Tomorrow’s Claims Today: Preparing for (Even More) AI Innovation

Welcome, everyone. The webcast is about to begin. Please take it away, Jeff. Thank you, Cooper, and hello, everyone. My name is Jeffrey Metz, and I’ll be your moderator for today’s webcast titled Tomorrow’s Claims Today, preparing for even more AI innovation. This event is brought to you by Property Casualty three sixty and sponsored by Origami Risk. Before we begin, let’s go over some basic housekeeping about the webcast console. This event is completely interactive and features many customizable functions. Every window you currently see from the speaker window to the slides to the q and a panel, either be enlarged or collapsed. So if you’d like to change the look and feel of your console, go right ahead. If you have a question for one of our speakers, please enter it in the q and a panel on your console for the live q and a at the end. We’ll ask as many questions as possible, so please ask away. And if we don’t get to your question, we do send these over to our speakers after the event so you may receive an email response afterwards. Lastly, there will be a couple of poll questions scattered throughout. So one one so when one comes up, please answer as best matches your experience. For today’s topic, we’ll be discussing the role of configurable core platforms and AI to supercharge claims operations, improve efficiency, and prepare your organization for the next wave of innovation. And today, to discuss this, we have our panelists. First, we have Chris Bennett, chief of strategy for Origami Risk and our webinar sponsor. Next, we have Heather Wilson, CEO of Clara Analytics. And finally, we have Scott Frosseff, a partner and expert service integrator with PwC with a wealth of experience in helping insurers adopt cutting edge technologies. Everyone, it’s great to have you today. Thank you. Thank you. Thank you. Okay. So let’s start by talking about the big picture. AI is reshaping the insurance industry, especially in p and c claims. But to harness its full potential, insurers need more than just an understanding of AI. They need the right foundation. So, Chris, from the perspective of a core SaaS platform provider, why is it so critical for insurers to have a configure figurable core platform in place to fully leverage AI? I I think there’s really a a two part answer to this. Right? And the first is understanding of the the core system’s place in the insurance ecosystem. As we think about the migration of technology, right, from paper files into actual, you know, organized claim systems that could really help facilitate the claims process, the role of the core system is to really serve as the central hub and to organize all of the activities throughout the life cycle of the claim. And and we’ve made, you know, huge technological advancements over the last thirty years in the area of claims, and and AI is only accelerating that. But the reality is the foundational life cycle of the claim has remained the state the same. You still have to start with intake and collect all of the data needed to, you know, properly work and adjudicate the claim, and it ends with settlement or closure. And along the way, you’ve gotta, you know, determine is there actual coverage, for this claim, you know, what type of reserves do I need to set, how do I route it to the right adjuster, you have to go through your investigation, ultimately, payment processes. But fundamentally, that hasn’t changed. Right? And the core system serves as the facilitator of all of the steps in the workflow. And and that, you know, as we think about AI, and it was really interesting last week to to attend the InsurTech conference out in Las Vegas and and walk the floor, the exhibit hall, and see the number of vendors in all sorts of area. And and Claire is a perfect example of this that are solving real world client problems in parts of that life cycle. Right? How do I make this part of that life cycle easier for the client? How do I get them better information, raise insights? And and just, you know, a a huge amount of innovation that’s happening. And the core platform sits at the center of all of that. Right? And and it’s the core platform that has to orchestrate, all of that connectivity. But but, really, it’s about taking all of those insights that are being raised by innovative vendors in the space, and then actually allowing clients to easily drive action based on those insights. And that is the true role of a of a core system, and it’s why flexibility, you know, configurability, you’ll hear no code talked about in in core platforms. The ability to rapidly adapt, is critically important, and and the core platform plays such an important role as the orchestrator of all of this this technical ecosystem. Thanks for that, Chris. And, Heather, as someone who’s been developing Insurance dot ai for years, what innovations are you seeing that insurers should be preparing for now? Sure. Thanks for the question. So it’s interesting. You say the word years. Yes. I’ve been at this for a couple of decades, and it’s it’s interesting watching how we’ve evolved from business intelligence to predictive analytics. And now we’ve added this amazing capability, generative AI to our arsenal. And these AI technologies, they’re not only driving operational efficiencies, but they’re giving the insurers the tools to lower avoidable losses in a material way. And that real magic happens when you combine all of these capabilities. So we have a lot of techniques that are happening on our AI platform, and you’re embedding them in the claims workflow that Chris was just talking about. That’s where the magic happens when the two come together, that that system of record and working through that workflow and then the system of intelligence working together. So if you if you think about generative AI can summarize the status of the claim, the predictive AI is telling you where it’s headed, and the business intelligence gives you the trending and the benchmarking, and it’s helping you understand how you’re performing. So you can really think about, we call it a lot augmented intelligence because we’re really taking the cognitive load off of the, adjuster, the claims handler, because we really want them using their amazing years of experience, in collaboration with the augmented intelligence that Clara brings. So you can think about AI now as giving you a virtual army of data analysts to tell you where you’ve been, where you are now, where you’re headed across your whole claims portfolio, and also on specific claims. One of the things that is very important, and I’ll talk about it again, is data quality still is really important to pay attention as we continue to adopt AI. And, it’s important because your AI models are trained on your data, and so we want you to get accurate predictions. And so the concept of garbage data in, garbage data out, we use the, concept of garbage data equals inaccurate predictions. So it’s still a big part of of of what you need to do for the right output. That’s fascinating. That’s fascinating. And, Scott, so you’ve worked with many insurers to implement both core systems and AI. What are some of the common challenges they face in this journey, and how do they overcome them? Yeah. Absolutely. Let me start big picture, then I’ll dive into some more tactical items. But I think the bigger thing that Carrie’s focus or ish have issues with is focusing on long term and short term goals. So how do you focus on quick wins, the pace of, you know, technology change changing very fast? Like, how do you focus on learning and incorporating those quick wins into your cohesive strategy or longer term North Star? I think with AI coming into play, this is new to a lot of people. They wanna know, one, where they’re taking us already with their goals and then then where they wanna bring AI to achieve those goals and why they’re doing this. So I think carriers struggle with updating their star to where they wanna go, then backing into the technology piece for where that technology architecture should be and design around that. And I think they get confused. I think there’s a lot of vendor overlap in market as well as proliferation. A lot of people say they can do what they do AI, and it’s up to them to evaluate how AI can fit right into their north star. Additionally, when I said short term goals, I think they struggle with how to experiment with the new technology to inform that strategy and have that continuous learning mind step to adapt to the north star and then and learn continuously. Getting more tactically, I think that, again, going back to the environment now is becoming more complex. Right? The core does itself, like Chris mentioned, but there’s a lot of other components around it. And this is hard to change because you need to integrate at pace, and flexibility is hard given legacy components, varying capabilities of of vendors and how they overlap and what’s the right one to use. And the more the two big ones, I think, for insurance carriers, one is data. And I think, you know, we’re looking at step four around AI and how to incorporate it, but step three and going back and and focusing on data and getting that right to set the the foundation to do this is hard. And that’s that’s where you struggle with integrations, getting the value out of these tools, and creating a holistic ecosystem for your claims organizations to work in. And then later’s people. So one, stakeholders don’t understand AI or what it the what it means. They don’t understand how to use it. And then they also change management so that your frontline workers. A lot of times, these are technology only projects. And how do you drop that in and understand how to evolve the skill sets, how they can use it, and then bring that product mindset into the AI experiments or into the core value selections or whatever you’re working on to make that more of value versus I get a lot of feedback that, hey. We’re running around with this one new tool. And GenAI is, like, the new one, but just overall, there’s a lot of data scientists running around not knowing the business and vice versa. So if you can bring them closer together, the ideas get better and the change gets easier, and then the trust in the eye raises, and then the adaption gets higher. It increases. So I guess, to overcome that, you know, there’s obviously partners like myself and those on the phone, but I think a lot of it is keeping that north star, keeping the cadences in perspective, focusing on your people, and knowing what we why you’re doing it. What are the objectives and how to do it? So thank you. Thank you, Scott. Thank you. And now that we’ve, kind of set the ground, just the basis for everything, let’s go ahead and get to our first poll. And so how prepared do you feel your organization is to fully leverage AI in your claims process? As I mentioned at the top, please just take a couple seconds to respond to this as best matches your experience or your perception, and we’ll go ahead and take a look at these results in real time. So, again, how prepared do you feel your organization is to fully leverage AI in your claims process? Very prepared, somewhat prepared, neutral, not very prepared, or not prepared at all. And we will take a couple seconds here to go ahead and just let you get those responses in, and we will then take a look at those results in real time. Okay. I’m getting a lot of responses in already. I’m gonna probably close this up in about, ten seconds. So if you have not had a chance to respond, please go ahead and do so. And, again, respond as best matches your experience. The question again is how prepared do you feel your organization is to fully leverage AI in your claims process? Very prepared, somewhat prepared, neutral, not very prepared, or not prepared at all. Okay. K. Let’s go ahead and take a look at these results now. And so it looks like plurality is neutral on this with next one is summer prepared, and next one after that, not very prepared. So right there in the middle. Is this about what everybody expected to see from this poll? I would think from my perspective, yeah, it sort of mirrors the technology adoption curve with AI. Right? And and I think insurance tends to fall squarely in the middle of the the the adoption curve for new technologies. We’re we’re an industry that mitigates risk for a living. Usually, you’re you don’t wanna be the very first to be trying out new technologies. You wanna make sure it gets in place and proven. So this this doesn’t really surprise me. It doesn’t me either, Chris. And I I think that, folks are, figuring out what is the right use case. I’ve got this great core system like like Origami. And so how do I start to think about that intelligence system and that analytical layer that I need to partner, with someone like Clara within Origami and then obviously have some consulting for PwC? But how do we start to think about the right use cases to jump into this AI playbook? And I think that’s something we’re gonna talk about today and where we’re seeing outcomes already for those who, jumped in, already. Yeah. I agree. I’m not surprised of the the neutrality of the answers. I think a lot of carriers are just dabbling or they’re taking a risk adverse position around employee policies on usage and other things and then waiting for the core systems like an origami or someone else to embed these in them versus try with them. And then also, given the regulation environment, they see a lot of carriers being tentative about going into this and waiting to see where things land, whether it’s NAIC, the Colorado laws, what’s coming up. And so this doesn’t surprise me at all. Okay. Great. Thank you. Thank you. And so now let’s get to our next topic, and that is AI integration. So let’s dive deeper into the integration of core platforms and AI. Heather, what are the specific AI capabilities that can be most effectively integrated into these platforms to enhance claims processes? Thanks. So we’re seeing a lot of carriers and self insureds jump in for a number of reasons. And, obviously, we keep regulation bias nonprescriptive. We wanna make sure that the adjuster, the claims handler is driving the case as they’re working in their core system as well as using a signal and the insights from the intelligence system like Clara. So some of the ways that we’re seeing predictive modeling for claims outcomes as well as using our machine learning or natural language processing or algorithms for raw detection are, you know, one, really looking at the very up upfront after FNOL, claim assignment. We’re seeing, carriers that, you know, making sure using the right algorithm of the right case coming in, the right experience being assigned to that. It’s really helping to get that right. We’ve had one, you know, one carrier that’s reduced their claim reassignment by eighty five percent. It’s a real productive productivity killer. And so many of our clients are using that model after first first notice of loss to just try to prevent that. I think the next area of just inefficiency that just carries a lot of manual and, just a lot of research time for claims, handlers are, know, they get asked a lot to prepare for claims reviews. It’s tedious. It’s, you know, the what, the the why, the how, and an adjuster has to drop everything when those questions come in. And this is where, you know, they’re spending hours upon hours reviewing claims documents, and this is where AI can and does and what we’re doing provide provides the claim summarization in a matter of minutes. They get the cliff notes of the claim. They get the documents associated with the claim, the summarization of the documents, whether it’s a there’s a legal document, a medical document, a police report, what have you. Again, you know, this is where Gen AI is taking an inefficiency, a very manual process, and giving them a lift and augmented, really, copilot here saying, listen. Here you go. At your fingertips, I’m giving you the data, giving you the analysis, I’m giving you the summarization so that you can get back to answering the questions and servicing the claimant that just had something happen. You know, we’ve been, looking at claims for over twelve years now. So our machine models, have millions and millions of claims in there. And so we’re trying to take the surprise out of of the the claim, the sleeper claim, the the stair stepping that’s happening. And claims are not usually linear. So, you know, our traditional diary systems are not working right now, especially those super high complex claims. You know, we’re triaging the case on a daily basis. If something happens, we’re asking, you know, them to go and look at this particular case. The explainability, here’s the various routes you can take just like a GPS ways. We want the, the driver of the case, the adjuster to stay in there, but look at what these insights are surfacing up. Here are the, options of your next best action. Here’s the explainability, and, here’s a way to help you, you know, twenty four seven that’s monitoring these cases. So it’s taking a diary system and just, you know, changing it on its head and saying, here’s your focus. And, you know, here’s ones that you don’t need to touch, you know, that may be really low touch, non severe cases. So, you know, we really use a GPS analogy saying in front of the case, what’s ahead, what could happen. But, also, when it comes time that they have to do research and answer questions, that’s also there for them in the in the matter of the moment. Thank you, Heather. Thank you for that. And, Chris, from your perspective, what are the key features that a core platform must have to support AI driven innovation in claims? So I I I think first and foremost, you know, the core system has to be able to easily consume all of the insights that are being raised. You know, Heather just just pointed out a number of different scenarios where, you know, Clara is doing really good work to raise insights, whether it’s leveraging generative AI to do a summary or leveraging machine learning to do scoring or reserve. You know, are are my reserves where they need to be? All of that information raises incredible insights. And the key to really being able to to use that information and to make it actionable is to be able to easily integrate it into your core system in a meaningful way. And and I think, you know, there are certainly some basic building blocks. So when you think about modern core systems and and most of the legacy vendors that out there have already been making, you know, updates to move this direction, certainly, you know, cloud based. Right? And and not just, you know, data hosted in the cloud, but but a true SaaS software as a service cloud native application where your vendors are supporting one version of their software instead of multiple versions. And you get the benefit of economies of scale. Right? So innovation that’s being driven that can easily deploy to an entire client base and doesn’t have to be, you know, especially configured client by client. That that’s certainly important. Obviously, the ability to easily consume this information via APIs. Right? And and it should not be a major project to integrate with a new idea, a new vendor, a new piece of of intelligence. Right? You you wanna make that process as easily pass as possible, which is where you get into kind of that no code configuration. The idea that I can configure with very little, you know, programming. I I don’t need a programmer. I just need somebody who understands basic structures of APIs. I can consume and integrate and get the data that’s coming in via these interchanges with companies like Clara into the right place in Origami where we can then start to drive action. I I think that’s another important thing to think about is how easily we can do that. And then certainly, the flexibility to actually modify the workflows in the system based on real time information and what we’re learning as we integrate the AI into the system. You know, if your process is rigid and it’s really hard to make workflow changes, then it’s gonna be hard to adapt as you learn new information because the pace of change is only gonna get faster. You’re gonna have more insights onto what’s happening, and you’re gonna wanna change and alter the way that you’re handling or routing claim to make sure you are getting it to the right person, right, to the right level of of authority into the person with the right experience to bring that human element to bear. You really want a system that can be easily, you know, configured to accommodate those new workflows. So, you know, can I continually drive innovation and change through the entire system and actually make it actionable? And that’s another key to think about with the core, system to to making this process in this ecosystem and and integrating with AI actually successful. Great. Thank you, Chris. Thank you. And, finally, Scott, what best practices are you seeing to truly harness both flexible core platforms and AI? And, really, really, what outcomes do you see coming out of these specific applications? Yeah. Absolutely. I think, Chris, you hit on some things there that I would say from a technical answer that are needed, right, to to make it successful, that flexible environment, API rich environment, quickly integrate and quickly adapt. Bringing it up kind of the back what I was talking about before, though, is success starts with really aligning the business outcomes. We have to know what you’re gonna do with both the core system AI, make sure technical and business alignment’s in place, and have cross functional collaboration. I think that’s one big thing that I see that has to be in place in order to successfully see. One, where you’re going with the transformation, whether that’s a core system, whether it’s AI or what you’re trying to do. And then, also, I think we’re starting to see an instance where if you think about the claims value chain, I think someone had on this earlier. We’re seeing carriers focus more on the vertical steps of the value chain and almost too much. And so the vendors are making each step very excellent. So whether that’s litigation management, FNOL, whatever step in it is. And so what you’re starting to see is the true end to end, like, the claim line owner starting to going away, and then you’re starting to see the excellence in the verticals. So the how do you start to think about who is the claim owner end to end and who really owns that? So the outcome to the customer and the outcome to the carrier. And this is also where I start to think of the traditional claim process needs to start to get to change. So I see a lot of either it’s AI or automation or suites. They’re just doing little p projects right now. But if you start to think about the bigger p that you can do and how to change the claims process itself, Maybe you’re moving to something more like a portfolio claim adjuster versus a line adjuster. So starting to think about the change management of these things. And I know I touched on change management lightly before, but it’s it is super critical. So, you know, how do you take a piece of technology? How do you have the ongoing support? How do you train them to know what they needed to do? How do you train your managers to manage whether it’s a core system or new AI tool? A new AI tool, how do you build trust with them? And then finally, you know, for for the Origami or or Clara. Right? Like, there is the vendor and strategic adviser support and partnerships that you need to leverage that, one, helps the implementation be smoother as well as you can influence road maps. You could influence what’s coming and influence, like, what capabilities you need in their systems, and so those are some of the best practices I’ve seen. As far as outcomes, you know, I think a lot of AI investments, the ROI and what you mean by ROI needs to be changed. I think a lot of people focus on efficiency only, and you’re seeing a lot of carriers, particularly with avenue of GenAI rising, not getting the outcomes they want. But then also forgetting about, hey. We don’t need to learn. We need to learn not only how to use this from a business standpoint, but evolve our business our technical folks how to integrate this into things and use things. So those are probably a few of the things I’d add there. Thank you. Great, Scott. Great. Thank you. And so now let’s go on to our next poll. And, so what AI applications are you currently using or considering in your claims process? And note, this is a multiple selection, so select all that apply as best as, match your experience, fraud detection, predictive analytics, intelligent document processing, raise insights to drive workflow or customer interactions, I. Chatbots, or none yet, planning to. If you could just take a few seconds there to go ahead and click on all the potential options or applications you’re looking to use in considering your claims process, and then we’ll take a look at those results in real time. We are getting some results in already. Again, what AI applications are you currently using or considering in your claims process? Fraud detection, predictive analytics, intelligent document processing, raise insights to drive workflow, customer interaction, I. Chatbots, or none yet of planning two. We’ll give that I’ll give it another about fifteen seconds, and then we’ll go ahead and push those results live. This doesn’t surprise me, the trending that’s coming here. This is exactly how the use cases are coming, to us. So we’re seeing the first bite of the apple coming in through intelligent document processing. Just the ability to have GenAI models, read these documents in a matter of minutes, it’s such a massive productivity savings. Fraud detection, the same way. We’re seeing, you know, the alerting, but it’s the relationships across, you know, lawyers, medical providers to be able to surface that up to SIU. And then all the transcription, the unstructured data, and the customer interaction, all of this is right on target. So I would say how we are approached right now. Yeah. I I would agree. The the the reading through of the documents, I mean, we just know how much time and effort. It is such a big time saver, so not surprising that that is really where people are are focusing. And, you know, and certainly, you know, predictive, is a little more of a mature AI technology. But, also, you know, you really have to understand what you’re gonna do with it. To Scott’s point earlier, you’ve gotta have a real plan. Like, this is how we’re, you know, this is how we’re gonna actually drive action leveraging predictive. Yeah. It doesn’t surprise me that the biggest time savings is is where people are really focusing. And then also, obviously, customer service is paramount. Right? And that customer experience, and so seeing customer interaction and chatbots, know, that’s an easy way to to get them quicker service, make them feel, you know, that attention without taking the burden of of an individual until it’s actually the right time for an individual to spend time with them. And what I’d add is that that raise insights and drive workflow is pretty broad. Right? So there’s a lot of use cases within that you could have, and I think you touched someone touched on, like, trial transcript summarization, claim summaries, demand packet summarization, medical management summarization, like individual summarization of files as you attach them in there. Those are all use cases that we’ve seen being explored at least as Gen AI has been more popularized for us. Okay. Great. Great. And so let’s go ahead and take a look at the next topic. And see here. It’s not responding. Give me one second. Okay. So now let’s bring this to life with some real world examples here. So, Scott, how are you seeing insurers find success in claims by leveraging flexible core platforms and AI? I think right now, a lot of insurers I see are struggling, actually. So I think what’s happening is there’s a focus on Generate AI coming in. It’s a shiny new tool, and they’re not focused back on that bigger picture where they wanna go, whether that’s the automation seat with gen the GenAI supplementing that. It’s called hyperautomation in some other places. So I see there’s, you know, a lot of vendors saying they could do this, and then they’re not harnessing a strategy how to do that as well as they’re not centralizing the focus on AI. I have a PwC, we have a factory we’ll talk about with AI GenAI, where we centralize our use cases and bring the product model owners in so we can reuse that use case and that pattern over and over again, whether it’s a summarization pattern, a generation pattern, a decision pattern. So I think they’re struggling really how to take AI, especially with its its gen AI has made it more popular, put governance, put strategy around it, and then experiment with it to see the ROI and value to it. I think a couple areas, I think, was mentioned where I see them focusing, though, where the the most popular ones in claims we’re seeing right now is fraud. I would call that for introducing the accurate ID, which would be through the quality of the referrals or detecting more. And I’m seeing more use cases involved in both litigation management, subrole, early settlements, reserve, and adjustments there. So, not probably the but, like, a direct answer to that question, but I think people are struggling, and that’s evidenced by we’re seeing lower number of use cases in production right now. Okay. Great. Great. So, Heather, how about, from the AI side? Do you have an example where AI made a significant impact on claims efficiency or accuracy and the associated outcomes? Not just one. I have several. So, thanks for giving me the next thirty minutes. I’m just kidding. So we you know, we’re we’re one platform, one integration, one UI with with Origami. And so, you get a lot of capabilities inside of that one platform. And that means if you’re having some, litigation management, litigation strategy, then, you know, using a capability to, understand, especially the unstructured data of those adjuster notes where there may be propensity that may be an alert about something that’s going on with the claimant and maybe use a settlement model to help to not even go into a litigious environment. We know there’s a lot of litigation financing going on right now. If it goes into litigation, then start using some of the models to help you choose the right defense counsel against some of those plaintiff law firms. So, again, you know, using all of the data, using the predictive, and then using next best action of the route based on, you know, the fogginess of what’s happening in some of the litigation environment. So that’s, you know, one area. I think the second we’ve talked a lot about, which is document intelligence. If you’ve ever seen the, before model of these adjusters printing out the the pages and pages and highlighting, You know, there’s a lot of problematic legal language in there. There’s a lot of medical severity language in there, and a model can get right to that bad faith or term limits or the demand and really put that right in center right in front of them and make sure that the rest of the claims organization knows some of the those demands that that have come in and it surfaced up even to, you know, a dashboard across the entire portfolio. I think the the other area that we talked about is the customer interaction. So those adjuster notes, that that voice, that text, you know, that data when you work with someone like Clara, we have unstructured unstructured data utilities. We know how to deal with any sort of data and then put that back to work in the models for any type of propensity behavior, propensity. It only enriches the the models. And now that we’re able to get at all of that data, it’s really helping to predict even more of what’s happening on that case. Scott talked about ROI. We lead with ROI, whether it’s around, you know, bringing down litigation costs, medical costs, you know, or on the expense side, bringing the claim duration cost in or the claims days in, you know, looking at the efficiency around the the the models reading the the documents. You know, five x to ten x is something that we deliver in the first year, and we lead with ROI because we’re delivering on outcomes based on some of these burning platforms, that are happening out there. And so, working, you know, and and having an easily integration with with Origami, one integration, you get all the capabilities on one platform. You turn on what you want. There’s a lot of use cases and scenarios right now where folks are getting a lot of good outcomes. Great. Thank you, Heather. Thank you for that. So from the core platform perspective, Chris, how is Origami, seeing claims experiences, orchestrated perhaps by integrating AI or other InsurTechs? Yeah. I would say I would take themes and combine things that both Scott and Heather said. So Heather Heather is right. There are some fantastic use cases that we have built, right, that that you can pull in to leverage AI. But to Scott’s point, it’s also slow to adopt across the broader client basis. Right? And and and it becomes sort of overwhelming. And so focusing on, you know, individual smaller use cases where a lot of clients have started. So, really, the simple things like summarization. Right? Can I summarize, and can I take unstructured data and pull structured data from it in an easy manner? Right? So can I can I understand what’s going on without having to spend all of that time? But then secondly, can I actually pull data from that that I can then start to take and put into my analytics engine internally? Right? So, hey. Am I seeing other claims that look like this in my overall experience? Does this client have particular issues that are being raised or surfaced? So, certainly, we’re seeing, you know, quite a bit of that. Predictive is obviously, one that that’s been around longer than the Gen AI. Right? So the machine learning is is a more mature AI technology now, and we’ve got a lot of clients that have put in place really good workflows leveraging leveraging Clara to to score and ultimately route those claims. Right? Getting them to the right level of experience in their adjuster because we hear from our clients all the time that that they’re doing more with less. Right? We know that that they’ve gotta be more efficient as a claims organization. They’ve got more volume and less people, and it’s harder to bring in new adjusters, and attract them into the industry. So they’ve gotta continue to focus on efficiency. And and where those tools have helped them do that, we’ve certainly seen success there. The other thing that I would like to point out, and this is this is, for me, sort of interesting, and and we’ve got some clients that are a little more innovative and thinking about it a little differently. You know, most of what we talk about in the claims process, specifically for insurers and leveraging AI, really focuses on reducing severity. Right? It’s how can I how can I handle a claim more efficiently, ensure that we are, you know, triaging that claim correctly, getting into the right person’s hand? Because, ultimately, if we can control the time that we’re spending on a claim and the overall time, timing equals money, we can reduce severity just by more efficiently handling claims and getting them and making the right decisions earlier in the process. What’s interesting to me about the potential of AI and what some of our clients are really, you know, looking to do, it is taking that structured data that GenAI allows them to pull out of, you know, all of those documents, even potentially images and video that that can then drive action and can focus not just on severity, but how do we actually reduce frequency? Like, how can I get this to the loss control staff? How can we work with the insured and with their loss control team to make sure that we clearly have an issue? We’ve identified something. AI has raised an issue. We’ve run the analytics internally and said, yes. I can see the trending. This is something we need to be thinking about. But then to actually take a look at how we start to to automatically create, you know, loss control surveys that can go out, and and recommendations and actions that those clients can act on to ultimately help reduce frequency. Right? If we can eliminate a claim, that’s that’s that’s the best way to control cost. So to me, that’s where the excitement really opens up because, you know, as long as claims have been around, there are two levers you can pull to control cost, right, of claims, and that’s frequency and severity. And if you can leverage AI to focus on both, I think that’s really impactful. That’s great. That’s great. So we do have another poll, before we get to our next section. What is the biggest challenge your organization faces in adopting AI for claims? This is single selection. The responses are potential responses are integration with existing systems, data quality and availability, cost of implementation, lack of skilled personnel, resistance to change, security and privacy issues, unsure where to start. So just go ahead and, take a look at those responses and just collect the one that best matches your organizational experience. What’s the biggest challenge, your organization faces in adopting AI for claims? And we will go ahead and push those results live once we get a good number of responses. Again, the selection possibilities are cost of implementation, lack of skilled personnel, resistance change, security and privacy issues, unsure where to start, and, of course, integration with existing systems. And what’s the biggest challenge your organization faces in adopting AI for claims? We’re gonna close this out in about fifteen seconds, but please go ahead and, just get your responses in, and we will take a look at those results. K. This is great feedback for where we’re taking the next topic, Chris. I gonna push these results live now. So it looks like, there was a plurality, though it’s pretty spread across integration with existing six systems. And then, look at all that that’s right there at the same. What do you guys think? Again, we we talked about it earlier that the ability to integrate, is really foundational to being able to leverage and use. Yes. And all of the above. I I I I feel your pain too, and we’ll we’ll talk about that as well. Great. Good comment from the audience. Definitely. Definitely. Okay. So why don’t we go ahead and get to our next topic? And so actionable AI. As we end the near or or near the end of our discussion, let’s talk about actionable steps. Heather, what should insurers focus on in terms of AI capabilities to ensure they’re ready for the next wave of innovation? It’s a great question, and I’m glad, Chris and I are here to talk about this because I think the first thing that you have to start thinking about in your mind is, okay. I’ve I’ve gotta get my core system of record that’s doing, you know, all of my workflow of operations. You’ve gotta get it ready for the systems of intelligence. And that’s the work that, honestly, Origami and Clara have done together. We have a a technical integration that we work Clara inside of Origami. So we’ve done, that, work for you. A modern core system like Origami is gonna give you the the advantage to take a you know, to jump into AI, without really a significant technical lift or investment. So that’s the first thing is, you know, recognizing that, you’re working with Origami. Origami has a trusted partner that has a system of intelligence, and we’re already ingraining together. And we can already start that conversation of what to, turn on. It’s very important to to know that, you know, in that core system, you’re gonna have a, intelligence layer like like a Clara that’s gonna come in and out of a core system. So when there’s actions that need to be taken, there’s insights. It’s you know, whether it’s an event driven, if it’s an episode, if it’s an outlier, that’s what that intelligence layer is doing inside of that workflow. I think the second thing we talked a little bit about this and having lived this during my time when I was, working in a carrier is that, you know, you do have partners like Clara that can help you with the data quality. You know, we do data quality assessments. We really help you to get that full feedback loop. That’s something that we do when we work with, an origami for an onboarding as well as, as others if some of your data is also, an internal in your lake house. It’s really important that we get that accurate for you. So having that focus is is just important. I think the other thing that I learned, especially when I left being in a carrier and was building a lot of things, people talk about build and buy, and I really talk about build and partner. And so we live in a world now that there’s a reason why organizations like Clara exist, and that’s just to really help you accelerate capabilities. You know, we’re a strong partner in AI with ROI realization. And so, you know, whether it’s, you know, giving you the platform that allows you to have a hybrid approach of us and you and then inside of Origami, There’s a lot of flexibility out there now, but, partnering is something is really important in having that trusted partner to, to accelerate. And I think the last thing that we always work on is getting those champions inside your organization. Those that just know whether it’s, you know, starting with document intelligence. Wow. I’ve got that. Whether it’s those going, I need to focus on litigation, whether it’s working on the severity modeling. You know, wherever you start in this AI playbook with an AI platform like Clara working alongside Origami, just making sure you have those, those adopters that wanna lead, this through. Great. Thank you, Heather. Thank you. And now, Scott, what’s your advice for insurers starting their journey towards AI integration? How can they ensure success? Yes. At first, they start small, but focus on scalable and where you wanna go. So what I mean by that is where you wanna go is address your AI strategy, address your governance, your third party risk, your compliance. And then but at the same time, you need to start to build this through innovation. How do you rethink innovation? How do you build sort of COI, or I called it a factory before? How do you start to embed AI into the overall organizational design and change of it as well? So how you develop, deploy, things like that. So I’d say start with smaller pilot programs, start small, test new technologies. And when I say technologies, you also need to know the vendor landscape. Couple of great examples here with me today, like knowing that there is and how to leverage them, whether that’s through a pilot, test and learn. They’ve done this before. I think, earlier on mentioned, you know, Heather, you start with ROI. Right? So they know the the use cases that may best apply to you. So that’s one good example to start with there. Internally, PwC tried to build some of these tools on their own. We ended up realizing that for some of our, you know, AI use cases, it was better to go with some of the commercial platforms and then develop with them and find their road maps and use them and leverage them. I’d say one other thing, like, when you build this, you gotta build a sustainable team. Don’t, you know, don’t pull people from the sides of the desk. Make a dedicated team or factory that’s gonna do this. Partner them with vendors to get their learnings, but also make sure you’re building your internal skill sets. You’re educating people, and that’s everyone. Right? So you have leadership buy in. Make sure they understand AI and understand GenAI. I can’t I mean, over forty percent of our internally submitted use cases weren’t even applicable to to these technologies. And then, again, clearly about communicating out to people, don’t treat this as a tech build only. Make sure you have employee engagement as you go. Develop them both the people developing it, using it, and then broader to your organization. Again, treating teaching them how to use that in a governed way. And finally, building in those feed loop back loops. So as you test and learn, how are you building in the feedback loops to what you’re learning, adapting the process, then making the next use case better. K. Great. Great. Thank you, Scott. Thank you. And so, Chris, for a final word before we get to the questions, what are the key takeaways for insurers looking to enhance their AI readiness through core platforms? What I would say, and and as a shout out to what what Scott’s been saying the whole time, is start with business outcome. Right? Keep it small. Start with business outcome. Know what your long term vision is. But if you start with the business outcome that that you wanna reach, and then think about the time that it’s going to take to execute on that idea. Right? I mean, the biggest barrier a lot of times to driving constant innovation, meaning perpetual, is is it gets really hard to invest the time and ultimately, therefore, the money into changes in your existing system. So think about, you know, with the technology I have, how easily can this thing be configured? Is the return there, you know, Tethr’s point about ROI? Am I can I expect a return? Am I saving enough pain with this use case that I’m implementing? Can I drive value to my user base to make this worthwhile? So I I I would start there and always keep in mind, you know, it it’s never done. Right? It’s not build one thing and go. It’s a constant learn, evolve, adapt, do the next thing. And always thinking through, you know, can I support these use cases in a really simple way leveraging the technology that I have so they don’t become big complex projects where we get bogged down? And I can start to show continual value in driving those use cases through my user base. Getting them used to accepting that new technology and a and a change to the way they do things. Change is hard. Change management is hard. We all know that. People people don’t adapt easily to new ways of doing things in their job when they’ve gotten so good at doing their job through time. Right? So show them the value. Show them how it makes their job easier and continue to demonstrate that with each use case that you go through. K. Great. Thank you, Chris. Thank you. And now we get to our q and a section. Again, as I mentioned at the beginning, if you have a question for one of our speakers, please enter it in the q and a panel on your console, and we’ll get to them as we can and as time allows. We probably have time for about three or four questions right now. But, again, if we do not get to your question during the live q and a, you, we do send these over to our speakers, so you may receive an email response as well. We encourage you to ask questions. We love questions. Please go ahead and put them in the q and a with your mind to console. So the first question I have, is your platform adaptable, for commercial property claims? Yes. Absolutely. So commercial property claims are something we’ve we’ve been handling for a long time. It it’s a big line of business for us. Certainly, it brings its own challenges. Right? It’s a it’s obviously a unique line of business. And depending on the organization that you’re working with and the size of clients to whom you’re selling commercial property, you’re dealing with very large property schedules and a lot of cope data that goes into those at the beginning. You’ve gotta be able to track against all of that data, on the backside with claims. It is it’s a very large line for us. Thank you, Chris. Thank you. And, this next question we have, so what are some of the biggest concerns insurers face from regulators with leveraging AI in the claims process? The first thing I’ll start with, don’t think regulators are hundred set up to speed on what carriers are trying to do, and they wanna know our use cases. They would not wanna know what we’re doing so they can learn as well. From their standpoint, it’s obviously bias in any decision. And and so when I say bias, like, are you letting us make a decision for you? Don’t let AI write the check you can’t cash. Right? So, like, how do you build that human in the loop to verify that, like, they attest to this decision or something’s there? So don’t let AI go wild. Second two things are explainability of the decision or the the any transaction. This is more around pricing and other other areas, but I’m claimed as well. If you’re making a decision, can you explain why you did that, and what is the data lineage to why you made that decision? K. Great. Great. So let’s see here. Let’s ask another question here for Heather. What emerging AI technologies do you believe will have the biggest impact on claims processing in the near future? Sure. I I really think it’s all about embedded intelligence across the core system with the system of intelligence. And what I mean by that is because we’re combining predictive models with gNI, we wanna give explainability. We wanna give the next best action. It’s really a a it’s a data game. And so, you know, within our platform, if you are getting summarization of what’s happening on the documents, whether it’s in a medical sense or on a legal demand package, that data needs to then be enriched back to the actual triaging of the case on a daily basis. What is then the summarization of that to the case and then the claim summarization, to that? So every point that’s happening along the case, there needs to be data enrichment that’s helping with prediction, that’s happening with the summarization, that’s happening along the way using GenAI techniques. Then you’ve got explainability that’s happening. Then you’ve got all the different routes that for next best action. And just like Scott said, you know, we have to very much make sure we stay, you know, bias testing. We do that all the time. We have to make sure we stay nonprescriptive so that we keep our human in the loop. But I just think it’s really all the the techniques that are out there right now with machine learning, predictive analytics, GenAI, then the, you know, business intelligence for the trending. It’s all coming together, and it’s allowing, really enrichment, across the claims processing and popping up whenever something happens, an event, to be in front of that adjuster. K. Great. Great. Thank you, Heather. As a reminder to anybody and everybody in the audience, if you do have a question for one of our speakers, please enter it in the the q and a panel on your console. We still have a couple minutes left. So you may hear us discuss your question and answer your question on here. Or if not, you may receive an email response afterwards. So let me see here. For Chris, how do you see core platforms evolving in the next three to five years to accommodate the rapid advancements in AI? So great question. I think, you know, we talked earlier about the the ecosystem, right, and the ease with which you can integrate with new ideas, that are being, you know, generated in the life cycle of the claim. Right? New vendors, new ideas from existing vendors, new solutions that are there, getting that data into the system. You’ll continue to see a huge amount of effort, and development, you know, across the vendor landscape focused on that part of the process. I think the other thing you’ll start to see, and I know things that that we are working on, are how do we facilitate the user experience for the adjuster and the claim supervisor and even executives and start to enable GenAI to make the way that they navigate, the way that they raise insight within the system, and ultimately, the way that they take action or drive action within the system becomes easier and easier. And so you have less and less, sort of ongoing interaction with with support to configure simple things like reports or setting a new workflow. Right? The idea that I could tell the system to show me all of the claims that are being handled by this particular attorney that was flagged by AI as a potential, you know, high risk plaintiff attorney. And then I can look across that set of claims and evaluate and say, yeah. I I I think this really is an issue. I wanna make sure that that I’ve got my eyes on all of these claims going forward and that that we’re really getting these routed to the right person. And instead of having to go to support and saying, hey. I wanna set a new workflow or I wanna create a report that shows me these things, I I just tell the system to raise an alert to me whenever we have any sort of incident or claim where this this particular attorney is involved. Right? And that idea that the system starts to the way that you interact with the system starts to change and evolve, and it becomes simpler to use. And it really helps that next step from, you know, both driving insight. But then, ultimately, the more important step is is taking action on the insights that are raised by AI. K. Great. Thank you, Chris. And, so now how often is the AI validated or updated? That’s a pretty tricky question depending on what AI you’re talking about. Like, the LMs, if you’re using commercial, it’s up to them. And if it’s yours, you’re doing as off as my and as you want. And, also, I don’t it’s a tricky question. Guess, Clara, like, or Heather Hoffen, do you see that in from your models? Like, are you doing that? Right. Absolutely. Let me, let me take that for you. So, we are constantly testing testing our models. We have, obviously, in our SLAs with our contracts with many of our carriers, and that’s, with our document intelligence, we have an accuracy level that we have to meet of, many times eighty percent or above for accuracy. So it’s very important, for that piece of it, that we get it, right, especially with the medical language and the legal language. On the accuracy for, our claims guidance, It’s at the same, threshold, but we’re constantly testing those, you know, if not weekly or monthly basis just to make sure, that we continue to get that right accuracy because there’s more data that we’re constantly refreshing that’s coming in and more and more cases that are coming in. So it’s a part of our, you know, our framework, be it from a regulatory, from a hygiene, from a accuracy, from an SLA contract. So it’s a big part of what we do. Thank you, Heather and Scott. And, I think we have time for one more question. But, again, if we do not have not gotten to your question, please go ahead and send it in, and we will get these over to our, speakers, and you may receive an email response after the event. So let’s see here. I just have a question here about security. Could you maybe give, our audience an idea of, some of the security considerations and the like or security vulnerabilities? Yeah. I I think that’s, you know I’ll start with us. I mean go ahead. Oh, no. No. Go ahead, Heather. So, obviously, just like with regulatory, you know, we we obviously have, a large set of global dataset, and we make sure all of our cyber certificates and it’s it’s beyond table stakes to make sure that we have the the strength and the hardening of that environment, and that’s something that’s, you know, is a part of our onboarding that all the security evaluations. And, we actually happen to have the former CTO of the FBI as our CIO. So we really do make sure, that, especially with all these claim cases, that we have the right, security protocols in place. Yeah. And I would I would add. I mean, obviously, as as a cloud native vendor, we were we were born and raised in the cloud. You know, security has always been first and foremost. Right? The we absolutely have to have tight security, not just in the AI that we’re integrating with, our partners that we integrate with the the language model that we’re using. It it’s in everything that we do. Right? It it’s it’s foundational to the system. AI is another layer. And and, certainly, you know, some of the ways that we’ve, you know, helped to address concerns with clients are are to build our own, you know, private LLM to allow the option for clients to take that and further build their own on top of that so that they’re not even using our collective LLM. They can use their own. You know, the the stories we hear about in concerns, especially with Gen AI, are around the public models. Right? People are concerned about private data making its way into a public model. The easiest way to to avoid that is not to feed the private data to the public model. Yeah. It’s really a point. Add is that adds on your third party risk, like, making sure how your third party partners are using your data. And and, again, back to Chris’ like, he’s following best practices, but maybe everyone isn’t. Making sure they’re doing that is very important. And we did to learn learn a lot through into internal employee policies and educating our people about how to use them and not put you know, we we have audit clients. Right? We don’t wanna be putting that these things in these these public models. And so educating people and updating your employment agreements are other good things to do. Right. Right. And so that does bring us to the end of today’s presentation. I wanna thank, thank you to our panelists for sharing their insights and to everybody in the audience for your participation and great questions. Remember, staying ahead of the curve in the PNC claims requires not just understanding AI, but also having the right partners, platforms, and strategies in place. We’ll be following up with a summary of today’s discussion and additional resources. If you have any more questions or want to dive deeper into any of the topics we covered, feel free to reach out to our panelists. And if you missed any part of this presentation, this webcast will be archived at property casualty three sixty dot com slash webcast so you can view it again or refer to someone else. You can also explore other upcoming and on demand presentations while you’re there. Thanks again for joining us today, and we look forward to seeing how you prepare for tomorrow’s claims today.