nickel said there about the experience people are having. And you could think that at a technology organization, everyone solves their own problems themselves. But I think really, that's only Arvind that does that. But Nick talked a few minutes ago about the IT transformation and broad-scopes and how that contributed to the numbers that you saw there. And we started to look through that and as part of getting all of our assets on the hybrid cloud, making it easier to build these solutions, we realized too, that we need to focus more of our time and energy in building these differentiated capabilities for IBM and AI. And so we start to look at that and say, well, why do I have this many people that are really focused on IT support, started to dig into the data and found out, for example, that only half of IBMers ever called IT support in any given year. We had benchmarked best-in-class numbers by about tenfold already with still having the internal staff. But we realized as part of being a digital AI-first company, that's not good enough. And so we took this as an opportunity as a challenge to see how quickly could we move with our IT support function and fully digitizing that. So we looked at the data. We used our products with watsonx Assistant, now moving on to watsonx Orchestrate. And started out within the IT team, pressure tested it on that and then deployed it to research, then through the course of 100 days, deployed it to the entire IBM population of about 280,000 employees at the time that we did this. This is a good example of being able to move very, very quickly with this technology, also learning as we go, and continuing to look at the data and using the top-down cultural change that Arvind and the senior leadership team were pushing, announced it to them as well as the bottoms up and getting the feedback from the team. Because just like Nickel said, every 6 days, they're releasing new HR capabilities into this. We take a similar approach with IT support. So then 10 months after we deployed it to all the employee population, I was able to turn off the phone lines. If you thought that didn't get me a little bit of hate mail at first. But you clearly haven't sat in my shoes. But what was interesting about this is that then we were able to pivot our approach because the team came to me and said, hey, our queue depth because people were still able to get to an agent to chat. And here, we did multilingual chat so that we could have a center of gravity of our IT support language -- language support and then using our watsonx technology to do that 2-way language translation, so you could talk to me in your native language, and I can respond to you in my native language even if you speak French and I'm only speaking English. Huge gain for us. But everyone is clicking that button, hey, I need to talk to an agent before telling us the problem, getting to unhealthy levels with the queue depth for the support agents. So we're able to turn the dial and say, Nickel, I need you to tell me what your problem is before you click that zero button in the chat interface to talk to an agent. But through this then, not only were we able to reduce the number of staff that are focusing on our internal IT support, and I back of an napkin benchmark with some other tech companies who are already like, I think the normal industry benchmark is like support agent to 400 employees. Other tech companies are maybe 1,000 to 2,000. We're about 1,000 to 4,000 in terms of how we support that, all of our population now. But now we're able to take that to the next level and redirect the time and energy and funds in order to develop the other AI assistant and agents that we're talking about today. I think this is critically important so that we can focus on what's going to be the value differentiators for IBM and spend less of our resources, but giving that higher level degree of productivity, just like Nickel said, giving time back to employees to accomplish the tasks that they need to do.

Great, Matt. Thank you. And I think what you're hearing is there's no shortage of data. So clearly, I talk about data as being a foundation for everything you do in an AI-first journey. So I think what we wanted to share with you, one of the big opportunities we saw was what we were doing in our whole finance planning and analytics. So Ed?

Yes. So it was just a few years ago, Jim had the vision to have a single source of operational truth for our company globally across all brands, all geographies. And at that time, we were highly decentralized in our data strategy. We didn't have data architecture. We didn't have data standards, and we said, well, how are we going to do this? And we figured it out. So we built internally an integrated data model that has all of our operational data that we use to run the company across finance, quote to cash, sales operations, marketing, HR, et cetera, in one integrated data model. We built this over time, starting literally with the first workflow within finance, with the financial forecast workflow. We didn't do a data dump. We didn't move the ledger into a data lake. We instead looked at what's the workflow that we want to transform. How do we simplify it? What are the building blocks from a data perspective of that data, and we move that workflow into EPM as we call it, our enterprise data model. We did that repeatedly over and over again. We started with finance. We then went to sales operations, then went to marketing, HR, quote to cash, procurement, treasury and many other of the functions in IBM to the point today where we have most of our operational workflows in a single source of truth, integrated at the refresh rate of the source system. So we pulled from about 50 different source systems. We have 30,000 IBMers who use it from the top of the business all the way through the business. We load at the atomic level, so you can pull at any level. It will all have referential integrity based on where you're pulling from. 300 terabytes of integrated data, and it's really changed how we operate the company. We've ended the data chase, used to be IBM, if you were a knowledge worker in IBM, you would spend days, nights and weekends chasing data and living in spreadsheets and eventually cutting and pacing into a PowerPoint for a review. Those days are over. We mostly work on the glass now. Our cadences are all run from our single enterprise data model, whether it's a meeting with Jim or it's a meeting with some analysts. It's the same data model, the same solution. And on that, we've been scaling our AI solutions. So with that foundation, a trusted source, we can scale internal AI solution within days. And that's really been the magic of our data story.

Nick?

Yes. And I would just to piggyback off that, I have the luxury of having the table set for me, right? We've spent a lot of time over the years from a finance organization, locating our resources in the right place, defragmenting our workflows, as Ed said, then aligning the data against those workflows. Where are we today? We're about half the size of a finance organization from where we were just a handful of years ago. And I see very clearly, we see very clearly a line of sight to do that again over the next handful of years. And we're at a point now where we have access to the data, aligned against the workflows where we can deploy at scale in rapid fashion AI use cases. Let me just touch on a few of them. Some are full on into production. Some are still in the experimentation stage and it kind of everything in between. Within the accounting space, within the accounting domain, RPA has been around for a long time. So like automation, robotics tools to automate journal activity. But historically, that's been a fairly cumbersome 3-in-the-box process. It requires your accountant or your domain expert. It has the technology, which is pretty heavy handed and code-based, think Python, and it requires a developer to sit in between the two of them to take a detailed description and write-off of what that actual journal entry process, that manual entry process is, hand it off from accountant to the developer to go translate into Python code. You can imagine for all of us that have dealt with developers in the past as domain experts, that takes an iteration cycle, a lot of back and forth in order to actually deploy these automations. Using Watson Orchestrate earlier this year, we launched a tool that essentially takes out that middle resource requirement and brings the domain expert directly to the technology in a no-code natural language process where the -- almost like a chatbot type flow-through, the accountant describes the process steps they're going through. They go to this system to pull this column of information, they map it against the same data set within this system, run certain calculations, do validation work, spit on an answer and book a journal entry, and they can walk through that in natural language without having to touch any code, deploy straight out of the system, get the approval from their manager, deploy out of the system and then monitor that activity over time. We still have today within IBM north of 200,000 manual journals that we do in a year. We've used Apptio, watsonx to analyze what all that activity is. And we have a very high degree of confidence that at least 80% of that can be automated here in rapid fashion. Within the finance domain, touchless forecasting. Any CFO conversation that you have, that's all I really want to talk about. What are you doing with touchless forecasting? For us today, in the enterprise, there's, I think, north of 70,000 data points that we, in an automated fashion, forecast. That's across geographies, across brands, across the P&L. And here's the really important aspect that I'd go back on something we said earlier. It's not a black box. I've been like railing away at this for a long time. Don't give me a black box answer. Even if it is 97%, 98% accurate, you need to explain to the individual, to the sales leader, to the finance leader, what are the levers, what are the drivers? Why is it giving this outcome? Got to go from prediction in a black box to prescriptive because we need to change the behaviors and adoption of these models. And we've advanced that pretty successfully across the company. Within the tax space, we've got our tax leader, Kanthi here. And yesterday or last week, we announced with EY, some automation work we're doing, leveraging IBM's AI product suite within the tax compliance space. Think of like 36 different systems that are being leveraged 1.2 million invoices a year that no human or no set of humans, no capacity could ever address that we can bring automation against. And probably the most advanced in terms of our AI deployments is within the pricing space. And we've been at this journey for over 5 years now. It has all the aspects of what we're talking about here today. Today, within IBM, roughly 70% of all of our transactional bids are handled in an automated, no-touch or low-touch fashion with no pricer interfacing with the sales leader and the technology. That's something like 70,000 bids on an annual basis. And we can see very clearly north of $100 million of incremental revenue realization that's come from two primary drivers within this space. One is the cycle time is a lot faster than having to deal and interface with a human pricer. When you increase your cycle time roughly 35%, you can get more bids out the door, more volume, more wins, more revenue. The second piece is that the tool uses 18 months of historical win-loss data across clients, across geographies, across the product set, think north of 100,000 bids, ones that we won, ones that we lost. And it's able to do an analysis, essentially establishing elasticity curve for that bid that you're coming to request. Optimization of price of course, mapped against propensity to win or win rates in order to give a predictive pricing outcome. Here's the really important part around behavior because we had this capability a few years ago, but it took time back on this point of explainability, back on this point of just teaching the seller that this is the optimal price and that the human nature is, well, I know my client, I know this product, I've got this intuition that I should price here. Well, guess what, the analytics actually does a better job at this. And with the adoption of actually using those prices and then being able to go straight through delegation and take it out to a client, again, with that cycle time, we can now see 3, 4 points of differentiation in terms of that revenue realization off of adopted bids that are utilized. That's unpacked within the $100 million. And you can see that all of this journey of all of these different capabilities from the technology, alignment of the data, driving behavioral change, moving on through to production, there's a long glide path of a lot of these initiatives to go from here. And we've got our table set for us, which is fantastic. So I think at this point, we're going to do a little demo. Matt.

We're going to do a demo here. Live test, see if the slides work. But let me -- before we go into that, let me pull together a couple of threads that you've heard throughout this. So in our productivity, IT has been a huge piece of this and just what I would say, IT optimization and transformation. As part of that, though now, I think some of you may have heard Rob's comments earlier on 1 billion new applications. So I'm looking at this as CEO is kind of dear in the headlight situation. You're telling me that I need to optimize my IT spend. I need to spend more in developing these digital capabilities to help humans be more productive with what they need to do. And now as you saw in the earlier slides, we've got 70 agents deployed at enterprise scale. That doesn't really count the ones that just aren't good enough to tell you about right now, but hundreds of those underneath the covers. So we kind of saw this early on. And as part of our Hybrid Cloud transformation, we had all of our IT expense going into Apptio costing. Now I could see where I have levers to pull in order to redirect investment like the IT support use case that we talked about. We start to bring in, as Nick alluded to, the non-IT information so we can track the efficacy of our digital workflows as we begin to measure the ROI of our use cases there. And then finally, it comes back to one thing. What is that human digital experience going to look like? And with all these agents all over the place, we realized that, that wasn't going to be a good experience if you don't know where to find something. It's obvious to go to AskHR. It's obvious if you're a seller to go to AskISC, but my workflows are becoming more and more complex. I need to be able to do context switching in the moment, okay, my better half just called me, how much vacation days do I have left, ask HR for that. There's a critical issue coming in, and I need to see what this next deal is going to be like. So how do we bring those together? And so what we're going to show you here early on in this, we launched something called AskIBM. This is a general purpose intelligence tool using generative AI that's built on our Intranet, which we call w3. So it is aware of IBM, IBM strategy. And the first iteration of this was solely just designed to give our IBMers access to these tools in a safe, controlled, secure way. We didn't want them going out to third-party tools and potentially risk exposing internally sensitive data. So we've put it up in a controlled way. But now what you're going to see here is that we bring together with the orchestration layer and watsonx Orchestrate, questions about HR, questions about IT, generalized questions, questions about sales and the current deals. So all stitch together for a data layer. So you've got your hybrid cloud, you've got your data layer, you've got security over that. And then our enterprise and AI platform powered by watsonx Orchestrate brings all this to life. So if you want to hit the button here, and I'll narrate as we go along. So first here, we're going to AskIBM. You can see options at the left there about chat, exploring, things like that. Fast, what is the capital of the United States? This is just a general question. AskIBM using the Granite models knows about this. It also understands about our Internet and you could ask it a question, help me prepare for a meeting with my boss about IBM's strategy. It answers that. But then they say, I need to know how to create an opportunity, this immediate context switching. So now it's going to AskISC, which is our seller productivity tool. It's telling them the steps of how to do this. This will mature over time for this in situ interface for that. Now I'm like, oh my goodness, I've got a problem with the device. How do I set up my Mac? How do I order a replacement device? And now we've got those steps on there. Behind the scenes, this is doing all the complex routing. I don't just have one application here as a chatbot. AskHR is managed independently from AskISC, from AskIT. It's bringing the knowledge. It's bringing the advanced business automation, connecting to our data platform and the systems of record and bringing generative AI to the forefront here to answer this question here about parental bonding leave, which is coincidentally because our leader for this space is going on parental leave. So her team did this as an acknowledgment to her as part of that. I think all of the IT organizations across the world are facing this complexity, hundreds and hundreds of assistants and agents. It's now part of our job to make sure we bring that human experience. And we see this as the orchestration capabilities as being core and foundational to where we're going next.

Great. Nick, do you want to share with us all the vision we have on enterprise capability of contracts?

Yes. Let's bring this home. And we talked about all that we've accomplished at this point, but we see very clear line of sight to a huge opportunity going forward. Now this is a little bit of a back of an napkin. So it's probably a little dangerous with an investor crowd here, but I just wanted to kind of paint a picture of what this can mean for an IBM, no different than any enterprise out there that has this challenge. So think about all the back-office infrastructure that all companies have this sort of G&A structure. What is the one kind of most important aspect that so much of those workflows surrounds, it's the contract, whether it's the commercial customer contract or it's a vendor contracts within IBM, we have around 700,000 live at any point in time contracts. Most of those customer, all of those written in natural language, many different languages. It's legal terms, so maybe it's not overly natural in terms of the language, but this is really hard nonstandard unstructured data where information needs to be extracted. Think for the purpose of cutting an invoice and collecting for knowing in a delivery context of how to fulfill that contract. From a legal standpoint, what is -- where the risk reside, what are the liability aspects within the M&A space, how do you do due diligence and identify toxic terms of potential targets that you're looking at across their entire contract set. Procurement, same thing, make sure that what we purchased is fulfilled that we've made or purchases on the right time. Tax already talked about that a little bit, 1.2 million intercompany invoices around the world, all of this through the language of the contract. So the technology exists to be able to extract that insight in an automated fashion. Today, within IBM, something north of 2,500 resources, $175 million is spent on the human glue to go into a contract from an accounting -- I didn't mention the accounting case, but actually to read the terms of a contract and ensure we're recording properly to the ledger. So much of this can be automated. And not only that, but the power of this is, yes, there's an efficiency side savings. But today, you only have so much capacity. It's no different than any company. You only have so much capacity to actually manually review these contracts. So you establish clip levels, you end up only looking at a subset of the activity. Kanthi's team can't look at all 1.2 million invoices. But guess what, robotics and automation infused with AI is capable of doing that. So there's a spend savings. There's an improved risk profile and risk environment because you can cover more, which is music to my ears as controller. There's a true win-win-win opportunities. And IBM, we're big, we're complicated, but we're no different than any enterprise out there. And you think of the, say, billions, but probably actually trillions of dollars spent in G&A huge opportunity to unlock 30%, 40% value. Very clearly, we see that at minimum productivity is a massive opportunity for us as a company to continue on this productivity curve, but also for us as a vendor, a trusted adviser for all of our clients. So Joanne, let me hand it to you to take us home.

Thank you. So our story on Client Zero really is bringing IBM strategy to life inside of IBM. So how do we become the best and most enabled Hybrid Cloud and AI company. I think you imagine that we are highly penetrated having delivered for Jim, the $3.5 billion with a much faster return to market and also an improved user experience. But we estimate we're about 50% of the way there. And why do I say that? Because what you saw from the team is there's elements of where we've transformed the whole workflow and then we're enabling the experience to be the most productive vision of ourselves. But there are areas where we've gone within a certain element of the domain and with more work to be done. I think the vision that we've got is that this year ahead, we're going to continue to accelerate our own productivity, that we are lighting up our team. I will say the one piece that we didn't share with you, I mentioned we were -- sorry, tops down. The bottoms up is a real movement. It's coming from the vision that all of us want to be a more productive version of ourselves. And we want to be curious. We want to be growth minded. We want to use AI. And so each year, we run a 1-week long rally where we ask you, what would you use AI for to make yourself more productive, getting incredible insights from our sellers, from our developers. And I have to say, because we feel so passionate about the vision that we have for shared services, we've changed our name to growth enablers because we believe when we do this well and we do this right, we actually will not only fuel the growth investment opportunity for IBM, but we'll also make our teams that really need to develop product and actually to be out with our clients and partners more productive version of themselves. Clearly, integrated value of IBM is bringing technology and consulting together as one. And so I wanted to share with you, last year alone, with this leadership team and a layer perhaps one level below, we united about 500 client stories. So where we go out with our sellers and our consultants and get to share our Client Zero experiences, predominantly focused on the domain that our clients most want to transform or the one that's the biggest pain point in their business. Is this going to ignite? I would believe so. Here we are having closed out the first quarter of this year, and we've done 450 in the first quarter alone. So this is a story, I think, that all of our clients and every industry wants to hear. They want to understand how they can ignite their productivity. I will tell you, my peer-to-peer conversations are they want to hear about the good, but they clearly want to hear about the bad and the ugly. And they do clearly want to understand the pain points and the opportunities and how do they avoid them and how do they drive the most impactful ROI because extreme productivity is what this journey is all about. So with that said, we'd love to turn it over to you, answer any questions or any points of clarification that we can do.

Why don't we start with Amit.

Amit Daryanani, Evercore. I guess I had two questions. Jim, as you're sitting there, how do you square the $3.5 billion of savings you have had with the PTI margins the company is going through? And if I think about this year, I think the guide is for 50 basis points, at least 50 basis points of margin expansion. How do I square that? Does that imply that a lot of this is getting invested back? Or you have a lot of levers that if demand goes sideways, you can actually defend your margins and free cash flow very well. Just talk about that investment cadence. And then just broadly for the group, how scalable are these solutions that you have today? Can you put a bow around this and sell it to your customers as the AskHR, AskIT? Or what is the big kind of challenge you have to do to get this to become a more scalable product for your customers?

Jim, I know you love multipart questions.

So thanks for the question, Amit, upfront. This flywheel we talk about. If you look at the shareholder value creation thesis that we laid out on February 4 at Investor Day, right, a huge component of our financial investment thesis is around productivity. Why? One, it enables that margin and operating leverage. But two, it generates that free cash flow generation engine that we've talked about that is important. And we've said for a long period of time that we've got the ability to toggle up, toggle down on how much we invest or how much we take to the bottom line. Typically, historically, we would operate about 1/3 goes to the bottom line, about 2/3 gets reinvested in the business. By the way, that reinvestment, as Nick stated, is both through innovation organically and inorganically. It's through go-to-market. It's through opening up IBM to ecosystem partnerships. It's through service delivery talent and consulting and on and on and on. Now dial back 9 months ago, what were all the questions on the earnings call? We love the Hashi acquisition from a perspective of the innovation value and the synergistic value of HashiCorp, IBM plus Red Hat together as an industry-leading Hybrid Cloud platform. But the question that was always asked to me, particularly as a CFO, God, how are you going to cover all that dilution? How are you going to grow earnings? How are you going to grow margins? How are you going to grow free cash flow? And we stated the acquisition in and of itself stands on the strategic value and the synergy, but this -- what we just took you through was the confidence that we had in driving the productivity that would enable us to come out in January and give a guide in the year all in, growing margins north of 50 basis points. And back in January, we took the Street well up on free cash flow to about $13.5 billion. So when you look at this year, we started out first quarter about 50 basis points of margin expansion. We said for the year, north of 50 basis points. That includes covering about $300-plus million worth of dilution impact in Hashi plus reinvesting all of our organic innovation. So you cut to the chase, this productivity, I think Nick talked about, first quarter was 350 basis points gross. We reinvested over 200 basis points first quarter. The rest of the year, we'll get 350-plus in gross, and that reinvestment is probably going to go up because we got a full quarter, a full 9 months left of HashiCorp going back forward. So that's part A. Part B.