Tom Warsop: Yeah. I don’t know that there’s a single vertical that drove it. I think the biggest growth because of the new sales that we’ve mentioned were the utilities and telecommunications segments. But we’ve seen good growth across the verticals that we support, but those would be the two largest.
Charles Nabhan: Got it. And if I could sneak one more in on capital allocation. You’re in a pretty good spot from a balance sheet standpoint. You had mentioned that your — the priority is buybacks. However, you do have that flexibility. So I was wondering if you could comment on M&A and if your appetite towards potential targets that would potentially accelerate your go-to-market in certain areas across your businesses? Anything you could say there would be helpful.
Scott Behrens: Well, I mean, I think we would always be opportunistic. But obviously, the — it’s been a number of years since we have done an acquisition. We did make the divestiture of our digital banking business in 2022. I would say just — I would look at 2024 at this point, very similar to ’23 and that the balanced approach of both share buybacks and de-levering is really the target capital allocation. I would just say, I think we’d always be opportunistic if there was something accretive, but the balance this year is targeted to share buyback and de-levering.
Charles Nabhan: Got it. Appreciate all the color.
Operator: Thank you. Our next question comes from the line of Joe Vafi with Canaccord Genuity. Please go ahead. Your line is open.
Pallav Saini: Good morning. This is Pallav Saini on for Joe. Thanks for taking our questions. First off, Tom, you touched on GenAI and large language models in your remarks. Can you give us some examples of how you’re using GenAI currently? And what are some near-term opportunities for you there? And I have a follow-up.
Tom Warsop: Sure. Three primary use cases for us with GenAI and LLMs. So first of all, it’s fraud detection and prevention, which I’ve mentioned before. I’m sure you’ve heard a lot about it. We’ve been using AI for over a decade in our fraud detection and prevention solutions. And we have patents and proprietary methods of creating algorithms and training models. And we have excellent products there and part of the reason is the use of AI and the continued use of AI. So that’s use case number one. Use case number two is customer service. And so as an example, we have — we haven’t completed this for all products, but for some of our products, we have loaded into Copilot. We use Microsoft’s Copilot for a secure environment. But we’ve loaded every piece of documentation we have for several of our products, our solutions.
And that includes FAQs, Wikis, inquiries from customers and the answers, all of it. And we’ve then trained the model, and we are now able to get very good productivity from our customer service representatives, the people that handle inquiries about our software products, people that deal with outages or issues that our customers are having. We’re able to get them productive in a fraction of the time that it used to take us with longer-term training. And ultimately, we’re going to make that same kind of knowledge base with the AI on top of it available directly to our customers, so that they can get answers to questions faster and it improves the productivity of our team. So that’s number two, customer service. Number three, probably predictably for a software company is software development.
And we have employed generative AI with our developers. And what we’ve been able to do — so I’ll just give you one example. We have several, but here is the one that I find the most interesting is we have created a way to extract logic from proven software. So some people call that legacy, I have banned the use of the word legacy inside of ACI. But our proven software, we’ve taken functionality out of it and created micro services in a matter of minutes. We created micro service in a matter of minutes. And we do that by using generative AI. And then we include a human in the loop, of course, because AI, you can’t just trust the output of a bot. And so we create these micro services and then we have our team tweak them and check validity. So we’re getting about 80% roughly accuracy in these micro services and then we’re taking that to 100% or as close as possible with our team members.
And so that — we’re getting probably — overall, we’re getting at least 30% productivity improvement by using AI. And in the case of these micro services, we’re seeing 10x or more, 10x, not 10%, more productivity from our team. So I don’t know that we’ll get that every time on every application, but it’s pretty exciting stuff, and it’s allowing us to move very rapidly. So those are the three primary use cases.
Pallav Saini: That’s great color. Thanks, Tom. And my follow-up is on digital assets. Are you providing any products and services on the crypto side right now? And how are you thinking about this space if you — and if you see any opportunities given some of the recent developments like the spot bitcoin ETFs? Thank you.
Scott Behrens: Yeah. Thanks. I was — just before I walked in here, I was watching the founder of one of those ETFs talking about the explosion that we’ve seen. But to answer your question, so many of our products are perfectly happy to facilitate a transaction in Bitcoin. It doesn’t really matter very much to us, what the medium is, whether it’s dollars or pounds or bitcoin or central bank digital currency, and it doesn’t really matter to our applications. And we have absolutely built in the ability to use crypto where it makes sense. So we’re fine with all of that. I don’t — I wish I was smart enough to tell you what all the impacts of the Bitcoin ETFs are. I might not be here right now. I might be on the beach somewhere.
