Fidelity National Information Services, Inc. ($FIS)

Good afternoon everyone. This is my favorite part of the Mizuho conference every year is the fireside chat with Stephanie. So thank you so much. Stephanie, I have the honor of having Stephanie Ferris here, President and CEO of FIS -- and welcome.

Thank you.

Thanks for being here. Well, I have a lot of questions, and then we'll see where it goes. But how about we start and kick it off, -- maybe let's talk a little bit about the demand environment, tech spending for banks. Has the environment changed at all with some of the geopolitical uncertainty center.

It's a really interesting time. What I would say in banking is we are in the middle of a generational moment in banking and financial institutions. There's 3 big things that are happening that are significantly driving a period of growth in banking. I think we have not seen since the -- before the great financial crisis. The first is the amount of deregulation. So if you think about -- and this is true globally as well in the U.S., there's been a pretty big rollback in terms of global regulation and local regulation so that banks can now participate in economic growth. And that in itself is pretty big and pretty important. The second thing I would say is the bank's ability to grow both organically as well as inorganically. And you've seen M&A come back at record levels. unlike we haven't seen before and probably over a decade. And I think that's been long overdue in terms of smaller banks as well as large banks wanting to grow acquisitively. And you have definitely a set of buyers and sellers. And so you have a great opportunity where the regulatory costs are coming back. banks can invest in growing their franchises organically. They can invest in buying and growing their franchises inorganically. And then the third pillar is with the advent of AI technology. And for those of you who haven't been in and around the banking system, they typically absorb new technologies with a wait and see. And what I would say is they're pretty excited about the opportunities in front of them with respect to AI. So those 3 things together are a generational moment. So when you think about FIS, Dan, as you know, really worked hard to simplify our story and serve financial institutions purely. And so the underlying market and client and customers that we serve are growing at record levels like we haven't seen before. And what that means is they have money to invest back in their business and their franchise. And clearly, technology is a place they're doing at.

And maybe let's touch on something very exciting, which is the entropic part that you have -- and can you help us maybe think through how your banks are approaching AI? And what is the appetite for adoption? What areas they're focused on? And is there more opportunity for cost cutting or revenue generation for them?

Yes. So let's zoom out and talk about AI within banks and financial institutions. I had mentioned that they're pretty excited about the opportunity primarily for 2 reasons. One is because they do want to use the capabilities to grow revenue inside the bank, whether they want to drive digital currencies that they're now allowed to participate in or tokenize deposits or grow through more in the top of the funnel, have more abilities to grow digitally at a lower cost. So think about them growing their franchise. They're spending a bunch of money like probably every company is in terms of on AI and driving those really high return on investment areas. At the same time, they're very focused, and this is more in the large banks because from an FIS standpoint, and we in the industry provide these capabilities out to regional banks and below, they're looking at costs that are inside the franchise that they've been unable to unlock for a long time. So think about regulatory and compliance. And the one example that we're doing with Anthropic is financial crimes. So in financial crimes, you're required to file suspicious activity reports. The requirement is no different, depending on a big bank or a small bank, it takes a lot of regulatory costs and human capital to do it and getting it wrong is quite punitive for a bank. So there's a really large cost structure inside financial services that has never really been able to be tackled. There's technology that has provided the reports to do it. But with what we're pursuing with Anthropic is really getting after unlocking a lot of that cost and then when you talk to banks, they want to take those savings and again, to attribute them back into building the bank. With respect to Anthropic, we're very excited about that partnership. As you know, Dan, we announced it at our Emerald Conference, which where we have over 4,000 clients attend, they're very excited. So Anthropic came to us -- they've been working with the largest banks, as you know. And they were very focused on helping unlock these technology costs with banks, but what they quickly figured out and why they needed to partner with us is you need a lot of regulatory and compliance know-how to be able to do something like that. So we have a partnership where they are contributing for deployed engineers, and we're working alongside them to learn how to build these types of agents. They are proprietary agents of FIS and then we will distribute those agents throughout our banking clients. And that's just first of many agents to come, but we're pretty excited about it. We have two bank partners that are launching with us BMO and amalgamated and early signs in terms of the cost savings coming out of this particular financial crimes agent is between 70% to 90% of cost savings. So we think this is a big opportunity for FIS to deliver and really TAM-enhancing capabilities into these banks. And next up, thinking about things like card fraud and disputes or item processing, places where FIS naturally resonates as our technology sits there inside the bank, and there's big laborious cost inside of a bank that they'd like to take out and redistribute.

And maybe stay on the topic of AI, which is obviously very topical. So how concerned are you about AI advancements driving banks to in-source and build more solutions internally?

Yes, literally 0. And here's why to think about that. If you're a very large bank, you're JPMorgan, you've already in-sourced our technology. You're not going to continue to take it in source. You've already done that. the banks below the SIFIs, really, they've outsourced their technology because we can do it better, faster, cheaper for them and on their behalf. So when you think about AI, why would they want to take that technology, unbundle the whole thing and then drive it throughout the bet. That's not where banks are focused. Banks are really focused on how do I take this cost structure that's inside my bank that's been permanently here because of the regulatory cost of being a bank and how do I unlock that? It isn't about how do I -- because they already have really cost-effective, cost-efficient technology from us. So that's number one. The second thing I would say from an AI standpoint, and you've probably heard this from some of your companies here today, deploying AI capability is pretty expensive. So tokens are very expensive, and you have to take out human capital to provide that level of savings to offset it. So that, combined with the fact that these banks are regulated. And so one of the things they rely on us for is can you provide the regulatory oversight for all the technology products and services and so when they start to think about deploying AI, they're focused on tackling those costs. They're focused on really trying to use their existing partners because when they do, one, it will be cheaper for them because we obviously leverage the cost of technology across our platform and two, because we're able to provide a lot of reassurances around this agent is compliant with all the regulations, it can get through your privacy. It can get through your cyber just like that's what we do for the entire industry today. So for me, innovation, whether it's AI or digital innovation, it's going to go through this industry. The moat -- I used to think the moat for this industry was data, and we do have a lot of data. It's our bank's data. I actually think the moat for this industry is really around regulatory compliance and know-how because every single thing that we do has to meet some sort of regulatory requirement inside and on behalf of the banks. And that's where when they start to think about, well, will I buy code something. Sure, you could. The question is, how do you make it compliant? How do you get that 3-year regulator? Everything has to be auditable, evidenceable, has to get through every piece of regulatory regime or else you're risking the franchise. So I think that's some of the benefits we give the banks as we deploy.

And will -- another concern has been a noise, I would call it, more than concern has been around Visa's Pismo which is entering the core card issuing market in the U.S. So what differentiates your products from Pismo offering? And what is it you think the market is not understanding about how strong your competitive moat is right now?

Well, as you know, we bought the total issuing services company from Global Payments. We closed that transaction in January. It is the premier largest credit card processor in the world, serving the largest financial institutions at great scale. They've done it for 25 to 30 years. Many competitors have come into the space and tried to move the likes of Capital One JPMorgan, Barclays, et cetera, off of that platform. So the first thing I would say is what differentiates us is that it is a large-scale platform serving over 90% of the card issuing banks in the world today. And that's not to say that it could never be disruptive, but Pismo showing up in the market doesn't necessarily mean a whole lot to me other than there's another competitor there, and we can compete. I think the second thing I would say with respect to Pismo and Visa, I respect Visa as a competitor of course, but they bring to the table in terms of a product set that can be very relevant, I think, more down market as they bring debit and credit together -- if you are a very large credit card processor, you don't need debit and credit running on the same platform. In fact, you're never going to have debit and credit running on the same platform. That is going to be a more of a down market opportunity. The other thing I would say potentially around large issuers is why would they ever have be single-branded with Visa that would allow them to have no leverage over Mastercard. So I think there's a lot of things that you need to think about in terms of the positioning of the TSYS acquisition, which are the top 20 banks in the country. So now how to think about it inside the walls of FIS. So that was TSYS if you think about it stand-alone and how has it been very competitive and been the biggest and baddest processor and why we were very excited to bring it inside the FIS ecosystem. So now think about it. We serve every large financial institution. We have all kinds of products inside all of these banks take core, debit, credit, card processing, trading and asset services. lending, there's a lot of capabilities inside all of these customers. So adding credit card processing to the capability set was very complementary. And very important for FIS as we thought about trying to grow in payments. You can't grow in payments without a pretty big credit card processing capability. But we believe that we bring a lot of value when you think about how we bring FIS and the power of FIS to bear, probably how Visa things they bring Visa and the power of Visa to bear. I think the final thing I might say around it is, I think like 65% of our revenue is renewed through 2029. So I think we're in a pretty good spot. Then I asked that question before. Can you tell Yes.

Well, we wanted to focus on biggest issues hopefully. That's good we're there. And maybe go back to Anthropic because obviously, that partnership is very exciting. Can you talk a little bit about -- give us a little bit more insight on how this deal came together? Why did you pick them? What did you pick anthropic, why did they pick you as a partner, which I think says a lot about FIS. And we've seen some of your peers announce partnerships with AI companies. So what is really unique about this partnership?

Yes. So anthropic came to us. I think you've heard Dario say there's 3 industries where you need a lot of deep expertise to go along with frontier models, financial services, life sciences and actually health care because entering these markets in these industries, you need to have a deep set of expertise inside those industries. And so Anthropic has been spending time with the largest banks. And they had tackled a bunch of these or tried to tackle a bunch of these capabilities inside banks. What they said was a couple of things. One, if we go 1 by 1 by 1 into financial services, given the size and scale, it's going to take us a long time because we have to get through the bank's cyber, we have to get through the bank's legal, but they could probably do that, by the way. What they said to us is we could rebuild, for example, we could build a financial crimes agent on behalf and with a bank. What we can't tell the bank is, is it compliant. Does it meet all the regulatory needs of the bank. Does it have the cyber capability that it needs? Can it be auditable, can it get through a bank's internal audit, external audit, regulatory audit. And so they said to us, you really bring a couple of things. One, you bring that regulatory compliance know-how and horsepower and you have a pretty big professional services organization. So we think you would have the ability as we build these agents to go configure them and drop them in your banks. So I think for Anthropic, we were a really good and important partner because while they tried to think about going direct, I think they found very quickly that inside financial services to get to FIs and banks you do have to go through the existing providers. We do provide a huge amount of value. So that's how it came together, and we were really excited to partner with them. I think culturally, we do get on pretty well. They do understand the importance of banks and banks to the economy and to society. We're obviously a part of Mythos and the cyber all the cyber tools that they're rolling out. And I might make a comment on that in terms of one of the other benefits, an important part of those of us who serve this industry is providing that cyber capability to these financial institutions. The majority of our banks rely on us to be their technology and operations organizations for them. And so they need us to keep them safe, and we take that responsibility very seriously. So we invest a ton of money in cyber and are working with all of our banks and talk about resiliency and cyber with them quite a bit.

In terms of the Agentic stuff, you're starting with financial crimes -- and the question is, obviously, why and maybe broader deeper here is -- when do you expect the agent to be in the market? And is this going to be a new revenue stream? Are you replacing something? Maybe just give us a little bit more color.

Yes, it's a great question. we focused on financial crimes for a couple of reasons. One, the TAM is huge. And it is a TAM that really hasn't been addressed and in agenetic form. There's folks that have had technology that they've deployed, but being able to take out -- keeping the human in the loop by the way, but being able to take out a certain amount of cost is something that's TAM enhancing to both anthropic and to FIS. I think the special sauce around -- the agent is scheduled to come out at the end of this year. We have 2 bank partners that are in design with us. Bank of Montreal, which is a larger bank and then amalgamate a bank on the smaller end. The reason why this agent works for us and why we chose it besides the fact that there's a lot of cost there is performing a financial crime. So filing a suit activity report is no different than any bank. It's not proprietary. There's nothing special about it. They need to get it right every single time. It has to be right when they file it. But it's not anything special that any bank says, to you, Dan, hey, I'm a better bank because I can buy SARs more quickly. SARs are required to be filed within great compliance. And so that's 1 of the things as we think about where we think we make a lot of sense where we should deploy agents, it's places where we can add value to the bank by saving them time and money where there's a big TAM enhancing opportunity for us. and where the bank is using a combination of our systems and their systems, so an agent would make sense. So we focused there. There's a lot of places inside a bank that makes sense in those types of guidelines. So it will become available at the end of this year. And I actually asked the team, why does it take so long? And the answer is, it doesn't take a long time to build the agent. It takes a long time to test the model because this is not -- this has to be exact. When you build a model that creates a SAAR filing, every single time, it has to be the same. So the model building, which is a proprietary model that we're building with Anthropic that -- and the agent, by the way, is our IP has to be right every time. So that is what takes a long time of training of the agent. And that's where the forward deployed engineers and FIS are learning how to train those models. And that's why it's so valuable for us to be partnering with Anthropic.

That was actually going to be my next question, but I mean, maybe we can -- so the agent is owned by FIS. Not by Anthropic. Yes, I think you answered it. And can you talk a little bit about the pricing, the -- how do you envision the pricing model going forward in the revenue?

Yes. It's a great question, and we're spending a bit of time on it because we're thinking about just like all of our products and services, what value does this provide to the bank. So it should create a very good return on investment for the bank as they deploy the agent. They should be able to, in a very safe and secure manner reduced cost inside their regulatory and compliance group. The question then becomes, okay, how do we price that agent. So we are working with Anthropic to see how other folks are doing it. What I will say, the other benefit of partnering with us as a bank, when you think about the cost of the token. The cost of the tokens will get passed through. So there'll be the price of the agent plus the cost of the token. One of the reasons the smaller bank came to us to talk about why they wanted us to work with us to build an agent is they had started to go out and build agents on their own. The challenge is a couple of things. One is this going to create a bunch of tech debt in their bank who was going to keep the agent up, how are they going to run it? It was running on top of a bunch of systems. But more importantly, the cost of the agent and the tokens were cost prohibitive for a small bank. So when you buy an agent from us, we've negotiated a token cost at our rate at a very large consumption rate. And so you get the advantage of working with a negotiated token cost at the higher -- just like in any technology versus if you were going to do it yourself? So more to come.

And some of the questions we're getting from investors these days is, okay, well, there's a lot for FIS here in this partnership. But what is in it for Anthropic?

A couple of things. One, they have a partner that helps them distribute agents out into the financial services industry. safely and securely, one. So we act as a distribution channel to we act as a product development capability for them. And three, they get the token -- the benefit of the token consumption as we deliver it through our existing customer base. .

And is this an exclusive deal? Are you...

No, we aren't exclusive to them and they're not exclusive to us. And I think we both felt like that was really important because we needed to compete on the benefits and the merits of our -- what each of us do really well. and the world changes. But I think for right now, we are each other's probably biggest partner in the financial services space, and we're very committed, and we're very excited about what's happening here. .

A few more minutes. So let's maybe shift gears. Obviously, FIS has a tremendous amount of data within financial services, deeply integrated solution and long-term relationships. So how do you respond to the concern that you're letting someone like Anthropic see everything and potentially take advantage of it.

Yes. They don't see everything. So the model is proprietary to us. And we have protections around the agent itself. It is our IP. It is our agent. So they don't -- I mean they see everything under the same thing as a cloud provider sees everything. So it goes through their token system. So I don't -- that being said, we're always very protective of when we partner with someone, and making sure we have the right protections because we do bring a very healthy set of product capabilities and regulatory know-how, and we've made sure we've protected that. I think Anthropic is continuing, obviously, to create more and more value, and we would expect to see them want to distribute to banks as they build out their products. But what we've agreed with them is if we want to take their products going forward, we're happy to do that, and we can drive that into the distribution as well. As you know, the space is evolving quite well. But right now, I'm not worried about them. We've protected ourselves legally in terms of what we're bringing to the partnership. So I don't think there's a risk there.

Okay. And I got a couple of more questions and -- this is very insightful. Maybe take a step back and think about AI more broadly. What are the areas in functions from both cost and revenue opportunity where you're deploying AI with success these days?

So zooming out. So when you think about -- there's 2 places in the revenue side of things. So we just talked about one that we believe is enhancing with the Anthropic partnership, which is an agent that gets after a set of returns and costs for clients we've never been able to get before. Then there is AI we're enabling inside of our existing technology and platforms that may significantly increase the benefit of that capability, cash flow forecasting, for example, has been significantly improved as we've used generative AI to make it better. So in that scenario that makes our existing product better, we can charge a premium price or possibly it just becomes this is the cost of the new product in market. So that's how to think about revenue. It's both existing and driving new revenue growth in the existing market, but then there's TAM enhancing revenue. The other piece, I would say, that's getting pretty material inside banks. And in general, -- and this is why, again, from an FIS and an industry standpoint, critically important, no matter what you're doing with AI, you need real-time data. You need real-time data. It has to be sourced. It has to be secured. It has to be householded. It has to have privacy. It has to have lots of things around it. So we are -- and we launched this at Emerald as well, we have created a data platform. You remember, we have 74 billion card transactions, 900 million core accounts that go across our platform. if you're a bank, you consume those and you have multiple data feeds coming from us. We have created a data platform that puts all that data together in 1 platform, feeds at real time. That could be 1 piece of data source of data. The second can be we can build a model on top of it. You can build a model on top of it. So again, these are data TAM-enhancing products that we're putting on top of our existing set of platforms. Now moving internal to the firm. So talking about how are we deploying AI inside of FIS, which is I would say equally as exciting as what we're doing externally. There's 2 places we are materially focused. One is in our technology organization to drive AI-first capabilities and it's pretty consistent with what you've heard everywhere else. How do we drive developer productivity and how do we drive more product out of our existing engines. The second place is in our client and support organizations and those 2 organizations make up the majority of the company's employees. And in our client support organization, I have a Chief Client Officer and what she and I are most excited about is less about the expense and more about how can we fundamentally change client support. So today, remember, we're a B2B company, and we swivel chair. So if you have a problem with my technology you're using my core banking software, and you need help or an issue, you will swivel chair over to a ticketing system. This is a little bit old school here in B2B world, but this is how we live. We envision a way where with agentic AI, we could take those same customer support capabilities and put them inside the existing product and make the support actually differentiate it and take down the costs at the same time. And we can do that because of the agenetic and workflow capabilities that were built inside the swivel chair over here and put them over here. So we have a lot of things we are exploring in the client organization. So both places, technology and in client and support. And both in terms of, yes, we think there's cost opportunities, but also we think there's some real value that can accrue to the firm.

Two more questions on my end, very helpful. I think earlier this year, you mentioned that you more than quadrupling AI spend. So just to -- for housekeeping, does that include Anthropic or not? And then how are you prioritizing the future AI investments, i.e., are you spending enough? Or do you expect that spend to go up over time?

Yes, it's a great question. When we announced that, no, it did not include Anthropic spend, so more spend. The way we think about it is, yes, we would expect to spend more in AI, but no, we would not expect to spend more in total. And so really focusing on return on capital, return on investments, and making sure that we get more productive in terms of our capital allocation and taking away from the lower returns and putting to the higher returns. So obviously, the AI piece is going to develop a higher return. But I don't see it significantly increasing the dollar of capital we would have. I do think, to be fair though, I do think the cost of AI is rising. I think it's underestimated what the cost of AI is. And what I mean by that is, if you are a large technology company, you do have to make these investments and then you have to make a decision how are you going to pay for them. They are not free, they are not cheap. And so you have to make sure you get the right return out of them. I think you're hearing people start to talk about the cost of tokens, they're expensive. And so if you want to make sure you have a commitment to capital allocation and returns, then you are going to have to make sure that you're either getting more revenue or you're taking out costs. So I do think the cost of AI, especially in large-scale enterprises, we are potentially underestimating. And I think what you're seeing, though, to be fair, is I'm not meaning to say there's going to be exploding technology costs. What I mean is you're going to see people start to become very, very focused on where the return is on the cost. So are you going to turn Claude code on for your entire organization? Or do you want to make sure you're getting your return on it, and it's delivering the revenue and the cost savings that you watch? So I think it's important, especially from a technology standpoint because by default, everyone thinks the cost of technology is going down or becoming free. I also think the cost of cyber should not be lost on anyone. The cost of cyber for every single company, every single bank is going up. It's going to go up astronomically, and we're all going to have to manage it. And it is a big risk for the banking industry and is on of the benefits of being with a large provider like us because we spend a significant amount of money on cyber and the cost of cyber is escalating given what the bad guys now can do with AI just alongside us as the good guys. We take that very seriously. And it's another reason, by the way, why if you think about your big bank going to bring all your technology inside, you just need to triple it with respect to what the cyber costs are going to be to protect yourself.

And then last question on my end. In terms of build versus buy or achieving your AI goals, like where do you see the balance? And especially as you look to delever over the next year or so?

Well, I'm focused on delevering. So there is no buying. So we are very focused on delevering from now until the end of 2027 because we need to get our debt paid down related to the total issuing transaction. So there is no buying per se of M&A. I think -- what is interesting is aqua higher that can sometimes be categorized as M&A from an accounting standpoint. It's interesting I think in AI, though, specifically, it's such a new area. There's not a lot to buy right now, and I think it's easier to hire folks. . But look, the industry is new, and we'll see what evolves over time in terms of capabilities and products and solutions. And one of the things I love about Fintech is it's very dynamic. And our job is to make sure we have the best products and services available to our banks. And if we don't, that we have an open ecosystem, an API environment so they can pick best-of-breed and then if it gets too good in the future, we need to make sure we either partner with them or buy them. But at this point, we're very focused on debt paydown.

Great. Well, just about time. So thank you so much, Stephanie. It was a pleasure to have you here.

I appreciate that. Thank you.

Thanks, everyone.