Hello, everyone. My name is Mayank Tandon. I cover IT services at Needham. I'd like to welcome Mike Thomson, CEO of Unisys. Mike, thank you for joining us.
Oh, pleasure to be here, Mayank. I love to have an opportunity to speak to the team. It's great to be here.
Terrific. Mike Thomson, I thought maybe for people who might be newer to the Unisys story or maybe not fully familiar with the evolution of the company, I thought we could just start with a quick overview of the business as it stands today, and then we'll dive into, obviously some hot button issues.
Sure. Yeah, look, it's always interesting to do these types of things and bring new folks on, 'cause we are 153-year-old company that spans a long time. I'll truncate that to maybe the last 50 years and our experience from a technology perspective. You know, clearly, back in the, I'll say early 19 or mid-1900s probably is when the mainframe server component of the business has started. We were a manufacturer back then from a history perspective. Kind of pushing over the next, I'll say 50 years into the early 2000s was, you know, primarily a mainframe-oriented organization, a lot of combination of applications of Windows and Linux and things like that. Really from the 2000 started to shift from manufacturing into services.
If I fast-forward us to today, we're fully an IT services company, really ranging from everything in the installation and break-fix of hardware, software, both in infrastructure and, you know, on the PC side of the equation, all the way through agentic application development and modernization. Think of us as a full service ITO, if you will, and really everything in between. Global footprint, have about 15,000 associates worldwide, service clients in roughly 50 countries around the world, again, doing full service ITO related technology.
Mike, I'm gonna jump right into AI because that's obviously the hot topic.
Yeah
for everyone concerned. Maybe you could and take your time with this however you want to address it, and I have some follow-up.
Sure
questions too. If you could just talk about how AI is impacting different segments of your business, what are the client conversations like, and from a revenue standpoint, have we seen any impact yet or is that something that's going to come down the pike? I know a lot in there.
Yes
I'd love for you to address the AI angle to the Unisys story.
Yeah. Well, thanks for that. Yeah, it's a pretty meaty topic, so I will kind of weave through it a little bit. Please, if you want me to expand on anything in particular, chime in and we can make this a good dialogue. In general, obviously it's the issue du jour, and it's all everyone talks about in the market. I really like to think about AI, and our company thinks about AI as a multitude of things. In the press we talk about AI and everyone just uses it as it's singular.
From our perspective, it really has to do with both physical and infrastructure related AI, and maybe I'll go through some of those components and then talk how our business kind of ingrains into that particular element, right? I've got physical and infrastructure AI. We've got data AI. There's governance and responsibility AI. There's security and design elements from an AI point of view. There's AI processing, right? If I think about and tie into how we started this conversation, we're a full service IT services provider for our clients.
That means as something as rudimentary or basic, and it's not rudimentary or basic, but the installation of AI and liquid cooling and immersion cooling for GPU chips, right, for large manufacturers and data center builders, et cetera. The managed service component of that infrastructure build, high-end storage, high-end compute, where that gets done, whether it's at the edge, whether it's on-prem, whether it's in a private or a public cloud. Everything in between and all of the Io,T devices, the procurement of, management of, maintenance of, predictive analytics around the self-healing and distribution, all of that would be in my mind, the component piece of what we would consider process AI or infrastructure AI, which is kind of a very foundational level.
The other aspect of that from our point of view, I think starts with your AI strategy. That strategy has to be determined really at every one of those levels, right? There's not one AI strategy for a company. There's an AI strategy for their process AI. There's an AI strategy for their data AI. There's an AI strategy for, you know, their infrastructure AI. They're all at different levels of maturity. Really, it's a very complex ecosystem, especially when you talk about enterprise clients, which is who we support, you know, typically public sector, state and local governments, large financial institutions, you know, folks like that. They have a very high complex ecosystem that needs to be managed.
When we hear about AI today, most of what we're talking about are frontier models. you know, there's lots of dialogue around the utilization of Anthropic and utilization of OpenAI or GPT-5, et cetera. The thing that people don't speak about, which is interesting to me. The ecosystem where we play is to support the entire stack, right? Although you've got these wonderful tools that can do some incredible things in the application or utilization of a large language model to write code, you have to take those applications that are written or those agents that you've written to do tasks, and you've got to embed them into an ecosystem.
I think about the opportunity for a company like Unisys as an integrated services provider, the stack that we support gets way more complex, not simpler. It's not as though clients are taking on this new agentic AI application development or agentic agent development tier and just dropping it into the ether, right? That has to sit in a hybrid ecosystem that has compute needs and compute ramifications, frankly, when you think about the utilization of tokens, how they get consumed, where they get consumed. The strategy on how you deploy those types of things, how they sit in your ecosystem, and how they interact with all of the other components of the stack that we support, is really where the opportunity lies.
Although all of the press and frankly all of the valuation is being applied to the chip manufacturers or the OEM component of that or somebody on the data center build or on the frontier models themselves, to deploy any of that, you really need the IT services provider because it's not as though you're talking about a stack that is native and you're and you're only applying it to one thing. It is a very small element of an integrated stack that will have to change over time. I see this frankly as the real growth vector for IT services.
over the course of the next 5 years, right? I think it's gonna continue to really be what drives the industry, I think, back to growth. I mean, you know this well. The industry has historically grown, you know, take the last 2 years out of the mix for a second, historically been in that 5%-6% CAGR growth. Over the last 2 years, especially for IT services, some SaaS companies, et cetera, we've seen a shrinking of that market and kinda headwinds to the short-term revenue because of the adoption of this emerging technology. The reality is the next 5 years, and we're seeing some of this growth start to happen already, the projections are to kind of revert to the norm.
I think it will be the main driver for growth for the industry to get back to the historic means of growth in IT services in general.
That's a great summary, Mike. Let me extend that question to some of your segments. Could you talk about, and I don't know if it's already impacting your revenue, but at least in terms of client conversations, how they're shaping up both in terms of LNS and XLNS. I'm asking you.
Yeah
a broad question, but maybe if you wanna take it by different segments on how AI is actually impacting conversations, 1, and, potentially, revenue opportunities.
Well, I'll start with conversations, and then I'll get specific to some revenue if it's okay. We don't have a conversation today with any client that's not started with AI. I will tell you the talk track is actually different now than it was even a year ago. We have had, over the course of the last year or two, been implementing AI in all of our solutions, and the dialogue was always around kinda where that AI is embedded and what it's doing for our clients, right? Now the dialogue is actually a lot more around the services component of AI and how that gets applied. It's not enough to just have AI embedded in your solution. That's almost table stakes at this point.
Those dialogues are really around what is the future use of AI and how does it either amplify the solution or how, what's your solution roadmap look like from an AI perspective? You mentioned our two kind of the ways we look at the business, and not technically the segment view, but LNS and XLNS, and I'll get into both of those specifically. LNS is our historic, a ClearPath Forward ecosystem. If we talk about the impact it's having on revenue in that particular business, and has been for a couple years, the availability of our ClearPath ecosystem to adopt modern languages, to use open source, to build applications, agentic and other, on top of our systems.
We have AB Suite and AI developer kits that really enable the use of a platform that has been in existence for decades, has tremendous structured data, and is aligned to the use of frontier models to actually build better outcomes for clients. We're seeing a significant increase in consumption from a utilization perspective. We talk about it in terms of tokens today. In that world it's really about MIPS, but it's the same thing. It's consumption use of data.
We've seen over the course of the last three years in that business, roughly about a $40 million a year uptick in revenue, based on consumption use because of the tooling and allowing it to utilize kind of the AI front end and/or the data exchange to AI native applications embedded with the traditional ecosystem, if you will. Really powerful there. That is a, you know, highly profitable business for our company. It's roughly 70% margin business when you think about that, and it runs about $400 million a year from a solution point of view. Seeing really strong consumption growth and longevity of that specific business because of the ability to have it interact, you know, with modern technology, right?
Consider it a modern ecosystem that has your traditional core, as well as the UX/UI component that are AI oriented, and a data exchange between the two. It really marries up nicely with the historic premise as well as the futuristic component of it. If I flip my hat to the XLNS side of the business, which is really everything else that we do outside of our IP within ClearPath, you know, that's where we talk about our digital workplace experience. That's where we talk about our cloud infrastructure and applications team. That's where our security practices, our apps modernization practices, you know, those types of things.
Clearly embedded in those solutions would be, as you would expect, AIOps and FinOps decisions on the management of your infrastructure, how you're looking at your security practice. I mean, you've heard all of the dialogue around Mythos and the types of things that it has brought to light in regards to security vulnerabilities, et cetera. You know, and the impact that has on folks like us that run SOCs for, you know, multiple clients, right, when we talk about managing their security perimeter, et cetera.
That is an impactful piece of the business and part of the point of spear by which we have dialogue with new clients around utilization of AI embedded in their SOC and the stack that they use to manage, detect, and respond to cyber threats, which, you know, obviously continue to get more sophisticated and much deeper. We view that as a pretty expansive component of our future state as far as the application of AI. We've talked a little bit around agentic AI and if you think about the adding of agents to essentially transition some of the workforce to be a digital workforce, to be married with the human capital that supports the client infrastructure.
Clearly the impact today on the creation of code and the ability to create bespoke applications that are fit for purpose for our clients are all elements of where our dialogues with clients are going, and where we expect to see the growth in those particular businesses to continue to extend, I, I'll say our footprint. We do plenty of apps modernization today, but that has a new flavor when you start thinking about utilization of agents to not only build, but run those applications. Talked a little bit about the compute. If you think about even the use of tokens and when you're using those tokens, where you use those tokens, in a public cloud, in a private, you know, scenario, whether you're doing it on the build, whether you're using it in your run.
Those are FinOps discussions that have to be thought through before you just start consuming, right? Because I think what people get a little shocked by is, you know, the cost and the use of those tokens far outweigh the labor cost of what you were using the people for. This is not a simple model of replace human labor with digital labor and it's cheaper. It should be, and there are elements to that, but you have to have that AI strategy on how you deploy that. That kind of sits in our cloud applications and infrastructure business segment.
Of course we talked a bit about the field services component. That's kind of our core offering in our DWS, but we've got an agentic service desk that has a whole knowledge curation and knowledge management, multi or omni-channel input, natural language processing, embedded generative AI for the utilization of that service desk. The knowledge creation, the knowledge management, reducing average handle time, bringing first call resolution to our clients, preventive healing, preventative maintenance, when looking at those data analytics and the corresponding of data elements, which all speak to experience level agreements that are embedded into. You know, ultimately it's about a better client experience, a more consistent client experience, a faster client experience at a value that is less than what they've paid for traditional service desk. Right.
That is just in an offshore location, right? That type of thing. All of our discussions and the conversion of our legacy base of clients to our new solutions, and then layering on these service elements is really where I think the industry is going. Again, these are extremely complex ecosystems. They're in highly regulated environments with decades worth of security redundancy, you know, transparency, auditability. You know, you really do need to marry the technical expertise with the industry expertise, as well as the client centricity and the understanding of the client's ecosystem to be a real service partner in the development of the future of our clients and both current clients and prospective clients.
Very helpful, Mike. I'm curious, on the AI side, just given the pervasiveness of the technology, have you had to change your go-to-market strategy? Maybe if you can tie in how you're leveraging partners as you try to, you know, gain market share, especially when it comes to AI solutions in the market.
Yeah. That's a great question. I would, if I could, I'm gonna add one more to it because I think it's before you get to go to market and partners, we really should be talking about the associates that work for Unisys.
Because it starts with an upskilling of the associate base at every level, at very different persona levels, right? The upskilling of our field service technicians to be the cabling and liquid cooling element of the installation, that's an upskill. When you think about a service desk agent, level 1, level 2, moving to a knowledge manager and really talking about knowledge curation and QC-ing of elements and building out those models, that's an upskill of the workforce. For us, thinking about it through the lens of AI first on everything that we do, from an accounts payable clerk to a distinguished engineer and everything in between, is kind of that starting point.
We, as you know, have had a capital light strategy for quite some time, which means we have a leverage into our partner ecosystem, which is super important when you think about, you know, people, process, and technology in general. We've made decisions years ago that we're not going to try to build all of this, right? We want to take the best of breed of what everyone else is building and bring that to our clients and be agnostic to what is most useful for them, right? We're agnostic to the hypervisor you use. We're agnostic to the OEM hardware provider that you wanna use. We will be agnostic to the frontier model that you wanna use. We're opinionated in what works best together, and we have the experience to use all of those tools, all of those hypervisors.
We bake those into what we consider a single pane of glass, and we have the API connectivity to all elements of that stack. To do that means you have to have those partner relationships with each element of that stack. We need to understand where our partners are going for, you know, their roadmaps of where their products are going. We need to understand our clients' objectives and where their roadmaps are going, and we need to be able to, you know, stitch that together and kind of be the glue to kind of pivot their technical debt over time to continue to take advantage of the emerging technology. All of that has to happen.
before you can even talk go to market. Clearly that is the kind of last leg of this, right. We use AI, frankly, in our go to market. If I think about how we target clients for opportunities, getting in front of those opportunities with thought leadership, understanding their intent from a buying perspective, understanding where they have weaknesses in their current ecosystem, and having positions on with our deep industry knowledge or our deep, you know, vertical knowledge of their particular business and great client quals, we're able to have that type of discussion that is meaningful and deep.
The area has really been around getting as far in front of the intent as possible, bringing in the emerging technology, using it as a disruptor to have a conversation. The, you know, the beauty for us is when we get through that first conversation, you know, I can point to my top 50 clients that we've serviced for on average over 20 years. I have wonderful client relationships, wonderful client quals that can illustrate everything that we do in production, right? These are not POC issues. This is how we apply emerging technology to production and have been doing it for decades. I think that's a really powerful story. That ability to bring that into the go to the market element is super important.
Using the technology to really focus on who we're going after and very succinctly why we're having that discussion and aligning that to our solutions and ultimately the value to the client and base that on the outcomes that they get.
Given all that, are you seeing any changes in the way contracts are structured and the pricing around these contracts? I know, again, I'm asking a very broad question. It'll obviously differ across the different segments. To whatever extent you can answer that.
Yeah, no. Look, it's really a good question and the short answer is yes, we're seeing changes in pricing for sure. We've talked about this pricing headwind, right? If I'm going to a client and I'm transitioning their traditional IT service desk to an agentic service desk, the expectation is it's going to be cheaper for them to do that because it's cheaper for me to run it because I'm using digital workforce in order to do that and the use of technology. There is this We're kind of in this weird period of we've got to cycle through the base of clients from a renewal point of view, and I'm talking about Unisys in particular, and for new logo acquisition.
Even though we're expanding our market share in that space, it is at a lower revenue price point. When you think about what's made up of that, it's not necessarily that just because it's cheaper for me to deliver that service. It's really changing what the outcome is, right? A lot of the dialogue we have is really about outcome-based pricing. We're not talking any longer around we're charging you for number of seats or number of tickets processed or things like that. It's really about the experience that your associates or the employees of our clients are getting. Their uptime, their productivity, their persona mapping, and what devices they get, and how many things we've deflected.
I mean, if we really do our job better and better and better, we should have less and less tickets. You certainly don't want to base your pricing model off of the number of tickets that you've serviced, right? We want to base it off of things like, you know, 40% deflections and, you know, 20% self-healing items, right? That your uptime is way more and we know what assets you should sweat, and we know what assets need to be replaced, and we do preventative maintenance based on analytic data as opposed to waiting for a breakage or an outage or things like that, right? I think all of the dialogue is really more around what is the output and is the value they're paying worth the output they're getting, right?
As opposed to kind of a commoditized number of seats, number of licenses, number of tickets, that type of thing. That is really prevalent today, and I don't see that changing. I see it getting smarter. I mean, we've been on this train for a little while now, converting SLAs to XLAs in our contracts, and it really makes sense, right? Most of our contracts in that space are on average 3 years or 5 years, something like that, or 3 years plus some option years. If you think about when you establish the SLA at the point of a contract versus where you are in year 5, measuring the same thing makes no sense, right? The experience level agreements, we're basically changing what that looks like throughout the contract life cycle.
When we get really good at something, well, we're not measuring that anymore. That's not adding additional value. Let's find something else that adds value that we'll use to measure that, and I think that makes good business sense for our clients and good practical sense for us.
Got it, Mike. Maybe I'll switch to, some number questions. You know, you guided, and you reaffirmed the outlook, just, what? A week and a half ago or so, if I have my timing right.
Wow, yeah. It seems like a lot longer than that.
I know.
Yeah, you're probably right.
It's like a blur. Maybe from a visibility standpoint, you know, how are you viewing demand? I ask that more because obviously we've all seen the news and there's been some geopolitical unrest and some companies have pointed out to macro, maybe elongating sales cycles. Have you seen that at all? Just wanted to get your thoughts around the overall environment.
I think, you know, if you followed our story for the last 18 months to 24 months, I think in general we've all seen a little bit longer sales cycles. We had indicated at the end of last year, Q4 last year, and also in Q1 this year when we had our dialogue, that it seemed like that was softening a bit and that we were getting some of that discretionary project work back into the cycle. I mean, if I looked at our quarterly results, our total contract value year-over-year was up about 33%. We had new business TCV up almost 45%. Our book-to-bill at 1.2 times. Increase in backlog.
I had mentioned the top end of the funnel was starting to pick up again as well. I think, and we've been saying this for a while, the conditions that fed the macro, at some point people had to get on with their lives and kind of go, "Okay, we don't know what's gonna happen in the macro, but, you know, we have to replace devices. We've got to do these projects. We've got to embed this emerging technology into, you know, our work if we're gonna transition our technical debt in the manner." There's just project work that has to be done. We're also off the heels of a very heavy, unusually heavy renewal cycle for us.
I think in, at the end of last year, we talked about $1.7 billion worth of renewals, which is probably two or three times our normal renewal cycle. I'm talking on the XLNS business here, which is our kind of managed service contracts and that types of type of thing within DWS and within CA&I. When you're negotiating a renewal cycle and you've got the complexity of what you just talked about, which was changing the pricing, changing the outcome based, you're not really negotiating new project work, right? You're trying to fix this three-year deal in or this five-year deal in.
On the heels of being through that renewal cycle, I think what we're seeing is a little reversion to the norm of the new scope and expansion work that we get in our existing base. 1 statistic I usually use when we're talking about that is that, you know, I've got about a $30 billion TAM in the existing base of clients that I service today and have serviced for decades to penetrate for services that we could perform that they're either doing themselves or someone else is doing on their behalf. You know, that gives us a lot of opportunity inside of the base of clients that we have to continue that expansion, which is really healthy, and we have, you know, wonderful NPS scores and really good connective tissue. You asked about go-to-market.
One of our go-to-market motions that we've changed over the last year or two is really about initiating much more C-level contact because what we found is that we have great relationship and great support in that client base for the people we work for. Perhaps the decision-makers are not fully aware of our skills and what we can do and how we can cross-sell, and the relationship that we've got with those accounts. You know, I mentioned to you in the pre-call here that I'm at a client forum. One of those client forums I have is a CIO/CTO forum where I have about 45 CIOs or CTOs in that in that forum where we have regular discussions.
I think those types of things help the decision-makers at companies understand what we do and how impactful we can be to support their organizations. Then you couple that with the bottoms-up view of what we have been doing for a lot of those clients, it opens up a window for us to really expand our footprint, we've had a good amount of success doing that.
Great, Mike. I know just given the limited time, we'll maybe do a quick switch to, the other big, elephant in the room.
The pension obligation. I think you've made tremendous progress on that. Maybe if you could give us just a quick update on where you stand, especially for people who might be newer to the pension story.
Yeah. Sure. Look, I don't even view it as the elephant in the room anymore. Look, I'll just go back and do a quick snapshot of where we were 10 years ago to where we are today. You know, 10 years ago, we were staring at a $2.2 billion pension wall of contributions that had to be made over that 10-year duration and kind of a buyout value of well over $2 billion. That deficit amount today is on the U.S. plan, is about $400 million. We've made tremendous progress in alleviating that deficit. And we've done a lot to mitigate any volatility in that pension program.
We've shifted the assets to kind of fixed income assets and aligned those to the liabilities, took the volatility out of the pension contributions in the future. I think we're on a trajectory to totally diffuse that pension obligation, you know, over the course of the next 3 years, 3-5 years, something like that. On a defined path and removed all the volatility. From my perspective, and I don't mean it to say it's simple, but it's on autopilot at this point. We have done 6 or 7 annuitizations, which is selling off those pension obligations to insurance companies, some of which we've been able to do below par, which is great, right? You're giving away more liabilities than assets.
Those come out of the trust, not out of the company, cash portfolio, if you will. They also come with reduced premium payments because you've removed participants from that plan. We've been able to honor the commitments that were made to our pensioners decades ago, and we've been able to mitigate that exposure. I think we're on, you know, kind of an autopilot path to full diffusal of the U.S. plans and feel really good about the progress that we've made, you know, shaving roughly $1.6 billion off of that deficit over that time period, which is, you know, I, I think kudos to the team and really aligned to supporting our capital structure.
Given the progress you made, and again, that's been really nice to see, and also some of the initiatives on the profitability front to improve margins, do you think you're at a point where you can maybe start to explore other ways to, you know, deploy capital?
With our capital? Sure. Yeah, look, lots of opportunities to do that. I mean, clearly we've had some acquisitions in the recent past. Prior to that, it was decades before we did that. Where we had a need from a solutions perspective, clearly there's a capital deployment to do that. We talked in the last call about a little bit of debt repurchase, that's obviously a utilization of capital. In a perfect world, I'd love to get to a scenario where we can do some equity buybacks and those types of things as a program. Clearly we want to be on the forefront still and always looking at M&A opportunities to grow the business as well.
When I started here, we were spending about $265 million per annum on just pension contributions. That was the cash outlay per year. We're gonna be, you know, under $100 million and probably more on average about $70 million per year from a contribution point of view over the course of the next couple years. That frees up a lot of cash to do things. You've mentioned the profitability movement that we have. I think over the course of the last 3 years we've had about 600 basis points of margin improvement in our XLNS business, so that's driving more cash.
The goal has always been, let's get to cash flow positive, free cash flow positive, not adjusted, not, you know, like actually accretive cash on the balance sheet, and put ourselves in a strong foundational position by which to grow the company with some inorganic and some shareholder return and getting the right level of debt refinancing. You know, that type of thing. It's nice to actually be able to think about how we're going to utilize that capital prospectively. Some of that is in today's world, you know, we're seeing our SG&A value, the core SG&A is coming down, but I mentioned to you the upskilling of our associate base. Well, you know, that costs money.
That's an investment, but I think that investment will pay off in, you know, clearly, the growth of the company and the support of our client base. Areas like that I'd much rather spend that capital in growing our business, than, you know, making a pension contribution when, you know, only about 3% of those pension years are actually working at the company. All the rest are kind of in retirement status. Flipping that narrative has been, you know, a long road for us, but I think we're certainly in the later innings.
Excellent. Mike, I know we're almost out of time, but I know you also have the Investor Day coming up soon.
Yeah
look forward to that. I don't expect you to give away the secrets, but will that be? I'm sure there'll be a lot of AI discussions.
Of course.
at that event.
Look, I, as you know, we do these every 2 or 3 years, we'll give another kinda 3-year run of where we think we're going to be and where that inversion to growth comes in. It's really about my goal is that the investor community and the analyst advisor community leave that session or sessions with the viewpoint of how we play in a future state market and how the role of AI is within Unisys, how we're thinking about that. I, you know, we still battle this, "You're part of a traditional IT services construct," and I really need folks to think about that's not really who we are.
We have to have a foot in both camps because most of our clients have a foot in both camps. Our future is really around an AI-oriented services company that has the skillsets to support everything from quantum to, you know, mainframe, and everything in between, and I think that's a pretty powerful story.
Perfect timing. That's a great way to summarize that, Mike. Again, really appreciate the time, the discussion today, and I'm looking forward to attending the Investor Day in a couple of weeks.
Great. Well, thank you for having me. Been a pleasure, and look forward to seeing you next month.
Perfect. Goodbye. Thanks, Mike.
Take care. Bye.