Claims technology expectations are changing fast. AI and automation are increasingly standard, while configurability, integrations, and cloud architecture are playing a bigger role in vendor evaluations. Celent Senior Analyst Harry Huberty and Origami Risk’s Dustin Barrington take a market-driven look at what insurers and TPAs are prioritizing in today’s modernization evaluations. Drawing on market research and buyer conversations, they examine where AI is delivering value, where deterministic workflows remain essential, and how modern architecture can support smarter claims decisions. What you’ll learn: Which capabilities are becoming baseline requirements. Where AI and automation are delivering measurable value. Where RFI criteria and final selection decisions often diverge. How buyer priorities may evolve through 2027. How you can pressure-test your claims modernization criteria against current market trends. Hello, everyone. Welcome to our webinar on what PNC insurers are really asking for, claims systems priority shaping RFIs in 2026 and 2027. Today’s program is being recorded, and a link to the recording will be sent out after the program. Questions will be answered at the end of the presentation, and you can submit those questions through the q and a function on the Zoom toolbar. I’m gonna turn the call over to Dustin Barrington to get the conversation started. Thanks so much, Aubrey, and hello, everyone, and welcome. Thank you for joining us today for what PNC insurers are really asking for, claim system priorities shaping RFIs in 2026 and 2027. I’m Dustin Barrington with Origami Risk, and I’ll be hosting today’s conversation. Claims costs are up, severity is up, and litigation isn’t slowing down. Insurers and TPAs are responding by modernizing their claim systems, and the RFIs they write show us where the market is heading. Today, we’ll look at what buyers are prioritizing, where AI is delivering value, and where deterministic workflows still need to hold the line. I’m delighted to be joined today by Harry Huberty, senior analyst at Celent. Harry, welcome. Would you, tell us a little about yourself and the research that you do? Yeah. Absolutely. Thanks for having me on, Dustin. I’m delighted to join today. As Dustin mentioned, I’m Harry Huberty. I’m a senior analyst with Celent’s North America property casualty team. For anyone on the call who’s not familiar, Celent is a research and advisory firm, a small firm specializing in financial services, and I serve with the insurance vertical, so solely devoted to insurance, in the North American market. My job has me talk to a lot of carriers and a lot of solution providers, solution providers to know their capabilities and insurers to know their needs. So really happy to be joining you today and excited to be talking about claim systems. Wonderful. We’re, privileged to have you here with us. Let’s start by looking back at the early twenty twenties. Why did pandemic era conditions trigger such severe long tail losses across personal and commercial property lines, and what lasting structural changes did they force on claims organizations? Yeah. It… It’s a great question and a really useful way to frame the discussion around what insurers are looking for from modern claim systems. So in the pandemic, in addition to everybody’s, you know, change to remote work, differences in driving habits, You also had a lot of supply chain disruption, and that created for carriers an environment of very unpredictable claims costs. So if you think of the costs that would normally go along with something that you could, you know, predict fairly well, like a personal auto claim. Suddenly, parts were taking a lot longer to come. That means that you have to garage the car at the repair shop for longer. It also means that you might have, a longer tail on your rental car. All And that added up into a situation where carriers found it very difficult to actually predict what their lost costs were going to be. And that wasn’t any simpler on the commercial line side of the, equation either. Again, lots of supply chain disruption, very hard to predict those outcomes. And that created an environment where carriers really wanted to control what they could, and that led for them to seek new modern claim systems that could help them, with their operational efficiency and in so doing, help them control their loss adjustment expense. So kind of the recognition was, well, we can’t really predict what things are ultimately gonna cost, but we can control what’s in our house. And that means giving our claims adjusters new and better tools to make them more efficient and more effective. Right. And that really shows up, I think, also in claims, cycle time. And when we talk to claims leaders after the surge, one of the first things that they were asking was that real time visibility into file aging, diary compliance, and vendor performance so they can step up before the delay compounds those losses. Thank you. Yeah. You you really nailed it. It really does turn into a balance sheet impairment, when you can’t control those claim costs and when those cycles tend to run long. Right. So we’ve also seen climate patterns structurally shift. Severe convective storms and record hail events suggest catastrophe season effectively has no end. How are carriers pivoting from trying to control loss events to mastering the variabilities that they have to manage? Yeah. It’s a really tough, situation to be to be in. We’re kind of in an era where cat season doesn’t really end, and those outcomes can be very difficult to predict. I was talking to one insurer, a very small farm mutual in Illinois earlier this year who said that, 2026 had set records for the number of hail events they saw over the summer, and 2024, saw the largest ever hail stone, fall in the state of Illinois. It creates a situation for carriers where there really aren’t those quiet quarters where you used to be able to use it to clear your backlog. And it also… You know, to your point about the severity, it creates situations where what would have been perceived as a secondary peril is now producing losses that are in the primary peril territory. You’re seeing higher frequency. You’re seeing more, severe individual losses. One thing we see a lot of insurers really interested in is a sense of their geographic risk concentration. And that doesn’t always live in the claim system, but it can do. And, they’re trying to manage their exposure, in these unpredictable, catastrophe environments by saying, let’s just make sure that we’re not, you know, overexposed in a specific geographic region. It also creates interesting times like 2026 where we didn’t really have a very, a very disruptive hurricane season this year. So a lot of property casualty insurers are gonna go into the next year with large surpluses. And I think the important thing to keep in mind is just because it was a mild 2026 doesn’t mean it’s gonna be a mild 2027. So really need some planful approaches to managing that risk and exposure. Right. We can’t really control the weather. It’s it’s really how quickly and consistently we can respond. And, you know, maybe with the exception of what you’ve noticed, just this last hurricane season, there’s seemingly no quiet quarter to catch up on work. You’ve really gotta be prepared for whatever the weather’s going to send you. And, I guess that really sets us up for the next question about what the core platform itself has to do to support all of this. So as specialized Insurtechs proliferate for fraud, aerial imagery, and automated estimating, how does a modern core system avoid integration fatigue and act as a true command center? Yeah. It’s a great question. And I think the way that you raised it is is really provocative. In some ways, we’re at a really happy point in, the history of of claims adjustment and education. We’ve got so many sources of data, that we have access to. So like you mentioned, there’s fraud, there’s aerial imaging, there’s telematics data, and, it provides a lot of information. The challenge of the core system is to make sure that information is available and useful and actionable. So one thing we hear from our insurer clients is that their claims adjusters do get a little bit of app fatigue. You know, there’s one pane for the core system. There’s another window for their telematics data. There’s another window for their geospatial imaging. And, to the extent that it’s possible, being able to pull all of that data into the core system can really make the adjuster’s life a lot easier and, can make the workflow a lot more efficient. So I think there’s, you know… Like you said, at command center, there’s kind of a need for a single pane of glass kind of approach that, presents all of that information to the adjuster, and in a way that that makes it easier to take action on it. So not just necessarily presenting the data, but also highlighting why it’s important. Yeah. That makes a lot of sense. In fact, I was just in an event, sitting across from, a a client of Origami who said one of the big things that have really changed is she’s no longer jumping between all these systems because so much of that data is coming into, into Origami for us. So, I understand what you’re saying exactly. But would you say it’s enough to display third party data inside the claim, or does more have to happen with it to be really effective? Yeah. Well, it kinda gets into that point about, you know, it’s not just making this data available and presenting it. It’s also making it actionable. And it provides difficulty for core systems providers like Origami Risk. You need to do a lot of work to cultivate those APIs. You may… You know, I know that you’ve done a lot of work on your partner ecosystem to help clean up, you know, some of that need for going into a separate window or, eliminating the need for double entry of data. You kind of need this kind of bidirectional data flow. So you get the third party insights in and consumable, but also it’s being made available in a way that, the adjusters can be responsive and take action of it. It also creates some complexity for carriers around that total cost of ownership question. So, you know, the the more sources of data you have, each of those data calls is an additional expense for the carrier, and it’s a a question of making sure you understand which of those data calls are gonna be valuable, and which are gonna really contribute to that underwriter’s efficient… Or sorry. Apologies. Claim adjusters efficiency. I I live in the underwriting world so much I sometimes slip up. But it does create challenges, you know, for platform providers like Origami Risk, to be able to pull in all of those data from such a disparate number of sources and present it in one pane that can make it actionable for the adjuster. Yeah. Absolutely. And this is this is a place where Origami has put a lot of investment recently. We’ve rolled out a new marketplace with over a 180, different partners and connectors that help build powerful workflows with AI inserted, you know, for that exact need so that the third party result that you’re getting back can go beyond just the information that can trigger tasks and review reserves or process payments, all those sorts of things that make it actionable. So, definitely a big emphasis for us here. And that really brings us to the data underneath all of all of this and what the system has to prove about it. As document extraction and intelligent processing take on more manual work, what technical features must core systems provide to keep claims operations fully auditable? Yeah. That’s a great question. From a governance perspective, we see regulators, approaching insurers use of GenAI and other AI tools, with more interest. So I guess I’ll take a step back for a moment. You know, you mentioned things like document ingestion and information ingestion. And I’d hesitate to say that that’s a fully solved issue, but we see a lot of providers in the marketplace who do a very good job with it. And one of the things that creates for core system providers is maybe a little bit of feature creep. So for example, as something like data ingestion and document ingestion gets more commoditized, there’s more pressure on the core system provider to bring that functionality into the core platform. So it’s a one stop shop, and, it’s very easy for the insurer clients, to manage it. I think it’s creating the opportunity, for features that would be nice to have, in prior eras are becoming… It’s easy to add. We were just… You know, I was on a webinar earlier this morning where we were talking about the way in which AI code generation is becoming a a force multiplier for a lot of carriers, and it’s the same way for solution providers. You know, you’ve got those. It’s easier to add functions and capabilities than it would be, before. But to your point about governance, it it also comes with responsibility on the carrier’s part. And that follows both on the insurer and the solution providers. So I mentioned, you know, regulatory agencies are moving into oversight mode. For example, the NIAC just issued some guidance. It’s not binding guidance, but it creates a really good first place for insurers to look. And, that holds them… That holds the insurer ultimately responsible for whatever their AI does, whether it’s an in house application or a vended system. And so that means that core systems are going to need to serve as that record for accountability. They’ll be… Need to be able to create audit, auditable logs, which insurers and regulators can then use to understand why the AI made a decision that it did. Yeah. I mean, auditability and governance is such a huge issue right now. I mean, in a very risk centered sort of environment with a lot of accountability in terms of reporting, we have to be able to trace what happened, why, and even, you know, years in the future perhaps. So that visibility into automated actions are really important, for our claims adjusters. And you’ve talked about, protecting carriers from technical obsolescence. How does that work exactly? Yeah. I mean, it’s… One thing that we sometimes joke about, over on the Sellit side is that generative AI is this wonderful technology. It just never solves the problem the same way twice, and, periodically, it will lie to you. And so, really, what carriers are looking for are those kind of audit logs and activity logs where the, generative AI component will log down what it did and why it made this… The the decision that it did. And then that creates a trail that the carrier can go through to investigate how the AI was working or why it solved the problem in the way that it did. But it also creates that regulatory, auditability so that when the regulators come in and wanna know why an AI is doing what it’s doing, the carrier has a a record that they can point to. And the real key is that it just… You know, we we always say, like, it can’t be a black box. That was true when we were talking about AI as predictive analytics or AI as, machine learning, algorithms, and it’s the same way with generative AI. It just needs to provide that kind of accountability so that both the carrier and the regulator can see what’s going on and make any changes that are necessary if that’s something that has to be done. Yeah. That makes a a lot of sense. Maybe let’s pivot a little bit to, you know, what what we’re seeing on the adjuster side. So recent sentiment data shows widespread adjuster dissatisfaction with early generative AI rollouts. In your opinion, why have off the shelf LLMs fallen short, and how must technology providers reposition AI to earn back adjusters’ trust? Yeah. Absolutely. I… I’m pretty sure I must have seeded this question because this has been, like, my favorite statistic to cite since I I ran across it. But, Glassdoor did a survey where they, you know, surveyed a variety of different professions, a lot of different professional personas. And they found, as you mentioned, that claims adjusters were by far the least satisfied with the AI tools that they had. There are a couple things that came out in that study as reasons why claims adjusters particularly hated generative AI. First was that it was being a heavy handed mandate, so they were being required to use it by their leadership. But I think what’s more important for insurance generally is that they found that the tool was too broad and too generalized to be effective in assisting the claims adjuster. If you think about it, that makes sense. Claims adjusters are very, you know, specialized professionals. They have a lot of domain specific detailed knowledge. They look at a lot of claims, so they know what the key pieces of information are. And if you turn those, claim summaries over to a generalized tool that lacks that specialized knowledge, often the output is going to be something that’s beneath the level that would actually be helpful to the claims adjuster. So it’s gonna come off as a summary that’s superficial or maybe doesn’t catch, the key data that the claims adjuster needs to to pay attention to. So in some ways, it’s like being told that now, in addition to doing your jobs as claims adjuster, you’re also gonna have this, you know, little toddler that you’re babysitting. It’s gonna do a bad job of summarizing your claims files for you, and then you’re gonna have to correct it. And, by the way, your boss needs to see that you’re using this at least four hours a day or else there’s gonna be there’s gonna be trouble. I think that’s a big area where we’re gonna see innovation in 2027. I think there’s a growing awareness in the insurance space that general purpose models need to be more specify… Specialized. And that for claims adjusters and underwriters, they need tools that have really domain specific knowledge. That may come in the form of insurers training their own models, like travelers, for example, very happy with their own AI, but they built it in house. It may also come in the form of solution providers creating a domain specific context layer that sits between the AI and the core system and can provide some of that, additional knowledge, to be really impactful and really powerful. But the bottom line, I would say, is just insurance knowledge workers are really smart people. They do difficult jobs, and the amount of specialized information that they work with is pretty substantial, and that’s one of the reasons why a general purpose LLM might not be the best tool for them. Yeah. That makes a lot of sense. So we’ve heard from our our own clients that need to inject their own LLM models and, you know, ones that they feel comfortable with into the workflows that they’re creating in Origami. And so that’s that’s been a a big request request that we’ve been able to to meet the market with the demand. To your point, what do you think that those generic tools are missing? Like, what do they miss when they’re when they’re used in that generic capability? Yeah. I think a lot of it is that the claims adjuster knows the information to look for and the tool doesn’t. So the tool can do a good job of summarizing information that’s there, but what it produces is kind of the first step for what the claims adjuster would need to do to start the claim, rather than a really helpful and nuanced, summary. So for example, you know, a general purpose LLM might get an accident report, and they can say, you know, this car was struck by this car at this red light. And the claims adjuster can find that out themselves within a couple minutes of looking at the claim file. So how much value is that really adding? There’s another step where, the LLM… You know, again, I think this is future facing, but the LLM can start to think the way a claims adjuster does. And if it’s trained up with that domain specific knowledge, it can become a much more, nuanced summarizer and can really look for some of that information that the claims adjuster really needs to take action. Something like, is this a claim that’s going to generate a long tail? Is there a high threat of litigation? Is this something that I need to close quickly? General purpose tools, I think, aren’t there yet. But once they get there, that’ll be a real help for claims adjusters. Those are the questions that they need answered right away when they’re first opening up a claim file. Yeah. Absolutely. I think there’s… There really is a need to go beyond the generic and into those that are really specific so they can be truly helpful for the adjuster and really for everybody in the whole claims life cycle. With the rise of agentic AI workflows, where should insurers keep relying on deterministic business rules, and where is it safe to let probabilistic AI complete tasks on its own? I think every insurer is gonna have their own level of comfort with that. I I joke sometimes that insurers aren’t allowed to have any agentic automation until they finish their process mapping. You know, having just said that claims adjusters are highly knowledgeable, highly specialized professionals, I think they don’t tend to think of themselves as process custodians. It’s a little bit more like, you know, in art, like, composing a symphony. They’ve got all their information at hand. They’re working through the claim file. They’re noticing what’s important. They’re reaching out to the claimant. Maybe they’re creating subrogation demands. And so the process when you’re in the moment can feel relatively unstructured. But there is a structure to it. There is a workflow to it. And I think insurers, when they start to look at agentic automation would do well to very carefully map out their processes and understand what each step in the workflow is. And to your point, when you do that, you find out that some steps in the workflow are deterministic and ought to remain that way. You know, it’s a highly regulated industry. We’ve already talked about how, you know, NAIC is issuing guidance around AI. Many, many steps in in underwriting process or a claims adjustment process do need to be deterministic for regulatory reasons, if not business reasons as well. But what happens is, you know, we… We’re coming out of an era where all of that knowledge was not hard coded, but it was configured into the core system in the form of business rules. And where AI can assist is, you know, for example, if you’ve got your tolerance for what can be processed straight through as a claim, And then a claim comes in that’s a dollar more than what you said is that threshold. Is that something that our claims adjuster really needs to take a look at, or is it close enough to what we decided was our threshold that probably the AI could could handle that, that is, I think, where AI is gonna start coming in and really having an impact, and and really creating some efficiency gains is those… We didn’t quite anticipate it, which is the other drawback for business rules. I mean, they’re they’re great. And, again, some things do need to stay hard coded and predetermined, but business rules only get you as far as what you’ve seen before and what you’re able to anticipate. And, know, we’re starting to see things like generative AI assisted fraud evaluations. And those are very powerful because they can identify gaps, and scenarios that the insurer hasn’t seen before. And they may be able to to flag discrepancies in a claim file that’s, unlike fraud the insurer has seen in the past. I think those are very powerful applications, where AI can can do some real assistance. But it does require some kind of… Again, I say plan full process mapping. So you really understand, here’s a step that we’ve already done in our business rules. We’re not gonna throw that work away. That’s just gonna happen the way it’s always happened. And here’s a more probabilistic step where the AI can actually step in and create some lift. Right. So what I’m… I am curious we don’t necessarily need to take some sort of business rule and shoehorn AI in there just for the sake of AI. Like, there are some use cases that are really powerful. Maybe you could speak a little bit more, and I know you’ve touched on this already, but where does AI really earn its place? Where does it shine and, you know, should be used in ways that may… Maybe haven’t been considered before? Yeah. I think the ideal use of AI in a claims process is summarizing that file and directing the adjuster’s attention right where it needs to go, right where its highest value, as soon as it opens. So you eliminate that kind of loading step where the file’s open for ten or fifteen minutes, and you’re just reviewing data, to get your head around exactly what you’re being presented with. I think, again, we’re seeing that, you know, general purpose LLMs aren’t the best at this. But an LLM that was, you know, purpose trained on claims data, I think, would do a very good job of pointing the adjuster, right to the the pieces of information that are the most relevant, and, that they need to notice first. There’s also areas like severity estimation, you know, that I think LLMs and and generative AI can be very helpful with. And the difference between, you know, a claim that you settle quickly and efficiently and a claim that goes to litigation can be, you know… It it it can be a pretty significant delta between those two things. So the ability for, an AI algorithm to say, this is one that you need to keep an eye on. This is one that maybe needs a little bit of high touch handling. All of those things, I think, are places where AI can meaningfully contribute to, to claims adjustment. Yeah. It makes me think a little bit, and I’ve had this thought, for for for a while now is AI can do do do so much. I mean, there’s all these workflows and, business rules that could really be processed by AI. But still, at the core of it, I think all of us, including in the insurance world, we still want that human connection. We don’t wanna feel like we’re con… Constantly talking or working with machines. Like, it’s about people. So maybe pivoting just slightly, you know, how can carriers balance that digital straight through processing for high volume, low severity losses with high touch, empathetic handling where, severe trauma or or when, severe trauma or catastrophe strikes and needs more of that human touch? Yeah. No. It’s it’s a great question. You know, again, we talk about places where AI can add value and flagging. This claim is gonna need some human touch, could be a really significant spot for for that value to be added. I think carriers need to come to a little bit of a balance. There are, as you mentioned, certain claims that are fairly low stress. You know, there’s a windshield. You know, maybe there’s a small fender bender where nobody’s hurt. And I think what claimants want out of those scenarios are, efficient adjustment and a quick and satisfactory outcome. There are also scenarios where, you know, you’re filing a claim because your entire life has been upended. You know, there’s been a catastrophic storm. You don’t know where you’re going to be sleeping that night. Those are situations where insurers can really play this role of stability in their policyholders’ lives, and be the the the force for helping people get back on the right path. We did some research last year on, claims outcomes, kind of around payments. But we did find that this has a real, real lift for policyholders. In normal claims, 25% of, of the respondents to our survey said they were so satisfied with how their insurer had handled their claim that they would stay with their insurer even if their premiums were to go up. And in cat events, that actually went up to 40. So 40% of people, you know, said, I was so grateful with how my insurer handled this that I’ll stay with them, you know, even if my rates have to go up. And that’s really powerful. You really see the ways insurers can create stability and, continuity in people’s lives, you know, making them whole, just giving them the chance to get back on their feet. Yeah. Absolutely. I love that. That’s an incredible study. Let’s just pivot a little bit to, you know, the financial strategy behind this. So in our current economic climate with investment returns on the float able to make or break product, profitability, how are modern claims architectures supporting flexible, granular reserving, and proactive litigation budgeting? Yeah. It’s, it’s something that we see not a lot of systems handling all that well, to be totally frank. We just published a refresh of our claim system research this year, and I personally sat in on, I wanna say, 20 or 25 demos of, insurance claim systems. It’s surprising the number of systems out there where the reserving is just basically a notepad, or, you know, subrogation and litigation management are basically just places where you log notes. More sophisticated systems will allow you to manage it, you know, like, with calendar invites. They’ll pull in notes from emails. And, to your point about reserving, you know, we are in an inflationary environment. The interest rates are high, unlikely to stay so. And so carriers can really affect their own bottom line by reserving intelligently. And I think we’re on the cusp of claim systems being able to give insurers actionable insights when a claim comes in saying, you know, this is about how much you wanna reserve For the the actual damage itself. This is what you might wanna reserve for the servicing, And being able to break down that reserve with that level of granularity can really help carriers in their financial planning. And every dollar that you don’t need to reserve… You know, every dollar that you over reserve is a dollar that you can’t invest. And so making that judgment very accurately and very quickly is beginning to become a differentiator for a lot of insurers. Okay. Yeah. That makes perfect sense to me. So let’s kind of look back to 2023 and 2025. And kinda with that context, we’ve seen corp, claim system replacement spike in 2023 and again in 2025. What operational pressures drove that second wind, and how did ensure priorities change between the two waves? Yeah. So the early, the early window in 2023, was really driven by a lot of those pandemic related, cost spirals that we talked about. So these unpredictable cost outcomes and unpredictable claims, outcomes, leading insurers to look for modern claim systems that could help them adjust more efficiently. The 2025 wave, I think, is really more about the onset of, AI at an enterprise level at scale. I think those are more insurers who are, you know, in addition to seeking claims efficiency, are recognizing that tools are out there in the marketplace, becoming more available, becoming more fit for purpose all the time. And insurers are eager to augment their existing capabilities, by adding, you know, some of these AI, enabled, features, that are now available in the marketplace. So the last couple of years, especially, you know, we’ve we’ve seen AI take off like a rocket. It’s basically been adopted like no technology I’ve ever seen before. So in ten years of doing this, I think RPA is the only one that’s had kind of, like, the same adoption curve that AI has. And Sellit has been surveying carriers about their use of AI since 2023. You know, now more than half, almost three quarters of carriers have AI in house, and almost a quarter of carriers say they have AgenTik AI capabilities in deployment right now. So it’s very, very, it… It’s a really steep, accelerated adoption curve. And I think that carriers are recognizing that what’s available in the marketplace is a step change from what they’ve tended to think of a core system, being. That really what’s available now is quite a bit different and quite a bit more sophisticated, than, you know, what they would have thought of as a core policy admin or a core claim system, you know, even in the late twenty tens. Yeah. That’s interesting. Really going to, be interesting what unfolds over the next couple of years as AI continues to accelerate and adoption levels increase. It’s it’s, a great time to be in insurance, I think. So, Harry, we’ve covered a lot from pandemic error loss severity, perpetual catastrophes, two waves of modernization, reserving AI architecture, adjuster trust, auditability, ecosystems, and the balance between automation and empathy. If attendees take away, just one thing for their own RFI or modernization planning, what do you think that should be? Yeah. Just just one takeaway. I know. That’s a hard ask. I would say I would say all of this is… It’s all exciting in terms of what it might mean for the future of work and the future of claims adjustment, but there’s also a very practical rubber meets the road element. And the value doesn’t actually get created until the claims process actually becomes more efficient. I think insurers, as they innovate, would do well to focus on very objective, very achievable goals where the efficiency and gains can be measured. And starting from those applications can build outwards to more success. So I guess to sum it up, to use very many words where perhaps a very few would be would be better, this has to get operationalized. And what that means is identifying use cases and then putting it into action. There’s gonna be some testing and learning. There’s gonna be some moments where carriers aren’t quite sure that they want to operationalize AI, and that’s where starting small, creating a a proof of concept that’s measurable, can show the value that’s being generated. So that’s that’s what I would say is pick your use case, optimize it, measure that you’re… Measure your value, and expand from there. Okay. I I love that. I mean, for me, personally, I’m taking away this idea that, you know, AI has… Is a tremendous opportunity. But to really use it effectively, you need to put it into practice where it makes the most sense. You’ve got to be able to govern it. You have to know what it’s doing. You have to be able to audit it. And while you’re maintaining that sort of structure, you also need to have a platform that is flexible enough and configurable enough that wherever the industry is going to go, you can adapt and go with it so you’re not left behind on, you know, some sort of legacy platform or or trapped into, technology that was was great today but isn’t going to serve the future. So it’s that balance, I think, that you have to really strike. And, you know, it’s Yeah. That’s a we’re always, like, thinking about. It’s it’s genuinely a great point. And I mentioned earlier that we’re hearing from a lot of carriers that their approach to the build versus buy equation is changing a lot. I don’t really see a future where carriers are vibe coding out their own core policy admin or core core core claim systems. I think that build approach can make a lot of sense for point solutions, particularly where carriers have specific business needs that they understand well. But to your point, a really strong foundation, you know, a strong core system is going to provide the most effective bedrock for that kind of innovation. And I don’t think that, I don’t think carriers… Like I said, I don’t think think they’re gonna, like, vibe code their way to a billing system. It’s good to have that foundation in place. To your point, it’s good that you’ve… If you’ve got a modern core system, which can enable you to be flexible in the future, and then that gives you a platform to build on top of. So I think the core system is still relevant for insurers today. Wonderful, Harry. That’s great. If we could, let’s just pivot to a few, questions that have come in from the audience. And, of course, if you’ve got a question that hasn’t been asked already, feel free to put it into the the q and a, and we’ll get to it. The first one is, how are buyers testing AI explainability and auditability during vendor evaluations? Oh, wow. That’s a great question. I would say, we’ve already talked about, you know, the need for governance, the need for these auditability logs, the need for these, activity logs. What I see as being most effective when it comes time for actually evaluating the technology is for carriers to, create a demo road map or a demo workflow that’s very similar to the problems that they’re solving in the everyday. And, you know, this… A little fraught being on a on a webinar with a vendor. A lot of vendors we see want to show what they’re great at and don’t wanna show what they don’t do very well. And so that’s where it becomes really important for insurers to say, here is what our average day looks like. Please model that out. And that’s where I think the, explainability and accountability can come in. It’s walking the system through an ordinary workflow and then having the the vendor or solution provider show you the output, show you where the data got logged. That is, I think, what’s gonna be most effective for for insurers. It’s very easy to say… Not you guys, of course. But it’s very easy to say, we do it great. It’s gonna work every time, and it’s more important for carriers to actually see the way the system would perform in their day in, day out kinda work. And that’s where the accountability and auditability is going to going to come through. Excellent. And maybe you could speak to, this next question. How do you expect buyer priorities to evolve through 2027? Oh, that’s also a very thought provoking question. We… We’ve talked a little bit about how the build versus buy equation might be shifting for a lot of insurers. I think… And as I said, and I’ll stand by it, I don’t think core core systems are going away. You’re still gonna need the policy admin system to be a system of record even if you’re doing more more work in the underwriting workbench or you’re doing more pricing in a standalone rating engine. I think that folks who haven’t started on their own AI journeys are going to be highly interested in what a modern core platform can provide, and they’ll be very interested in the new features that have been added in the in the last couple of years. On the other hand, I do expect there to be a little less interest in some some of these point solutions. And I think carriers will build their own systems guided as they are by specific business needs. And every carrier is gonna have their own level of comfort with that. The ones that I’ve seen that are very bullish on, we’re gonna build it all ourselves, have mature IT functions with great leadership. And so they can afford to approach that with with some confidence. There are also a lot of insurers that we talk to who are relatively immature with their technology functions. We talk to insurers that are still on mainframe systems. And, I think those insurers can make their lives easier taking advantage of what the vended solution market has already solved, so they’re not, you know, repaving the cow path or recreating the wheel. Yeah. That makes a lot of sense. In your opinion, are there any claims capabilities that perhaps were differentiators at one point that have really now transitioned to baseline requirements in the RFIs that you see? Yeah. Yeah. I I… Like I said, you know, I come back to things like reserving and litigation management and subrogation. I was really surprised watching a lot of claims demos that many, many systems don’t handle those well and are relatively immature in what they have on offer. Subrogation always kinda drives me a little nuts just because it’s such an ability. You know, insurers don’t wanna pay for anything that they don’t have to pay for, and you go into these, you know, billion dollar organizations that are managing subrogation on Excel. I think that’s an area where claim systems are moving into, more standard features, but it’s it’s still a work in progress. Okay. Wonderful. And maybe just the last one is if you’re just starting a modernization effort now, what do you recommend as the first step? That’s a great question. For just starting a modernization effort now, what we recommend as the first step? You know, I mentioned, in talking about, automation that you really have to understand your processes and your own workflow. I think there’s a similar step to be taken where you take your own technology inventory and you really understand, what your stance is right now with the capabilities you have and chart that against the needs that you are going to have in the future. I think that kind of gap analysis makes the problem evident, and that allows you to, embark on a modernization effort that’s about, creating features and capabilities, not just doing technology for technology’s sake, if that makes sense. I think it all has to start with an underlying business need. And so understanding where your business need is and what technology you need to help you close that gap down, is gonna be the really crucial bit. Wonderful. Harry, I really appreciate the time to chat with you. I think you’ve given us a lot of insight on what the PNC insurers and the industry in general are are asking for, especially looking toward, ’26 and ’27 and beyond and being prepared for what’s coming. So thank you for being here, and thank you to our audience. I’ll turn it back to you, Aubrey. Okay. Yeah. Thank you. I thought that was a great conversation. So I wanna thank our audience for joining us today, and I’d like to extend a special thank you to our speakers, Dustin and Harry. As a reminder, we will send everyone a link to the recording once it’s available. So thank you all again for your participation, and have a wonderful day. Thanks, everyone. Wonderful. Thank you. Bye. Thank you.