The Path to AI-Powered Insurance Strategy: Building the Data Foundation for True Program Visibility  

One. Welcome to Red Hand Advisors Risk Tech Webinar Series. I’m Patrick O’Neill, the founder and president of Red Hand Advisors, and excited today to have Ryan Cantor from Origami Risk who Origami is sponsoring our webinar today. We’re gonna talk about and I’ve gotta look over at the name of the it’s a long name, I’ll get it right. The path of AI powered insurance strategy, building the data foundation for true program visibility. Before we get started and ask Ryan to introduce himself, we were having a little we were having a little fun back and forth pre call about the woes of living in the south and getting hit with storms and ice. And I know no one really probably cares too much about Atlanta and Dallas. You know, when it get cold here, we usually have pretty nicer weather, but we’re both, we’re both, living through that right now. Glad to see we both have our uniforms on to keep ourselves warm today, Ryan. Doing doing whatever we can, Patrick. Thanks for having me on. So Ryan Cantor, Origami Risk. I’m Chief Product Technology Officer. I’ve been with Origami for almost two years now and really I live in Dallas as Patrick alluded to. And I’m a native New Yorker, so I’m used to it and I think everyone else is being crazy right now, but we’re working through it. We will survive. And I know no one in the north feels sorry for us whatsoever. But, no, I’m really excited to spend some time with you today just talking generally about kind of program management, data structures, how to get ready for AI, how people kind of start to think about things the right way. Appreciate you having me on. Yeah. Well, I’m glad to have you and I am excited. Thank you for, Origami participating and sponsoring our webinar today. So, Ryan, you said two years. I feel like that’s flown by because I remember I remember first meeting you at, I think it was one of the Origami conferences. You were you were still relatively new at that point. But one of the things that struck me in that conversation, I don’t know if you recall. I’m sure you’ve had many conversations with with clients and customers and and other people is we you and I got right into talking about kinda artificial intelligence and kinda where technology was going and how that fit into kind of how Origami moves forward. And maybe that would be a great place to start is just talk a little bit about that journey kind of how Origami sees AI fitting into, I guess, you know, for our topic, broadly speaking, insurance strategy and more specifically insurance programs. I think that would be a great great place setter. You know, listen. I think there’s a lot of things AI can do, and and sometimes people get really, really stuck on, you know, kind of the the the future, you know, the the end state, which is great. We should all have ambition. We should all have vision about where we wanna go. But when you think about, you know, specifically about AI, it really should open the door to kind of better insurance decisioning. Right? Where Origami is a big data platform. Risk managers have just immense amounts of data that they need to collate and manage on a day to day basis. Right? And so thinking really about AI in the most simplest form to get started. Right? Best best way to do is get started. Just think about it as a trusty assistant that doesn’t talk back, doesn’t call out, doesn’t get sick, and quite frankly, has a higher degree of accuracy than than really any human. Right? Give it give it a document and say, here are the data points I need from it. It finds those data points and it brings it back better than a than a human would. And so I always like to say, I talk about this all the time with my staff, when they’re thinking about I mean, because obviously we’re building AI features for our clients, but we’re also using AI internally for efficiencies and building better product encoding and helping our clients and serving our clients better as well. And the way I refer to them is the same way I’ll refer to to anyone, and that says, think of all the things that you hate about your job that are tedious, time consuming, and quite frankly, aren’t really commensurate with the level of skill or the strategic thoughts that you’re doing, that’s a great place to think about, well, how can AI do that task for me and make me more efficient? Right? And so think about, you know, the challenges though that we’re gonna talk about today. How do you get from here to there? Right? So once you frame the eye and you kind of know what you wanna do with, you know, we’ve talked to a lot of our clients. I’m sure you have as well. Right? But risk managers have data all over the place. Right? Spreadsheets, emails, it’s not a lot of this data is still is not in a system of record. So the idea that you wanna now use AI to do something strategic with it, we’re kind of missing a building block there. The other one that’s really interesting to us, this kind of took us a little bit aside for a minute, is that when you really talk about kind of insurance placement, it is a project. And lot of our clients that we talked to, a lot of our managers weren’t talking about it that way. But if it kind of looks like a doc and it sounds like a doc, it probably is a doc. Right? And so it has a process, it has statuses, stakeholders, feedback, questions, loops, decisioning. Right? It sounds like a project. Right? So it is a project to be kind of managed. Right? It is a project by far and away. Right? So even though people but sometimes it it’s good just say it out loud. Right? So people hear it. They think it. They be it. Yep. Go go ahead. No. Sorry. I was clearing my throat. I should use the mute button. Oh, no. You’re fine. So and then we think about okay. So if it’s a project, then how let’s let’s start at the beginning. Right? And and if we can use AI to ingest the data to remove that human barrier, right, and we’ll talk a little bit more about this, but that’s really the first hump. Right? It’s it’s that that throttle. It’s that burden of how do we get the data in the system in a more meaningful, more accurate, less kind of, you know, menial way. Right? And how can we use AI to automate that so that then in the future, we obviously can do all kinds of analysis and trends and tracking and help our risk managers, you know, make better decisions? I guess my question for you a little bit would be, you know, what examples have you seen, right, where, like, this thing is these things are breaking down for risk managers? Like, we’ve talked to some of our clients, but what are you seeing kind of in your interactions with people where these kinds of issues are kind of either in the way or blocking people? And are they, or how are they starting to think about tackling them? Yeah. There’s a couple things that come to mind, and I and I wanted to also just comment on the whole concept of it being a project. And and I I’ve said this I I know I’ve probably even said it on prior webinars. The activity I’ve seen in just, I’d say, the last two plus years or so of of clients, we get involved in either helping optimize selecting a system for the first time, and maybe even selecting a system is a better way to approach this comment, is being driven by the fact that the complexity of the renewal process has gotten overwhelming to them. Right? And I don’t think, historically, people considered it, and they may still not use the word project. Maybe maybe, fundamentally, you’ll you’ll ingrain that in them today, and they’ll they’ll come back and they’ll come away and say, yeah. I’ve gotta really consider this a project. But the process that’s involved in a renewal and I think, historically, it it may have been more straightforward, but the amount of information that markets and brokers and insurers require for to go through a a renewal and an accurate renewal and get the best outcomes, it’s all around data. And we’re and we’re gonna dig in in a bit more into that. So that that would be the first that’s the first thing that I see. The other piece that gets in the way here is it’s not a it’s not a single it’s not something that one person or one a team can own. You can’t say this is the risk management responsibility, and we own all the pieces. There’s there’s a lot of moving there’s a lot of other parts. There’s the broker. There’s the markets. There’s your own operations. Right? Because you may not have all the data that’s required to go through that process. And I think that’s where the complexity has led to organizations saying, hey. You know, I’ve done this. Even and this is true. Even organizations that have that I’ve seen that have had REMS systems in place for years that didn’t manage their renewal process or their program within their REMS. They kinda primarily focused on the claim side, are starting to say, this is getting beyond what I can do structurally in a spreadsheet or or a document or even a shared document with a broker. So that’s kind of that’s that’s the big things that I’m seeing in the market. And then, you know, kind of so in your conversations with these people, right, I I’m assuming I mean, obviously, you know, we’re talking about AI here as well. Right? AI everyone has AI ambitions. Maybe it’s internally, you know, ambitious. Maybe it’s your your supervisor, your leader saying, what’s your AI strategy? And quick come up with one. Right? Whatever it is. Like, that that Right. Yeah. Where do you see, like, you know, kind of examples of the kind of the gap between kind of readiness versus ambition? Right? Like, where they wanna be, but, you know, do they understand kind of the the the the foundational items they need to put in there? And and I’ll I’ll put you on the spot a little bit. Do they understand how to make the case, right, like, strategically that some of those some of those foundational items need to get get in place in order to achieve those AI ambitions? Right? What what do you see? I I’d say most organizations don’t don’t have it figured out yet. Right? They they get they get excited about the the opportunity of what AI can deliver, but realize that it’s just you know, it’s not like a light switch. Right? You don’t just throw the switch and all of a sudden, you know, we turn AI on in Origami, and all of a sudden, it solves all of our problems. Right? That there’s having good data. It’s having the right data. And I think even more fundamentally, and when I work with my clients on AI strategies because my clients are getting asked that same question from their own organization. Right? They’re getting asked, what is your AI strategy? And you think about typically an internal risk management or claims department. Their focus typically is, you know, it’s an internally focused support center for the organization. They’re probably not getting all the right attention they would get from their larger organization. The or look. Their organization, I’m sure, is is focusing on AI. They’re thinking about how it affects the way they work with their clients, whatever their business is. So they’re turning the clients I talk to who wanna figure out what their AI strategy is, always try to take a step back and say, what are you trying to solve for? What are those pain points? What are the things that you’re not that you you think you could do more effectively? Let’s forget about AI for a second. Let’s figure out what we’re trying to solve for, and then let’s put the right pieces in place. AI can certainly help, and then we’re gonna talk about, you know, data ingestion. You’ve got it up there on this on on one of on the slide already right there, the fact that it can help drive ingestion around data. That’s a big piece of it. Then you can kinda start plugging in where AI can actually start to solve problems, and not just AI for AI’s sake. Right? It certainly has efficiency opportunities, and we could all be using the chat GBTs of the world. But there are there are more complex use cases that I think these systems can handle. So maybe it’s time for Ryan, if it’s alright. We’ve got a couple polls that we’ve set up. I think that will kinda maybe level set this discussion and see where where where our audience is with that. So, let me launch this poll on the screen for everyone. So if everyone could just take a minute and answer the question that’s on the screen, how would you describe the current state of your organization’s insurance policy data? So one of what we’re asking about is kinda one of the fundamental building blocks of the renewal or insurance process, is your data. So how how how good do you feel about that? We’ll give people a couple minutes, couple seconds to answer. We’re getting a lot of answers already. It’s good. Give it another second or so, and I’ll close the poll, and then we’ll share the results. Ryan and I Ryan, you and I, when we talked last week, we were we were kind of I think we were putting some bets on where we thought this would fall. We’ll see we’ll see where we get in a second. It’s actually fallen generally where I would have thought. Alright. We’ve got about eighty percent of the respondents. Let me close that poll and share the results. So here we go. So where do people that are on the call today find themselves? So centralized, structured, and actively used. That’s fantastic. I I I mentioned Ryan, when you and I practiced, I said, it we should give kudos to the folks that that that fall into that category. That’s a lot higher than I expected. And I would tell you from our historical Our historical research, even you know, we do an annual survey for our report at the we’re doing it right now for this year. But about a year ago, it would have been it would have been much lower. So I I’m I’m I’m pleased to see that. Partially organized, fragmented across systems, kind of the big two, I mean, just partially organized with gaps and inconsistencies, almost fifty percent of the data. I guess I’m not surprised by that. I’m actually but what I’m what I’m actually the most surprised about is that and maybe and a good and a good way because I I’ve got some thoughts on this, that it’s not largely managed by your broker because that that certainly sets some limitations. And I think particularly not having data ownership in your own data in that process is real important. So I think it’s encouraging to see we’re kind of somewhere in the, you know, fragmented to partially with a with a percentage of folks kinda moving towards, you know, more centralized structured process. So, you know, Ryan, maybe maybe the next place is, you know, how do clients I guess, when you talk to your origami clients, where do you see that where do you see your clients now, and where are you trying to get them to? Yeah. So we we asked if you flip flip this slide. We actually Yeah. We surveyed and we talked to ten clients on the spectrum. This is our interpretation of the spectrum of our complexity in terms of various policy and program management capabilities. And obviously green is easier and red is really, really complex. Right? And we just asked them, where are you at today? This wasn’t a seek of ambition of where you wanna be. This was just an accurate you know, first step is figuring out where you are and where you wanna go. Right? And so this was where are you at in kind of the life cycle? And what you’re looking at here is, you know, you know, four of the five four of the ten are in that orange category. But I will tell you as a caveat, those four clients, a lot of brute force to get there. Right? Like a lot of manual processes. There’s a lot of commitment. It’s actually unclear from an enterprise perspective that if key individuals left the organization, if they’d be able to kind of keep maintaining that. Right? So there’s definitely kind of a challenge there. But if you flip forward one slide, then we ask the ambition question. Right? And we said, okay. Well, where do you wanna be? Right? Because this is the first step is finding out the delta between where you are and where you wanna be, and then we can start unpacking what’s preventing you from moving from point a to point b. Right? And what you see is overwhelmingly everyone wants to live and have these kind of capabilities in there, right? Obviously there’s one lonely client who is completely complacent and totally fine with just being able to manage their quotes and have at them. If that’s what makes them happy, no problem. But I think we would agree that ninety percent living in a place where they really would love to have more keys to the kingdom, more capabilities. And again, our primary unlock for this was data. It was that the nature or the manual nature of imposing all of your kind of policy program data from quotes to endorsements, everything, every artifact one by one between accuracy, checks and balances, time consumption, sifting through, you know, huge documents to find kind of needle haystack problems to find the pieces of data you’re looking for. They just don’t have the time. They don’t have capacity. They don’t have the resources. So it wasn’t a lack of ambition. It really is a blocker on capability at this point, which, listen, as a software provider, it gets me really excited because that’s a problem we can solve. Right? So, the thought this was really, really telling, as it kind of, you know, illustrates a little bit of, like, what we’re seeing and of where people are going. And so if you go to the next step, we then kind of helped unpack a little bit further, and we started asking questions about, like, well, if data is the primary blocker, how do you view data? Like, are you viewing this data as a strategic kind of in what we’re calling infrastructure? Are you viewing it as a strategic asset? Right? Because what you figure out is we get all these kind of historical assumptions. And, again, you can you and I can kind of tag team a little bit because I’m eager to hear what you’re seeing and and if this resonates with you. But, really, like, this policy data is treated as static documentation. It’s kind of a moment in time. It’s kind of static. I know where the file is. I can kinda go get it when I need it. It’s there, but it’s not kind of living and breathing. It’s not standardized. It isn’t morphed in or loaded kind of into any system at all. Is that consistent with what you’re seeing as as kind of how people manage this at scale? Yeah. It is. I think that’s I think, you know, again, back to the whole complexity of this and thinking of this as a project rather than just kind of a one off situation, I think historically people have treated this more very transactionally and have kind of had that that one off kind of here’s the dot know, here’s the policy, and take the old school. Here’s your policy, and I stick it in a drawer. Right? And I move on, and I live till I live till the next renewal. But there’s so much you know, part of it is just unlocking, and we’ve already we’ve we started already touching upon data, unlocking the data that’s necessary to really get the insights you need from from this information. Right? So the the opportunity most organizations, you know, when they’re tracking what I see is when they’re tracking this data in a centralized system or a Remus like Origami, they’re tracking some really basic header I’ll call it header information about the policy, and then they’re attaching a PDF. And historically speaking, that didn’t really do you much good. Right? You can always refer to it, but you didn’t really have those answers kind of at at the tip of your finger at your think tip of your fingertips to get answers. And I think that’s that’s another another thing that I’ve seen quite a bit that I’ve heard challenges from clients talking about, like, that whole, source of truth. I think that’s the word you used a few minutes ago about about having a source of truth for your risk data, for your program and policy information, is that far too often I hear risk managers say that they don’t have the answers they’re looking for at their fingertips. And so if you take that example that I just gave. Right? So it’s a policy specific question. You know, do we have certain limits? What are our exclusions? And it’s buried in a PDF right now that, again, technology can help us with that now. AI can help us with that. But historically speaking, and many organizations haven’t taken that leap yet, is that data is not right there at their fingertips so they can say you know, I think far too often what happened is they get that question and they say, hey. I’ll get back to you on that. I don’t know if it’s included or not included in the policy. Right. Which leads to the kind of this now the second and the third. Right? I think these two are kind of intertwined. Right? And that is whether it’s in a system or not, it’s still fragmented. Right? Various stakeholders have different parts of data, different parts that they need or have early access to. So you just gave a great example where, hold on. I’ll get back to you, and I go, you know, spend time scourging through files and trying to find you answers. Right. Question is, you know, it really people our clients were self aware. It’s limiting their ability to make decisions. And in some case, were some examples. I don’t know if you have examples where this data is actually actually lended to coverage gaps or or or, you know, kind of or or they’re missing trends that are happening underneath it. I don’t know if you have any examples of that, but, like, our clients were able to kind of talk a little bit about how the fact that the data is sitting in pockets, that it’s not organized, it’s actually leading to risk. Right? I mean, it’s it’s leading itself to to higher higher cost of risk. What what are you what are you seeing? Yeah. I think there it’s definitely leading to to gaps. There’s a couple things at play here. Right? First, the whole, like, the whole concept that all the data is not in one place. Sometimes some of the data is is managed at the broker level, and it may not be at the client level. The other piece and and the the connectivity isn’t there always is that, you know, relating that policy data to the claim data so that you can see kind of, you know, where erosion’s taking place and actually so pretty standard capability, but people not connecting the dots. So that that’s another another area. It’s the this whole, you know, fragmentation of the data and having again, taking a step back for a second, what what are the things we wanna be monitoring? What are the things that what are the analytics we wanna be looking at? Then that solves the problem. Okay. What do we need to do? We need to connect our claims data with our policy data instead of just, you know, oh, well, we’ve got we’ve got these great capabilities, but do we have the right data to to answer the questions? That’s great. No. And, again, all of this I mean, mean, maybe this is maybe this is a little hitting too close to home and and people watching are feeling little triggered because we’re we’re kind of uncovering some pains. But this again, this is just what our clients have kind of shared with us and uncovering kind of know, we talked about where they were, where they wanted to go, and really it’s it’s getting in here, and we’ll talk about it for a minute. But, you know, the next part was how do they make the case that that that these are kind of essential things then, right, that we need to kinda make some investments or or invest some time, you know, slow down to go faster kind of things to get the right framework in place. But I think I think you have another poll. Right? We were thinking about the other Yeah. We got another poll. So let me pull it up, launch the poll. How confident are you in your organization’s internal visibility when preparing for insurance renewals? So let’s see what people think here. I’m excited to see a whole bunch of very confident people. I think we maybe maybe we have some very forward thinking folks on our call today, but we’ve got we’ll give them a couple seconds to respond. I’ll throw out one other thing while we’re waiting for the poll to respond. Something that came to mind, Ryan, when you were just talking is and I I kind of alluded to this earlier, but the renewal the renewal process and what I’m what I hear from my clients is just getting more and more complex. And that the amount of information and data that the markets require is is growing exponentially. It used to be real simple to just provide some basic exposure data and claim his historical claim data, and you you could go through and go and get a renewal. And you can do that, but the confidence level that you get and the type of renewals and quotes that you get when you have limited amount of data or data that you’re not completely comfortable with, it causes significant problems. And I think that’s why it’s become a real important, I would say, top of mind for many customers looking to manage this process more effectively, get the right data in so that they ultimately can have the right outcomes. So let’s see let’s see where we are with this. We’ll close the poll. Thank you for responding, and I will throw it up there. Alright. More confident than I thought. That’s great. So forty percent, very confident in on where they are. You know, another almost half, though, really, in the next one, somewhat confident rely but rely on, you know, external filling gaps. I think the what the question probably doesn’t get at is is how how simple this is, right, or how you go through a complex process to actually get to this point. I think that might be. I’m not suggesting that that these folks aren’t very confident in the data, but there’s still a lot of room. There’s there’s well more than half of the audience that are that are still in a place where, they’re not confident about, the internal visibility going through the renewal process. And I think that speaks to that example that I just gave about, you know, having the right data, for for the whole for this process. So, next step, what do they do, Ryan? Like, how let’s let’s give some let’s give some recommendations and suggestions to the audience on how they can maybe get there. Yeah. No problem. So let’s let’s let’s talk a little bit about, well, what is the foundation? We talked about the foundation and laying it right. So first, centralized organized boss record. I mean, that sounds obvious. Right? That’s fine. But we have to put AI ingestion in there. Right? Orgami has been working really, really **** ** that. We have some things coming out, which will be really, really exciting to to do that. But whether whether it’s us, whether it’s someone else, whatever the case may be, this is where you need to be thinking about using AI. Right? And and it’ll be you know, I will tell you is gone are the days of OCR, gone are the days of, you know, using providers or partners who are sending your things offshore taking days to do it and come back. Right? That that that’s that’s that’s that’s that’s gone. Right? AI is really, really great. Good news. We know what data we need. Right? And if suddenly tomorrow you’re like, oh, I want this new piece of data. No problem. Right? We have the base artifacts that are coming in that you need to ingest. It’s about using AI, about having an automated way of saying, listen. Here’s the data I need. Here’s the structure I needed in. Hey, AI. Go review these PDFs, documents, files, Excel files, whatever. Get me the data that I need, put it in this structure so it’ll automatically put it in the system of record, and have a little bit of human in the loop quality control in there. Right? Because, again, AI, you know, we don’t want you to have to go page by page by page and find this data. It’s a lot easier to have it all summarized for you in a nice tight package. You kind of look through. AI isn’t perfect either. It might have missed something. It’s lot easier to go find one piece of data than it is to find the whole thing. Right? So have that human in the loop process ingested in there, but think about that process. Think about how we’re doing that. Think about, you know, do you have all the documents organized today that you’re gonna wanna put through that process? Right? So get organized, get those things ready to go because this kind of technology is coming. Right? Now you have kind of data becoming in you know, ingested in standardized ways. So now you have a consistent structure across your programs and over years. Right? So now we start getting into trend analysis and being able to uncover or unlocking the ability to get insights because we’re not changing the rules of the road every year or every single renewal process isn’t using a different strict schema or a different structure of data. Right? We standardize those things. Now you’re enriching that data. Again, I think you you’ve pointed out whether it’s linking your claim data, other other other company insights, intelligence, third party data, whatever the case may be, you’re enriching your data to make incredible insights. But again, you can enrich it because it’s centralized and it’s standardized. And so now you’re appending to it. And then we get into kind of the fun part. Right? Everyone loves to jump to the last chapter, open up the last page of the book. Oh, cool. We’re there. Well yeah. And by the way, this whole process doesn’t have to take years. It doesn’t have to take even quarters per se, but it is it is methodical, and and there is no cheating. Right? There’s no going from step one to step four and skipping the things in the middle. How fast an organization goes through that, it’s entirely up to them. Right? But getting organized, having the platform and the the technology to unlock this capability is, of course, essential, and then moving you towards this idea that, okay. We have analytics that can use AI and and and can do types of analysis where I could ask questions of my data in real time and get answers, right, to to those questions as opposed to even, you know, static charts and graphs, right, because different situations, different stakeholders, different circumstances are gonna prompt you to suddenly have different questions, right, that that you have to ask. What if scenarios, for example? So I think that’s either from a structure perspective, how people should kinda get their minds organized going through this process. Does that help at all, Pat? Just kind of the way you can go with No. It does. And I think I think the exciting part I know and I know some organizations are starting to do this already, right, using AI to ingest, you know, that starting place, that first building block that you have there of ingesting the data because there’s so many you know, I’ve seen so many iterations of, you know, organizations trying to use OCR to collect the data, going through a significant manual lift to get that data in the system. And and most of those kind of fall have fallen short, or they’ve looked at the significant amount of investment it would take to actually get that data. And there’s some there’s some meaningful reasons to have that historical data in there. You know, if you can you know, it’s great to be able to do kinda what if scenarios with your with your current program, but to be able to look back and say, okay. Well, here’s what we had in place historically. These were the were the costs. These were our actual lost costs, and actually starting to look at what our total costs were for a particular, year or program, I think, is really important. So I think the the opportunity of just AI helping us get to a place where the data that a lot of the data is more structured, I think, is is critical. Making sure that you’re capturing all those so that you can, you know, put the right analytics in place. I’d also say even if you don’t have to get everything in a structured manner. If you think of my example before, like, what are, you know, what are the exceptions to this particular policy? What’s excluded? That that could easily be a prompt. Right? You can prompt the system to to to look through the policy, and and it can tell me what the exceptions are. I’ve had a lot of success just playing around with that to see if it actually works, and it works great. I’m sure people have lots of examples and think about things the way you’ve used LLMs to to do that type of stuff, and it it can work great. So we don’t have to have everything in structured in a structured format, But having real time access to those tools and, to that data, I think, is critical. And I think it’s I think it can be it is game changing. I’m not even gonna say I think it’s game changing. Because I I know that the the ability one of the things that the organization’s been many organizations have been trying to do for a long time is when they go through the renewal processes is really to do these what ifs. Right? What if I had this program structure in place last year or five years ago or as my the way my organization’s changed so that you can make the right decisions, going through this process. There’s not there’s not always one you know, there there’s not one structure that works for every organization, and and having that insight into your your your loss history, your program history is critical, for making those decisions. So it I think we’re an exciting place because what’s been limiting us for so long is that first step, is getting I I think we’ve been talking about having accurate and the the right data for so long, not just across policy. Just think about just managing risk in general. And we’re getting we’re getting we’re we’re taking a leap right now, which is, I think, a great a great thing. I think it’s it’s changing a lot of the dynamics that are going on. You you said something. I wanna dive into a little bit. Like, again, if I’m someone listening in right now and I’m thinking about how I get started, you did mention would you agree, though, that just at least getting started today going forward would already be a big step forward? Right? Obviously, it’s nice staff important historical. But if someone’s overwhelmed or resource constraint or maybe having a challenge with some executive buy in on time or investment to get going, would you agree though that at least getting started today saying, well, this is our go forward strategy would still Oh, for sure. Yeah. For sure. Yeah. I it doesn’t it doesn’t it still isn’t, you know, it isn’t, a zero time gain that we can just get all our historical data in place. I have clients who don’t even know where they you know, they they may not even have access to the files we’re talking about. Right? But having a go at a minimum, I think a go forward strategy is, you know, set set set a foundation for what you want that to look like going forward, and and work that in place. And if you can then go back in time for if it makes sense it might not make sense. That I think that question ultimately becomes a organization by organization decision. Is there is there enough value in that process? So, again, that’s part of that taking a step back on what are we trying to solve for, you know, to to to ingest all your historical policies just for the matter of ingesting them, and they’re not gonna actually play any role in in any of the analyzation and analytics that you might be looking to do, then that doesn’t make sense to do that. But we definitely wanna this is not something I think everyone should be starting to move in this direction because I haven’t seen many organizations that have their policy and program data in a structure that allows them to have kind of real time access to the data. Again, having, you know, significant amount of structured data that we can that we can use for reporting, for analyzing, and for monitoring. I think that’s critical. No. Okay. And and, like, again, I always like to figure out that some of the hardest things when I went soon as we talk about AI and data, sometimes I hear people’s eyes glaze over, and they get they get, you know, they they get paralyzed with complexity or this is this huge project, and am I gonna get funding or approval? And I’m like, listen. You know, when you’re in a hole, the first step is stop digging. Right? Like, sometimes the hole and then you just kinda say, moving forward, we’re we’re gonna we’re gonna chart a new course, here’s, you know, in the coming period, the value that’s gonna unlock. That’s Brian, the other yeah. The the other thing the the other way I think about that is if you just started with where you are today, it’s a test case. Right? It can show you, and it would help with that, you know, with the with the buy in that, hey. We started with our current program. Look at what it’s done for us. Look at the things we’re able to do. And then you it’s imagine if we had x, y, and z, what that can do. Or and in some cases, we’re we’re really not touching a ton on this at this point, but it’s it’s it’s certainly one of the the the pluses of going through this process. It’s gonna it’s gonna significantly improve the process itself. It’s gonna make you more efficient. You know, the this whole, you know, this whole process should be an efficiency gain as well as, you know, a an insights gain for for your organization as well. But I I always tell people when they start any AI program, you know, figure out kind of what the the pilot is. And and in this situation, what we’re talking about, I think the pilot is your current program, is your current year. Maybe it’s in maybe it’s even, like, just a specific policy within your program. Maybe that’s the place to start. Because then you can then you can tweak it and get it right. Because you do want what you do wanna get right is you wanna get that you wanna get that data structure right, not that that’s so hard to change. But you don’t wanna have to go back and say, well, we’ve done it we’ve done five we’ve got five years in place. Now we realize we should be car capturing x y z, and we’ve gotta go back and do it again. Not that that would be a complete redo, but, you know, really being thoughtful about what that structured kinda standardized data that we wanna capture is that’s gonna be helpful for making for decision making because that’s ultimately what you’re trying to do. That’s great. No. I mean, that’s that’s that’s well said. Yeah. So do we think everyone’s ready for this? Well, you and I you and I took bets. We’ll see. We’ll see. We’ll see how we’re doing. Last one, I promise. And thank you everyone for for filling out our polls as we go. So how ready is your organization to apply advanced analytics or artificial intelligence to your insurance programs. And if you’re if you’re already doing it in a significant way, I’d say your answer is ready, the top one. But look let’s see what we get here, and people are quickly filling this out. This is one that has been changing very rapidly, Ryan. You and I know you know that. We’ve asked a we’ve asked artificial intelligence questions in a in a meaningful way for the last two. This will be now the third year in our poll in our in our survey, our our REMS report survey. And if you go back and just look at the change from this is twenty six from twenty four to twenty five, it was it was pretty significant. I mean, I think if we were asking kind of the same range of questions, you had a lot of people in not currently doing it, mostly unsure. And last year, we started to see people being kind of in that not ready mostly and then somewhat ready, although we didn’t use these exact categories, and not not any significant amount of folks in that ready. So I guess I’m not gonna be surprised at what we see here, but there’s this is something that’s changing so rapidly, and it’s it’s being, it’s being implemented, by organizations so quickly. So I I guess I’m not gonna be completely surprised. So let’s close the poll and show the results. And yeah. Okay. About the same. I guess each one of them has come out pretty similar. Ready ready. Our data foundation supports it. That’s great. Somewhat ready. Improvements need it. Not ready. Foundational work still to come twenty seven percent. So still a significant portion. Actually, this is one maybe I’m a little surprised just be just based on I’m I’m really interested in our survey that we’re doing for the report on where people are a year later, and I guess I’m expecting we’ll see more folks that are that are there. I think there’s a lot of it it it isn’t hype. Hype’s not the right word, I think, for AI, but there’s a lot of there’s a lot of excitement around what AI can bring to organizations and to managing risk and insurance programs. And and maybe we’re just at that I I thought twenty five was the year we’re gonna we were gonna see more operation that being operationalized. I think what I realized is most people were doing the things that you and I have been talking about kinda behind the scenes to get ready for that. So maybe this maybe this kind of meets us where where I’d expect people to be if I if I think through that. Any thoughts? You have any thoughts on it? I would say this doesn’t surprise me at all. I mean, I think even when I look at the somewhat ready and the not ready, I do think there’s a little bit of a delta. There’s a couple of deltas. One, where is your current data, and how is it organized, and how what is its proximity to AI capabilities? Right? So some of this is, okay. I might think it’s really, really structured, but is it in a system I can connect to AI, or how do I actually physically get the do the thing. Right? So, like, there’s a little bit of a logistical tactical problem I think the people are doing. Then there is, like, I call it the CIO problem, right, which is I have ambitions. You know, I talked about this in a webinar we did recently where regulators and states are starting to really maybe a little, you could say, overreact or not react, right, but it’s an emerging technology and so regulations are coming out very quickly around proper uses of AI and in different segments and different fields. That’s something to be mindful of, right? And so then you have a CIO problem which says, okay, well, you wanna do what? And where’s our beta? And what’s going on? And what’s security? And what’s the access controls? And, like, and you keep going, and you’re like, I just wanted to look at this document and answer some questions for me. Right? But here we are Right. You know, now and struggling with how I move the needle from point point a to point b. Right? So all of those things, I think, play into that question. Right? Like, readiness isn’t just always, you know, the the risk manager might be ready. It’s the rest of the organization that’s not ready or the infrastructure isn’t ready or the CIO isn’t ready or whatever. Right? I think there’s some other factors playing in there as well. Yeah. And and your and your and your core your system of record is is a big piece of that. I mean, one of the things that I’ve heard, and and I’ve certainly communicated this to to your organization as well as to your your your competitors, is most of my clients are waiting for their core system to provide these capabilities. They don’t right now, and I think, you know, I think as integration, continues to improve, we’ll probably see more things being bolted on and and and accessing data that might be the you know, again, that that central source of of truth, and that might be the Remus. It should be the Remus, I think it is, in this space. But it doesn’t mean that the capabilities have to come direct every capability has to come from the Remus. But right now, the what I’m hearing in the market is that most organizations are saying, that’s kind of my starting place. I wanna see what capabilities that because it’s already, obviously, would be integrated. It would be integrated with the dataset. But then the question is, you know, are we gonna be doing more complex things? Is that maybe not my ex my expectation maybe isn’t that comes from my core system, but maybe that is something that I do with the right connections and and and integrations because that’s something that technology is making, you know, maybe maybe not as sexy as artificial intelligence, but the integration capabilities that are available today, with the right systems, with with more modern technology is, is is certainly exciting on what what’s possible. Makes sense? Do you agree? Do you agree or disagree with my comments? No. I I know. I told I mean, listen. I totally agree. I mean, listen. If you go to the next next slide, I think, a little bit. Yeah. There’s, some kind of just steps or things to think about. I think it ties very nicely into what you just said and that is, listen, some of these things we’ve already kind of covered, start understanding current state policy data. People on this call have already stated overwhelmingly that I think they have a generally firm grasp on that. Step two here is, again, calling it a project, thinking about it as a project. But next part on step three is a little interesting because, I mean, you could argue that the poll for this this this session maybe is a little biased because if you were energetic enough to join, you know, a webinar to begin with, you might be on a certain side of the spectrum to begin with. Right? So Right. There’s probably a little built in bias in the poll, but it’s still great to see those numbers. But I will say this, and that is I you know, listen. I lead a software development life cycle here at Origami, and the closer you are to the process, the more you think it’s awesome. Right? And what has AI done for us is the same thing that AI is doing for risk managers, and that is it’s making you rethink your processes. So even the people who are like, I’m really confident in my processes, you might be confident in the process as which you have defined it and you’ve been running it that way for x number of years, but maybe it takes forty hours and it should take ten, or maybe it takes eighty hours and it should take fifteen. Right? And so there’s this little bit that I’ll challenge everyone to be thinking about, and that says with the emergence of technology as your core platforms continue get better, as origami gets better, our competitors get better, this all happens, you got to open up that mind, right? The first step in fixing a problem or fixing a process, admitting that there is a problem to begin with, right? Or admitting that there’s at least opportunity for us to create efficiency, right? And these all carry opportunity costs. You could be doing something else with that time. So I would just say think about that because when I talk about step three, identify the most challenging parts of that process and put plans in place to start systems. That means be willing to adapt or modify even if you think things are kind of working for you today. How could they be working? What does optimal look like? Right? How does what does better look like, not just comfortable or that I’m I’m I’m good with the current process? So just, you know, kinda be on the lookout for that. And the last part is just it’s a mental exercise, right, to be thinking about the use of AI in these systems. Right? That that’s and and the best way I could do that is, you know, you’re running your business every day. You’re doing tasks every day. And in the back of your mind, you should be thinking about, could AI do that? Or this thing just took me three hours. If I had x or if I had y or how would I do this? How could AI do this thing that I hate doing? And, again, I the mental exercise I do with my team all the time is pick the things you hate doing because those are usually the ones that you’re willing to be like, man, wish I couldn’t do this anymore. How could someone else do this? Well, just take the someone else and put the word AI in there and then work backwards and figure out what needs to go in place to help solve those problems. But it sounds simple, but it’s the hardest part with all of these things is knowing where to get started. That’s why I was asking you, like, even if you to your point, I think you said it great. Like, if you just start with your current programs, if you just start with your current policies, your current renewal cycle, just get started. Right? It Yeah. It gets better the longer you do it. Right? Even if you don’t have the poll you know, the bandwidth or the resources to to to load historical. Same thing. Mental models on, like, how are you gonna use this? You just gotta start somewhere. And then I would say start with things that are the most frustrating, most time consuming, or the things you hate personally doing and figure out how how how to work through those. Yeah. I I think I guess I’ll just I’ll just add a tweak to that because I do think starting I’m just thinking that step three, identifying the most challenging parts of the process is important. Right? That’s that what I was saying before about taking a step back and kinda coming up with, you know, what are what’s that list? What are the what are the things that are the most difficult, most challenging? What are the things where we don’t have an answer right now? Right? But once you do that, particularly if you’re just starting, you can prioritize that. And, you know, one of those items might be the one that would give you the biggest impact. It also might be the biggest lift. And if that’s the case, I would tend to recommend that’s not where you start. Right? I would start with something where you can start to see some positive momentum, things that might be little little more low hanging fruit, but still challenging. And, again, we we actually didn’t spend a lot of time. We just touched on it a few times talking about the efficiency gain in this whole process. I’m I’m actually glad we talk more about the the actual results than the the efficiency gain, but the efficiency gain is so important as well. And if we talk about, you know, this part of the the risk process, the the renewal, the ins the you know, managing your insurance program. It’s it’s one of the most complex parts, if not the most complex part of of the of the cycle. And while, you know, you call it’s interesting. I love the idea of calling it a program, but it almost doesn’t have an end. Right? It it it it shouldn’t be a it’s not a three or four month cycle. It it’s the it’s the pre and all the way through that that insurance program. And then it’s again, obviously, then it starts over again. But it’s not kinda like a one and done, whether that’s two months or three months and you put it again, put it in a drawer. There there’s monitoring. There’s so much more that that can be done. So I guess my I think we’re beyond now the whole, you know I like the idea of people kinda playing around and figuring this out. Just my example before where, you know, I took an insurance policy and stuck in a chat GBT and started asking questions. I mean, I I couldn’t find things that it would it couldn’t answer for me. And that’s just working with one document. But I when I do talk to organizations that are kind of on the beginning end of the scale of using AI or maybe haven’t even haven’t even got there at this point, that’s the place to start. It’s just start to experiment a bit. Start to see, you know, Origami, you have capabilities that are already out and available. Start to, you you know, start to use what’s available and see what type of results you can get, and then start thinking about how you tweak that or what other data do I need or how could to your point, how could I apply that somewhere else that it would make an impact? I think those are all those are all great ways to get started. Let me ask you one question. I think we could probably go to q and a this year. Yeah. We’ll go to q and a after Yeah. I have one question for you, and that is we just talked about a lot of things. I’m an individual. I’m watching this. Right? How in your mind can they frame effort or investment that they may need to do to get here as more of a strategic initiative and not just kind of clean up work, right? Like how do people kind of get buy in on the value of what this will do for them? That’s a great question. Because if I’m if I’m sitting on the other end of this thing, I’m like, great. Yeah. I wanna do all these things in my free time. Right? Like, so how do I there definitely is is some some investment needed. Right? And so the question is, how how do they go get buy in for that? Yeah. I I think so I I think part of it first of all, I think I think the the the larger buy in from your organization is probably there. I think most organizations are are grappling with the same question around different parts of their organization. So I think taking the initiative of figuring out, hey. How can I use AI in my practice and in my responsibilities to improve that process? I would also make this part of your overall this should be this should be a team effort and not one particular one particular person. I had a conversation with a client recently who said, you know, they brought in someone who’s kind of, you know, who’s gonna kinda drive technology and hopefully move them forward with artificial intelligence capabilities, but kinda disconnected from the rest of the team. And I don’t think everyone has to be an expert. These tools are getting easier and easier to use. And so I think making it a a a project for your team, for your whatever your organization looks like, for you everyone that’s responsible for this, everyone has insights. Because you may not know if you’ve got that person who has that responsibility. They might not actually understand where the real pain points and the things that we don’t to your to your point earlier, the things we don’t like to do or that take a really long time. And then surfacing those ideas and then figuring out how you can solve for it. And, again, I think if you can if you can knock off a couple of low hanging fruit or solve some problems, I think that starts to build that buy in and and build the case. Something I heard at a conference, and this was just about a year ago, so it’s it’s kinda interesting how things change so quickly. Someone in the audience I was on a panel talking about artificial intelligence, and someone in the audience asked, you know, is this something I really need to be concerned about? Is this something I need to be using within my organization? And the one of the panelists that was sitting with me took the answer, took the question, and said asked reverted the question back to the audience member and said, you know, I I’m not gonna ask you how old you are, but are you you know, if you’re not within a couple years of retirement, this is gonna change the way you do business. And if if you don’t start to start to think about how it’s gonna be applied with your organization, you and your organization are gonna get left behind. And I think that’s the case. This whole concept and I know I’m going on on this one thing. But as I talked earlier about how complex the renewal process and the data that’s required, I think AI is also gonna make that that requirement go up even more substantially. So if you’re not getting around your data and understanding what it’s telling you, you’re gonna get left even further behind. But I think it’s those early wins, Ryan, that are gonna help, you know, get buy in from your own team, your organization, what have you. I think that’s the to me, that’s the place to start. Fair enough. So q and a, I should have mentioned this at the beginning of I’ve got a few already in the box. In the, in your control panel for the audience, obviously, everyone’s on listen only. There’s a q and a or questions bar. You can type your questions in there. We’ll try to get to we’ve got some time. We’ve got a few minutes to answer those questions. It looks like some people are typing, but I’ve got a couple I’ve got a couple here that we can start with Ryan, and we’ll see we’ll read them, and we’ll see who who should take them first. So first one, you talked about the importance of strengthening your data foundation before expecting more value from analytics or AI. For a team that knows this work needs to happen but has limited time and resources, where would you suggest starting? I actually think I kinda touched on this a second ago. Anything else? Can you think anything to add to that where where you think people would start? Yeah. So listen. I I think I think you your next renewal, right, or or your next quote, right, or or just start with what the work is that you’re already doing to at least get some initial value. Show someone what the value is, and then, you know, at least you then have a business argument. Maybe there’s some temporary resources, other stat you know, other things we can do to accelerate or go faster. Right? And then it’s business decision. Do we want to slow adopt this over time during the normal course of doing business? To your point, do we want to put some historical data in or at least get all of our active policies and programs up and running? Those are all decisions. But the answer is all of it. Right? You can do anything, whatever your budget, time, resources kind of allow you to do, but it does make a decision that says, hey. Let let’s lead here. Let let let’s say starting today, this is what we’re gonna go do. And even if it doesn’t give us, you know, kind of an immediate thing, which I think it will, you know, you’ll be able to get some value proof points pretty quickly to articulate what it would mean for the business for people who may not understand why you you wanna you wanna invest those resources. Yep. Agreed. Alright. Next one. You mentioned that policy data is often fragmented or broker controlled, which affects renewal strategy. How can risk teams improve their own visibility and insight while maintaining strong collective broker relationships? That’s a great question. I was actually gonna comment on this earlier, and I and and moved on from the point. This even if I at least, I’ll I’ll take a shot at this. Right? Sorry. Even if I think there’s a there’s a case to be made that this should be managed by the organization. It’s your data. You know it the best. There are certainly roles and important steps that a broker has in this process. They in a lot of cases, they have your historical data. I’ve seen many organizations where when they wanna when they’re starting the renewal process and looking for the exposure data, it’s the broker they go back to to get that renewal, that exposure data. I think that data you should own that data. But that’s not to say, I think what we need to see is more stronger collaboration and not just by trading documents and and spreadsheets. I think I think all the parties or at least most of the parties should be on the same platform. So if you’ve got a Remus and you’re using it for your renewal process and renewal management, I don’t see any reason why you wouldn’t have your broker to be part of that. I’m not sure I’d give them full access and edit ability, but getting them in that process so that they can be part of a a workflow that might be built, a review process that’s in place, I think that’s important because there are too many systems that have to be looked at. The other thing I would I would suggest to organizations is to, when you strategize about this, have your have those parties, and particularly, I think the broker, being part of that conversation. Because even if you have the most complex systems in place to manage and manage your program, collect that data, ultimately, it is gonna get to the broker has a step in that process. They’re gonna have their own tools that they use with the markets, And you wanna make sure that those tools work properly together and that you’re giving them if you have all these great insights, that they have the ability to use that data, import the data, and maybe there’s a new solution. I I’m not I I I guess I won’t preclude saying that maybe there’s a better way to get this data to the markets. There are certainly tools in the marketplace that are starting to look at that and trying to improve that process. Brokers are trying to improve these processes, but I would say, ultimately, ownership has to has to land at the organization level, the organization that’s whose program we’re talking about. Brian, do you have any thoughts on that? I just think that, you know, two brains are better than one. Right? And I think to your point, owning it as a strategic asset, just leveling the playing field a little bit, just again, from risk mitigation, levels the playing field and, you know, is your business. You should own it. You should have those insights at your fingertips. And to your point, I’m not Broker is a valuable, you know, fully integrated partner in this process, but partner means partner, right? Not at their mercy either. That’s right. And I think for too many for too long, and I’ve been in the industry a long time, I think organizations have been at the mercy of brokers. And and some of that was just because there was no other there wasn’t an alternative. They were the best source of the data, historical data, and and the complexity of the process wasn’t as it is today, and that that’s changed significantly. And and even when brokers are kinda managing the process, the amount of and I’m sure everyone on this call can can relate to this, is that you’re just getting that those those questions, those demands. I need this set of data. I need to go deeper here. I need to understand what what what was going on with these particular in this loss history. What’s going on in your out of your in your operations? What does this exposure look like? All those kind of questions. So whether they’re driving it or not, the need for that information is is critical. I think we have time for one more question, do we? I think we do. Alright. Last one. It’s actually the last one on the list. So you gave examples of where AI can already support insurance programs when the data is in good shape. What are some signs that an organization is actually ready to apply AI in a more meaningful versus an experimenting way or or with it? Ryan, you wanna take a shot at that one? Yeah. So I’ll be a little controversial. Right? You’re ready. You you the answer is everyone is ready. There there there really isn’t, you know, what you can get out of it. That’s different. Right? Like, again but are you ready to start using AI in meaningful ways whether it’s ingesting data, asking it questions? I mean, you used some basic examples of just using AI and ChatGPT, which is, I don’t know, like a four dollar subscription. So I’m pretty sure everyone can figure that out. Right? But here’s what I’ll what what I’ll what I’ll tell you is, you know, we’ve done a lot of analysis, and one of the blockers is what I call kind of the technical ego. Like, you you you just said that I, you know, threw a policy into ChatGPT, which is, you know, a multibillion dollar model and improving itself at record speeds. And, you know, it’s funny. I read a document internally yesterday where we were we were taking a position on this, and there’ll be some, you know, a thought piece or something coming out from Origami about it. But it was interesting because it referred to kind of legacy thinking, and I kinda joked because legacy AI thinking is, like, eighteen months ago. Right? So I I kind of laughed. Right? I was like, okay. It’s important. We chuckle about that for a minute. But there was definitely a phase, right, where, you know, AI models were coming out and people thought, well, I had to have AI models on my data or built on my data or sophisticated around my system. And we’ve done a lot of benchmarking on this, and and we went out and we talked with partners who said, I have built AI models specifically on millions of claim records. Great. And then we compared them to the Sonnets and the Clods and, like, all the different billion dollar all the money is going into these models and they’re improving in one week what a private company or a private dataset would take a year to involve. Right? And when you look at the pace of that and then we benchmark accuracy insights, these out of the box models were in the plus ninety two percent range or higher in terms of accuracy. And this purpose built model, this company, this poor company, who built themselves and raised all this money to say, I’m gonna have the world’s best model on claims in this example, was in the seventy percent range. Right? They lost. They don’t even and I feel bad. Like and we we transparently told them that you’re in a really bad place right now, and we gave them all that benchmark data. But I’m saying that to you because when people ask these questions about am I ready for AI? The answer is yes because you can get started. There isn’t this predication on I have to have my own model or I have to have my own thing and, like, tell all the, you know, the technological snobs to kinda sit down for a minute and just get started and get you know? And that’s my biggest advice. Yeah. And I would just add, Ryan, to your point is the data you get from those models, you you always should question it. You shouldn’t you should never take it as is. You need to apply your own thinking. That that’s where the human in the loop comes in in this process. So, you know, does I I I agree with you. You and I actually had a philosophical conversation about that a week or two ago on a different call about, you know, client models and and and the models that are out there. But, you know, question it, look at the data, understand where it’s coming from. Those are the important steps. So I’d add that to the point. So I know we ran over. I’m glad we covered that question as well, and one or two more just popped in. So we’ll we’ll follow-up with with folks that that that had that that asked questions and we didn’t get to. So I appreciate it. Just in closing, I’ll throw up some contact information up there. I wanted to thank Origami for for sponsoring the webinar today and Ryan for joining. I appreciate it. I think it was a great discussion. It was good back and forth. I enjoyed this. Although, I can’t see you the way I have my slides. Haven’t been able to see you the whole time, but this was great to do. We should do it more often. So anyone has any questions for for me or for Ryan or wanna check out what Origami’s doing or how we can help at RedHand, the information’s on the screen. We’d love to we’d love to have that conversation. So, Ryan, thanks again. Stay warm and stay safe. Thanks Pat. Appreciate it. Have a great day. Thanks everybody. Thank you. Thank you.