Verisk Analytics, Inc. (VRSK) Earnings Call Transcript & Summary

Got it. So I think across these sort of hobby horses of ours, we've covered now that the investors should feel pretty good about the new pace of innovation, and we have not used the word AI yet in the [indiscernible] we've been doing this. So maybe that's a record for this event this year. But let's touch on AI. The number that I think resonated most with us coming out of the earnings call was that there are 7,000 licenses on your exact AI solution, which, sort of, is very much often contractor facing, not purely carrier-facing. But it's the widely -- it's the AI version of your widely used exact were claims estimation suite. Talk to us about what's resonating so well in exactly with exact AI in the marketplace. Is it is the strong data foundation, workflow capabilities? Or is it just the user base is particularly prime there a 7,000 is a big number.

Yes. So it's a combination of a couple of different things. And 1 is, yes, the value of the product and the desire for more efficiency in that space. So the exact AI tool can bring Gen AI tools on top of -- we talked about the software and the underlying pricing detail that enables the carriers and the contractors to develop this pricing estimate. Now you've got Gen AI tools built into it. So for example, one of the most popular features of this product is automated photo tagging. So you can take a photo of the damage that you're assessing. And then the AI will populate -- okay, here's the estimate. Here's the price list, and so here's what the estimate will cost. Now that, of course, requires review from the contractor, it may hypothesize that it's a certain type of drywall, certain tension of it and actually it needs to be modified. So there's carrier review. But that's one of the most popular features and it significantly improves the throughput of number of estimates that can be reviewed by an adjuster or by a contractor in a given day. So that's the popularity of it. That's one driver of the growth is the strength of the product. The second though is actually gets that something fundamental, which is the rate of adoption of AI products in different sectors. And so the rapid growth that you're seeing on the number of licenses is primarily on the contractor side today because they look at the product, they see something with a real return and a benefit for them. And the key thing is they are sort of able to adopt and switch on relatively quickly. In the carrier market, we also have users on that product. I think we said we have 1 of the top 10 using that. But what you see there is on the carrier side, it does take a bit more time for AI-enhanced products to go through a review process to go through an AI governance board to go through legal and contracting review for both sides to understand data use rights and what exactly is being shared and what content. So I think what we see there is the ability for adoption to be rapid when there is #1 of value proposition and that whatever kind of governance and contracting hurdles can be worked through quickly. We think it will also continue to deliver value over time on the carrier side, but that it is -- and there's quite a bit of interest. It is still going through, in many cases, the review process.

Yes. We see that as a few infra services firms that especially serve regulated use cases. The sales cycle with an AI-enabled product is, it's not short right now.

Yes.

But maybe to switch to -- obviously, the other thing we're seeing a lot of the information services space is the launch of MCP connectors. And various CaaS too. One is your forms rules and loss cost data. And the other, again, is that exact business, if I have it right. Obviously, insurance adopts technology in a deliberate way. Is that what you're seeing play out with your MCP connectors as well? Or has that been a little faster, has it been any way different than what you've seen in the exact or piece? .

Yes. I think on the carrier side, yes, it is on the similar cycles of that. There is a lot of interest and some adoption on the MCP connectors. Right now, today, those are MCP connectors on flood. It's not exclusive to Anthropic. So we can launch other connectors in the future with other large language model companies. But -- so that is -- a customer needs to be already a customer of Verisk as well as a customer of Anthropic in order to begin using the connectors.

Understood. And maybe just thinking through how you've distributed your data historically. Can you refresh us how to think about your historical distribution channels? And how does Verisk data distribution change in the age of AI?

Yes. I mean if you want to go back to historical distribution channels, the loss cost, the forms, rules and loss costs used to be mailed to our customers in a stack of paper circulars. That was how we started. Obviously, that's -- more recently, it was downloadable PDFs. There was still a surprising amount of workflow that -- of manual workflow that went -- took place on our customers' side between downloading a PDF, looking at the loss table and then potentially manually keying that into their own internal systems or their policy administration systems. Our goal is to distribute in however a customer wants to take it. And believe it or not, there are plenty of customers that still access our content by downloading PDFs. That will continue to change over time. For those who have the appetite for it, obviously, a better 2020's way is to provide the data through APIs. They could also get it through -- we have connectors built through most policy administration systems so that they can access it directly through a Guidewire and Duck Creek or many of the other policy administration systems that they may have in place. There are many that build their own connections. And then, of course, in today's world, connectors may be the latest way to access the data. Eventually, we are doing some trials where we are building agents based on our data, and that may be a way for our customers to interact with our content.

And I know that the core part of your core lines of your Reimagine program was making sure you are distributing the data as in the most modern way that your clients wanted so that they can get rid of any associated tech debt because the spending on ingesting your data has always sort of been more than the spending on your data itself. So as you've seen -- as we enter the age of AI, which is made sort of in some ways, reducing tech debt, a much easier task. Have you had to go back to the drawing board and say, "Hey, look, this is our road map is Corlanor imagined. But actually now, we're doing this thing instead or we're doing that thing instead. Had you guys had pretty good visibility that this was the road of travel, maybe towards that.

Yes. I think we have pretty good visibility. I don't think it's instead of what we're doing with core lines we imagine. I think it's in addition to and on top of. So maybe -- maybe if I can jump off from that. As you were talking about distribution of content, and we said downloading PDFs, part of core lines we imagine and reinvesting in those that forms, rules and loss cost content is -- has been about bringing the content together on a platform of core.verisk.com, where they can get all of it together and get associated insights and analytics around it. that all of our content will be available on core.verisk.com by the end of 2026, which is essentially that kind of essentially marks the completion of what was initially envisioned with core lines we imagine. I think on top of that is where we can begin to build other ways to access the data that they can use, whether it's API led or whether it is AI-enabled. And maybe if I can also comment on this -- our dialogue with our customers around the AI road map has been very active. They will have their own AI road maps, which is generally around workflow efficiency and how they manage their own workflow processes on the underwriting value chain on the claims kind of process workflow. Verisk has always been fairly workflow agnostic given that it's so different across all the carriers. And our value is that we bring our data and content to bear for the customer at their point of mission-critical questions. The mission-critical questions being, should I write this policy or not? What is the -- what are the risks associated with this policy? What is the actuarial expected loss cost of this policy? What am I taking on? How does this fit with the rest of my portfolio on the claims side, is there likely to be fraud associated with this claim or not? Those are each points and questions that they will come to us regardless of whether they have manual tasks that get them to that process or even in a genetic world, -- they will still need to answer those questions, and we largely believe to use our data and content to answer those questions. So that's why the AI dialogue with the customers is whether or not they are rebuilding their own workflow from an AI standpoint, the conversation they're having with us is how can that AI-based workflow interact with your content in an efficient way.

And we've heard a lot of talk lately about token costs, right? And we've heard examples of software businesses that had start changing their economics as they ratcheted up the amount of AI that was available to their clients. In your instance, walk us through sort of how both those branded products like Exact AI, but also the cloud-enabled products, how -- what does the tokenomics of those products look like?

Yes. Tokenomics on the different products will vary according to the pricing model of those products, which is going to vary as to what fits the market best. For the connectors, we do not bear the token cost the client does. I said the client had to have a anthropic subscription as well as the Anthropic subscription. So they handle that separately. That doesn't impact us for our cost base. On the Exact AI side, yes, the license -- there's a license price for that product that yes, should cover the token cost. And if that changes, then ultimately, the license costs might potentially have to change.

Got it. But I don't think we need to worry about the insurance industry, token maxing particularly.

Yes, that hasn't been their opportunities so far -- or their experience so far, I think nonetheless, whether they are at max or not, so to speak, and maxing. They are absolutely like every other industry looking to ensure that they have the best ROI for what they are spending on insurance, and we intend to be a partner in that discussion.

Understood. Let's maybe shift gears to capital allocation. versus this nice model of pretty fairly consistent 6% to 8% with margin expansion on top of that, driving good double-digit EPS growth. but you recently introduced a target to return 75-plus percent of free cash flow to shareholders. I recognize there's some fluidity around these discussions, but you also recently had a large acquisition that midsize acquisition, let's say, whose status has changed, as on and now was off and now it might be back on. Again, just walk investors through maybe what they should know about that AccuLynx situation and what that would -- what impacts that would or wouldn't have on your capital return framework?

Yes. So on the AccuLynx acquisition specifically, it was an acquisition that was signed just over a year ago in late July. For a number of reasons. It was terminated at the end of 2025. And there is now active litigation around that termination. So I won't give a lot of comment on that other than to say it's under appeal at the moment. And it is being handled on an expedited basis as litigation goes. So the time line, we do expect within -- on the order of months to have some visibility into outcomes there. As to our overall capital allocation, it continues to be -- first beginning with the generation of our very strong free cash flow and then looking to invest in our business organically and with acquisitions, probably at the moment, focus more on tuck-in size acquisitions. And then returning capital to shareholders in the form of dividend and share repurchases. Yes, with that commitment of over 75% of free cash flow on an ongoing basis, given the strength of our investment-grade balance sheet and our 2 to 3x debt-to-EBITDA range, which even today, we're at -- in the middle of that at around 2.5x debt-to-EBITDA. So we have significant balance sheet capacity to fund whatever tuck-in acquisitions we may be interested in.

Got it. But that would include -- if this were to go a certain way, that framework would not be materially disrupted by this acquisition?

So yes, so if the situation were to be that we were going to end up proceeding with the AccuLynx acquisition, number one, we would have the capacity to do that on an all-cash basis, probably in the immediate aftermath of that, that would reduce the 75% -- that would temporarily pause the 75% of free cash flow return to shareholders. But we would -- I think we would relatively quickly delever and be able to return to that position.

Got it. That's very helpful. One thing that's sort of, I would say, unique to Verisk are appreciated by the long time owners of the company in public markets has been year in Lee, your CEOs focus on return on invested capital. We like to look at some how these are structured and designed and measured by your Board. And about a year or 2 ago, they switched from measuring your performance on an incremental ROIC basis to an absolute ROIC basis. And I know it feels like it's in the weeds, but ROIC is very much in focus for you guys. So maybe you can touch a bit on that transition from that incremental measure to the absolute measure. The target for you guys has been impressive I think is the target number in the proxy. So maybe you can just play out for us what prompted that change? And -- and how you guys are thinking about ROIC at this juncture of the business?

Yes. ROIC is very important for us, holding ourselves accountable to what returns we're driving on capital is absolutely, we think things we should be held accountable to, whether it's organic investments or inorganic investments. And our Board has supported that focus. Investors have supported that focus. -- the change from incremental ROIC to just absolute ROIC was one that largely came actually in our discussions with shareholders. The feedback that we heard was we like the incremental ROIC concept, but it is difficult to measure. There is complexity to it, and it's hard to compare across peers as we looked at the peers and of course, with the Compensation Committee of our board and looking at metrics and stuff. There weren't many others out there that had an incremental ROIC target. And so for simplicity as well as the fact that we were now kind of post the divestitures that we've talked about, the feedback was for simplicity to have the ROIC metric, which also is maybe more intuitive both how you measure it and also kind of what those numbers mean, like what good looks like. And so that was the -- what it says about us is that we were just kind of open to that investor feedback in a more simple and a more simple metric to continue to drive accountability.

Got it. Got it. And obviously, the simplest way to hit that number is to grow organic revenues as fast as you can. The next way to hit that number is to continue your margin expansion, which obviously you did well on the last few years and have committed to sort of progressing forward on that. another piece of that would be, of course, are you deploying capital towards M&A effectively? Are you reducing the amount of invested capital through buybacks. But -- maybe on that M&A piece, again, talk about the inorganic opportunity in front of you. What do you look for in an acquisition? And in the past, you sort of said, "Hey, look, we feel like we're -- there's a lot of opportunities right now? Are there a lot of opportunities right now?

Yes. So within acquisitions, we said we've kind of got our priorities or roughly 3 priorities. One is proprietary data sets and anything that would enhance our own proprietary data -- sorry 0 -- I should have given 0, which is the filter of is this relevant? Is this acquisition relevant to our insurance businesses or not? -- we are solely focused on the insurance industry. Within the insurance industry, priority 1 is proprietary data. Priority 2 is expanding the total addressable market even within the insurance industry, whether that's geographic or maybe segments of the insurance industry that we hadn't addressed before or, I don't know, excess and surplus would be an example. And then priority 3 is anything that can enhance the workflow or the efficiency or the use of our data for customers. So those are sort of the priorities within the insurance industry. We do see opportunity out there, not as much that's attractive from a scaled and transformative business, but we will always continue to look and stay active in the market. It helps us stay sharp as to what's out there and what we should be keeping in mind for our own businesses.

Yes. And a perception that I've always had, and correct me if you think I'm wrong here, is that because you have these 98%, 99% retention rates, you're in 100 of the top insurance providers, all hundreds of top insurance providers in the U.S. for P&C, is that you really have the platform to scale a solution that might be too early stage for another company to acquire. And so as long as it has product market fit and is good the bones of the acquisition are good. You don't need it to be a material revenue generator be later in its maturity curve. You're willing to buy earlier stage companies. Is that a fair assessment? And is there any other thing we're missing in?

No, it absolutely is, and it is one of the drivers of success for us. We talk about, for example, the life insurance business that we acquired when we kind of entered into the life insurance space has been a software product that sells to life carriers, it was something in which we saw a great product and remain -- and continues to be a great product. Their growth had been kind of mid-single digits when we acquired it, and it has delivered strong double-digit growth every year thereafter, in part because of the customer footprint we are able to bring to the table, even in the life insurance space where we hadn't historically had a presence, the Verisk name carries weight and as well as just the scale and investment-grade balance sheet of a strong partner for the insurance industry, for the life insurance carriers. So yes, we drive growth on acquisitions that way. And I would add -- I think it is also an underrated strength in driving our strong returns on invested capital because even our organic investments -- if we get a product right, if we get an investment right, we can scale it very rapidly across our footprint. We don't -- Joe, we don't need to go take a lot of people out to lunch to get to know Verisk and know who we are. We can bring them innovations right off the bat, which has helped us scale our innovations.

That's really fantastic. We've got 30 seconds left. So maybe I'll end up with this question open ended. But Elizabeth, when we have you here in this room next year, what is one thing investors are not asking you about now or aren't talking about with respect to Verisk now that will be highly relevant to this room next year?

Great question. I think you are asking the question now, but I think a year from now, I'm really excited about what traction and further updates we can give on some of our AI product developments and starting to see the adoption more broadly across the industry.

Great. Thank you so much for your time today.

Thanks a bunch. Great to see you. Thanks for the attention.