Great. Hello, everyone, and welcome to JP Morgan Boston TMC Conference. My name is Alexei Gogolev, Head of Vertical SaaS here at JP Morgan Research, and today I'm delighted to be hosting CCC management. We've got Githesh Ramamurthy. Githesh has been CEO of the company for many, many years, decades even. We also have Katie Coleman. She's vp of Finance at CCC. Welcome to both of you. Thank you for joining us.
Thank you.
Githesh, if I may start with some framing of the story. Coming out of 2025, you've crossed $1 billion revenue. AI is now at 10% of revenue. You got EvolutionIQ on board. What materially is now different about CCC's strategic posture going into 2026 compared to last year, and what changed that investors might not fully appreciate?
Sure. First of all, Alexei, thank you for having us here. I would say, you know, as we rolled through 2025, you saw that when we reported fourth quarter of 2025, is since we went public, we've been investing very heavily on our AI solutions. As you know, we've been doing AI for about 10 years. You know, what happened as we rolled from 2025 into 2026 is that the fundamental change has been that AI, all our AI products moved from evaluations, testing to really much more of a production where it's really starting to scale customer adoption. We have literally customers for every single product, including some very, very large customers. I would say that's been the big change.
It's one of those things where you work for years and years to what looks like overnight success. It is one of those things where I think initially, you know, we underestimated the time to build the trust for the AI. We feel now very good coming into 2026 that the trust in our products and the AI is really there.
Perfect. Githesh, given very high penetration with major insurance carriers and various repair facilities, where do you see the biggest growth opportunity? Would it be deeper wallet share in existing carriers, or maybe some adjacency expansions, or maybe some further repair monetization through OEM part ecosystem?
Sure. We see the largest opportunity for us continues to be to really identify the much broader insurance process. When we look across the entire ecosystem from the front end of the claim all the way through the back end, you know, the adoption that we're seeing, for example, the customers that have adopted our core solutions now as they adopt AI for auto physical damage, that's a significant uplift in terms of both the ROI the customers are seeing, and the revenue that we're seeing as a result of it. It's bringing along renewals on the core as well as this addition. Then we're also expanding to many adjacent parts of the process. For example, subrogation is now achieving scale.
That's starting to roll out. There are many other components for in the auto physical damage process itself. With AI, you know, there's a significant expansion of TAM with our existing customer base. One very large adjacency is the casualty business. There was actually a very interesting crossover that took place in 2025. For decades, the spend on auto physical damage, which is repairs and total losses, was higher than the medical claim spend. In 2025, that crossed over. Now the spend in casualty is actually more, slightly more than auto physical damage and is inflating at a higher rate. That's also an area where we've invested very heavily in casualty. We think that's a growth engine. Repair facilities continues to be a growth engine. Clearly the parts ecosystem, the work we're doing with OEMs.
We've also done a few things on the repair side, like consumer financing for repairs, 'cause 25% of all repairs are now paid for by consumers, in addition to the acquisition of EvolutionIQ, that gives us expansion into disability claims, workers' compensation claims. These are all the places where we see the next several years of growth.
Perfect. Githesh, you describe CCC as an interconnected decision engine. Maybe can you talk about decision points in the claim life cycle, where you can deliver the biggest step change in outcomes over the next couple of years? How will you measure the success? Is it going to be through cycle time or maybe reduction of leakage or some severity CX?
The answer to the second part of your question is yes. All of those points have significant impact. In terms of the first part of your question as to where we see where are decision engines and what are some critical decisions. For example, when the consumer, you know, is first reporting the claim. The decision on should this car be totaled or should it be repaired. That has substantial implications in terms of the consumer experience, in terms of the cost of the claim, and everything else. Today, with a single photo, our AI can help decide.
Yeah
can help not decide, but provide a recommendation. The decision is obviously in the hands of the adjuster, to decide what to do. That's an example of a decision point. When a car is in a repair facility, what diagnostics procedures should be used for this particular car on this particular day with this particular options, these particular cameras. That's another set of decisions. How do I repair this car? How do I disengage this 800 volt rail? That's another decision. At every step of the process, there are decisions. Same thing with casualty. We can also infer from photos, you know, whether, you know, what's the potential implication on medical claims for this car. Electronic part ordering.
That's also a whole series of decisions that the repairer might have to make or the insurer might have to make. There are decision points across every facet of it, and what really helps us are really three things to make this work. First is bringing contextual data for every decision. Second is having the workflows that deeply connect these pieces. The third is teeing the decisions up through AI or traditional deterministic means.
Githesh, as you sell more integrated solutions versus those point features, how do you decide what to bundle, and how do you avoid value dilution so customers clearly attribute outcomes to CCC and expand accordingly?
I think there's a couple of ways we do that. We just announced, you know, a large customer who just did that in the first quarter, where they felt that the value of choosing all of these solutions as an enterprise license was much more valuable because rather than evaluating individual solutions and pieces, they could look at the whole bundle. In many instances, when you use three or four solutions, the value is greater than rolling these out individually, and the time to ROI is also quicker. That's really more and more how we're migrating.
Mm-hmm. Yeah, makes sense. Githesh, in terms of AI deployment, without revisiting first some prior adoption stats, what does scale deployment mean operationally inside this Tier 1 carrier? What has to change in process or maybe staffing or governance, for AI to move from pilot to actual business as usual for these customers?
I would say, you know, many of our large sophisticated customers have been evaluating testing over the last couple of years. That process has led to how do we deploy, how do we train? We've been doing a lot of that work along the lines. In many instances, these tools are AI is integrated into our solutions. For example, I'll give you one example. For example, some of the re-inspection solutions we have. Customer might already be using the re-inspection solutions. Now there's an entire AI layer on top of it, which is using photos to generate incremental ROI for the customer. In terms of training, and some process changes, we usually uncover these during the pilot evaluation process.
That's why we're now starting to see the deployments, get to be much larger.
Mm-hmm. Okay. Githesh, in terms of governance at enterprise scale, so as automation expands, who owns model oversight, and how do various compliance teams engage, and what governance artifacts are increasingly required for broad deployment of your product?
I would say first and foremost, because we deal with a very regulated industry, so the consequences of getting these decisions wrong, Alexei, exactly as you point out, are very, very high. It actually starts with the cleansing of the data, the anonymization of the data. It starts all the way from how we prepare the data for training. There's, so there's enormous amount of work starting with the data that even goes into the training of the models, the traceability back to how these decisions were made. There's internal teams that are responsible for every facet of this, all the way from training and also during deployment. We also have to manage drift, the accuracy of the AI. That means you're getting daily feedback, making changes so that the accuracy is maintained.
What we're also seeing is that our customers are also getting increasingly sophisticated, to your point, which is, they all have teams in terms of AI governance. We have to pass all of the AI governance tests. In the past, it used to be just cybersecurity tests. Now it is AI governance. How is your data built? How are your models built? How are they maintained? What are the privacy issues? We engage very early on with all our customers to get through this before we get to a final contract and implementation.
Okay. When a carrier wants higher automation penetration, what becomes the binding constraint that you see most often? Is it data sufficiency or maybe some carrier rules that you talked about? What are you doing to remove some of those bottlenecks?
Yeah, I would say it is very dependent on customers.
Different customers based on the geography they operate in, their own books of business, their own philosophy on how they work with their clients. They have different views in terms of how they operate this. We have a high degree of configurability in our solutions. We've been doing this for now, you know, well over 10 years, in the last five years, much more intensely. We've started to understand the degrees of configurations that customers require. Different people might have different thresholds for approving something. you know, the human in the loop, how and when does something have to be looked at by a human being to be approved? There's a lot of different variations. It varies by customers. We've learned how to work those.
Okay. As you introduce AI agents across claims and repair workflows, which tasks are best suited for agentic automation today, and where is their industry at risk of deploying agents prematurely? Like maybe it's compliance that you refer to or some liability customer experiences.
Yeah. I think in general, I would say even when we started building out our AI, we've been very conservative, right? If you remember, when we started rolling these solutions out, we only said, "This can only do front-end hits.
Okay.
Can only do back-end hits." We've been very, very careful because it's better to be on the conservative side. By and large, that's really been our posture, is to be more conservative. Back to your question about where can agents be deployed, there are just tons of places where there are handoffs, multiple, you know, work that can be done asynchronously by an agent.
that comes back. The agent might fork four other agents to go get some of this information, come back to it, and then tee this up to an adjuster and say, "Here's the information I have for you. Can you make this decision if needed?" I would say, we are seeing the deployment of agents in repair facilities, in the claims function. Remember we had the IX Cloud capability.
Yeah. Yeah, yeah.
The IX Cloud is really turning into an orchestration engine across the enterprise, across the ecosystem. That's really what that is morphing into.
On the point of IX Cloud, I understand that it's not directly monetized, so how do you ensure it becomes this strategic anchor in customer conversations, and how do you decide which apps are on top and should be marketed first?
you know, a lot of that is done in very close consultation with our customers. We just finished our customer conference last week.
Yeah.
Right? You've been to some of them before. We have multiple advisory councils. We have advisory councils for subrogation for every area of the business, and we get very, very direct feedback as to what is creating value, what is important, how do you prioritize. That governs our development plans and our strategy, so we get very clear answers for that. Some things like the infrastructure, the IX Cloud, some of those pieces, that's just infrastructure that is an event framework that runs across the entire ecosystem.
We think, the right answer as we work with customers is looking at each of these applications that create specific ROI. Whereas for the platform itself, it's hard to figure out what that is, and we've chosen to make it easier for customers to adopt.
Okay, that makes sense. In terms of your durability of your moat in the frontier model world, as foundation models get cheaper and better, where does this differentiation concentrate? Is it going to be around the proprietary data that you have or maybe some network effects that we discussed? Which part is hardest for a well-funded entrant to replicate?
You know, the way we see it is we're using all of the frontier models. Early on, we only built our own models many, many years ago.
For the last several years, we're using every major frontier model.
Okay.
Not only are we running these frontier models, some of these models, smaller models, are actually running on the edge. We're even running some of the models on your phone.
Right? For those cannot be large, complex models. They have to be very lightweight, very tight, very specific, because the phones now have you know, some of the core capabilities.
Yeah.
We work across all of these different models, and we think our real advantage that our customers are telling us, at least feedback we're getting from customers, is the quality of the dataset we have, the way we've trained the dataset, therefore the accuracy of our models is very high. Second thing is that the embedded workflows that we have, where we are now connecting an insurance company to a repair facility. In the instance of a car, should I tow this car or repair this car? The connection to a tower where the tower is taking the picture and helping decide should this go to a repair so all of these different connection points in the network between a salvage yard, a tower, a repair facility, 7,000 parts provider, almost every OEM, where you're getting diagnostics procedures from 10 different partners.
All of these different decision points are all working in sync, and we think that is a huge advantage and therefore a huge source of ROI for our customers.
Okay. That makes sense. Githesh, moving to implementation velocity as what you call competitive weapon, what have you changed in tooling or maybe some playbooks and staffing to reduce time to production? Where can you still compress timelines meaningfully over the next few years?
We've learned a lot in terms of change management. I would say the biggest thing we've learned through the process is change management for AI is different than for traditional deterministic software.
Traditional deterministic software, you put it in, it's gonna do the same thing over and over and over again.
Yeah.
With probabilistic software, which is AI, you know, there are more gray areas. That means the thresholds have to be different. There's been an incredible amount of learning that we've had from a change management perspective, and we have taken this learning and applied that to the new rollouts that we're doing. We've also brought on board a number of talented people, who are now working, you know, in helping make these deployments easier. Plus, the customers have also gotten much more experienced at deploying these solutions.
Okay. You and I talked about casualty for some time now. For large casualty migrations, what are the highest risk operational gates, like regulatory configurations or maybe some user readiness, and what have you learned about sequencing these to avoid disruption?
I would say, for casualty, you know, in almost every instance, we have spent months understanding the customer. Months, in some cases might be even longer. In many instances, the customers have also been either piloting or testing the solutions. We understand all the nuances in terms of how they operate and technologically what our product's capable of. We take these two pieces, and then we sit down with the customer to envision in the new world with the capabilities we have, what should the process be, right? Today, with, you know, some of the core capabilities that came out of our work with the EvolutionIQ team, the ability to synthesize medical documents is an extraordinarily important piece. Today, an adjuster might spend hours trying to understand 200 pages that are in a medical claim.
That can now be teed up very much to the front using generative AI, but that also means some changes in the operations in the process. Again, so all of these things is understanding the existing capab- how they're using, what they're doing, what are the new capabilities are, and then stepping forward and saying, "With these capabilities, what would you do differently?
That deliberate process and that learning has started to accelerate the rollout.
Okay. Githesh, we spoke about payments, a lot in the past. You've described payments as a longer burn opportunity. What leading indicators like adoption or maybe some attach data unit economics would cause you to lean in more aggressively versus kind of keeping it more gradual?
You know, I would say, where we sit today, we think payments continues to be a very large opportunity.
Still a very, very small piece of the business for us.
Yeah
today. We did add one component, for example, consumer financing.
Yeah
for the repair facilities. That is actually starting to get traction. We're starting to see that where because 25% of claims are now paid for by the consumer.
You know, there is not a magic formula for us relative to payments. We do think it's a slower, longer term opportunity, and we have clearly prioritized the rolling out of AI for physical damage, casualty, subrogation, you know, workers' comp, you know, disability, all of these components. I don't know, Katie, if you wanted to add anything to that on the payments side.
No, I mean, it's definitely an opportunity that's out there. When you think about the estimate that we're producing, you know, that is really currency for the insurance economy. You know, facilitating then the actual payment is an opportunity that's out there. As Githesh said, there's a lot of other opportunities that we've been focused on, and we've seen really good success in rolling out and seeing adoption of AI solutions as well as prioritizing the integration of EvolutionIQ over the last year or so.
Okay. Thank you, Katie. Githesh, you just touched again on subrogation. As you move from smaller customers into the top 10 carriers, what changes in product requirements and integrations, and what milestones signal the categories moving from early adoption of subrogation to more mainstream?
I mean, bottom line is adoption. We are now starting to, in the early phases of that with a very large carrier. We will not name them publicly, but starting to use this at substantial scale.
You know, and the results are very good, both in terms of what the customer is experiencing, the ease with which we've been able to start to do the rollout. We are now, you know, we are a couple of years into this effort, and we are now starting to see that make progress. We'll talk more about that at the end of our Q2 call.
Okay.
I would also highlight we're not building a solution that's different for smaller carriers versus larger carriers.
We have one solution. We've talked about it being configurable for the rules or how a specific carrier runs their business, I think that's what really differentiates us. I think that's one aspect that I would just call out as well.
Okay. Thank you, Katie. With regards to repair workflow, after invoice reconciliation and related admin steps, what are the next time sync workflows you're targeting, and how do you balance investment between insurer ROI and then repair facility throughput?
First of all, we have dedicated teams. There are teams dedicated to the repair facility market. There are teams dedicated to insurance. We've made sure that each of our customer-facing markets have their own resources as well.
With the repair facilities, you know, we are in and out of repair facilities every day. There are opportunities around the estimation process. There are opportunities around the repair plan. How do we create a repair plan for this particular car? You know, what is the customer communications looking like? We're also working with making sure, for example, one of the things our repair customers have told us they don't like is having more total losses in their shop, right? We're working on how do we make sure that total losses don't come into the repair facilities, and there was more repairable vehicles. There's still. We're also applying this into electronic parts ordering, where agentic capabilities to say, "If this is the estimate, these are the parts that I need.
Can the AI do more of the work to do the electronic parts ordering, the follow-up? Did it show up? Do the parts arrive? There's just many, many aspects where we are agentifying the platform.
Okay. Githesh, in terms of the consumer mix, and consumer pay mix that you referred to, with self-pay growing and new financing embedded now in CCC ONE, as you referred to, what have you learned about how payment friction affects conversion and maybe cycle time, and how do you help shops win work without changing CCC risk posture?
I think part of it is a big part of it is the partnership we have.
Right? The partnership we have is with a financing company. It's very important, and this a partner that's been in consumer finance for a very long time, and they do it at substantial scale. Again, this is another example of, you know, in many, many places, we'll have partnerships to actually help solve many of these problems. In fact, at our customer conference, we brought several partners in the showcase, where in the tech showcase, we had several of our partners from consumer financing to diagnostics to other areas. You know, this is where specialized expertise and risk is particularly important.
Okay. Katie, maybe this is a question for you. With very strong free cash flow and active buyback program, what conditions would you or would lead to you to kind of pivot more incremental capital either towards more M&A or maybe towards more organic investments into growth or maybe faster deleveraging?
Yeah. We've been very disciplined in our capital allocation approach. We invest heavily in the business. You know, over $1 billion invested in R&D to develop a lot of the AI solutions that we've been talking about today over the last several years. That is always gonna be our number one priority. I would say on the M&A front, we've been very selective and will remain highly selective on M&A. I think where you see the business and the stock trading right now, it's a very attractive price point for us to be buying back shares, which we have been deploying most of our free cash flow towards over the last couple of years.
In the last six months or so, we've bought back about 10% of the outstanding stock. You know, in terms of where we're operating from a leverage perspective, we're about 2.7x levered. We feel like we're able to confidently stay underneath the 3x net leverage position and still maintain adequate capacity for investing in the business with an optionality with deploying additional free cash flow for buybacks or, as I said, being very selective on M&A.
Okay. In terms of margin algorithm, Katie, so as you invest in new solutions that can be somewhat margin dilutive, I guess, early in the process, how do you decide the right pace of launch versus scale? What internal guardrails let you innovate aggressively while still expanding margins over time?
We talk about delivering about 100 basis points on average of EBITDA margin expansion per year. We've been able to do that over the last several years. You know, we are investing in internal development, but the profile of those solutions, the margin profile of those solutions is not any different than our core solutions. It may look a little bit dilutive from a gross margin early on as we start to depreciate the capitalized software of that development. There's no structural difference in the newer solutions that are being developed versus the more core solutions that we have.
Okay. Githesh, maybe a final question for you about the EvolutionIQ acquisition. How do you avoid two product cultures? What integration choices matter most for you, maybe in terms of data model or workflow integration? What would be an early warning sign that integration is creating friction rather than leverage?
You know, it's an excellent question. In fact, this is one of the reasons we brought on board Josh Valdez as our Chief Product Officer. Josh has tremendous experience, you know, in terms of scale at which he's done product management, product strategy, product leadership, and Josh is actually deep in the middle from a product and strategy standpoint. We've now taken some of the EvolutionIQ team is actually, you know, leading some of the efforts on casualty activity. Some of the core capabilities are now integrated inside over here. There are certain things we do from a scale, security, and other standpoint that traditionally the CCC team has worked through. We're applying some of those capabilities over there.
Gradually, we've been blending more and more of the capabilities and in fact, when you click up to 1 higher level, if you look at the last, I would say 15, 18 months, we have promoted 35 leaders from within.
whether they're from, you know, New York or EvolutionIQ team or within the traditional CCC team. We've also brought on board 25 leaders from the outside. The mix So all of this is getting us ready as we think about the next five years, what are the skills, what are the capabilities that we need? It is really starting to work very, very nicely in terms of capabilities.
Githesh, in the last 30 seconds, anything that you want investors to walk away from this meeting? Anything that you feel is important to mention about CCC's future?
Yes. I would say, you know, the single most important thing is years of development in AI, in the customer relationships and the innovation is now truly starting to work at scale.
Yeah.
I think that's the fundamental point that we are at today, and gives us a lot of confidence about where we think the business will be in the next five years.
Perfect. Thank you very much, Githesh. Thank you, Katie, for being with us today.
Alexei, thank you so much.
Yep. Thank you.