Martha Dreiling, Co-founder & President: Reserv: Rethinking claims: from ‘boring but brilliant’ AI to real industry change (399)

Hello, welcome. Or welcome back to the Insect Podcast if you are a weekly listener. It's all here as always and this week Robin is joined by Martha Drailing, co-founder and president of Reserve. What if the most valuable use of AI in claims? Isn't the flashy stuff everyone's talking about? In this episode to be explored how one of the fastest growing players in the TPA space is quietly rethinking claims from the ground up, using data, technology and a very different approach to risk. Martha Drailing has spent her career at the intersection of fintech and insurance and now leads Reserve's commercial strategy as it scales rapidly, challenging legacy models along the way. So what actually changes when claims data starts feeding underwriting in real time? Why might efficiency alone no longer be enough and are we about to see a clear divide between the insurers who are still planning and those who are already executing? Poyles off a coffee and enjoy the conversation. Well on this week's Insect Podcast I'm joined by Martha Drailing. She is the co-founder and president of Reserve. Martha, thanks for joining me. Thanks for having me Robin. I'm really excited to be here. Well so, because if I'm entirely honest, I'm a bit bored of talking to CJ. He's got non-planetable name and never trust a man with no syllables in his surname. I think there's a good rule to go by. So we don't know you quite as well, but we'd like to just tell us a little bit about your current role at Reserve. I know you're president, but what's the sort of day to day job involved? Yeah, so I own the kind of commercial strategy for Reserve. So day to day I'm talking to our customers. I'm learning about where we're adding value to their insurance offering and then also where we can add even more value to their offering. What were you doing before Reserve? So I didn't start an insurance. I actually started my career at a FinTech company in New York City called Deccapital. On Deccat the time was this kind of pioneer in online lending to small businesses and we used traditional and non-traditional data sets to help expand access to capital for mainstream businesses in the US. I was at on to for eight, eight and a half years and I fell in love with two things during my time there. The first was just startup life. There's nothing more thrilling than building a company. I'm knowing your customers really deeply and just fixing a problem day in and day out. And then the second thing that I just really loved about on deck and it's really been the three point for my career is I love thinking about how you can use data and technology to make better risk-based decisions. I've spent my entire career thinking about risk. How do you manage it more effectively? How do you build technology that actually supports good risk judgment rather than just automating a bad process? And even at reserve in the beginning, we're figuring it out as we went but we knew the market, we knew the people and we knew the problem because we'd all built some companies beforehand. That's fascinating. But that makes this question. If you were visiting a Fintech and CJ was visiting Snapchat, how do you come together to found a company? Yeah, reserve has a super interesting origin story. So we were a venture incubation from being capital ventures and all-time ventures. And after my time at on deck, I had entered the insure tech space and worked at a couple of high growth MGS. But I knew the folks at being capital ventures exceptionally well. In CJ, he had spent his entire career in claims technology. Snapchat was the pioneer in virtual audio inspections. And then they eventually built a really great claims platform. So he had spent his career in claims technology. I had spent mine between Fintech and insure tech, but always thinking about data and risk management. And then the two of us kind of joined forces in this venture incubation. But what's always interesting, I think, about us is we're co-founders, but we're not the classic two kids and a garage who went to college together. But the whole group behind reserve was just a set of really seasoned industry professionals with deep context. And everybody was just genuinely passionate about fixing this one broken part of insurance. Actually, many of our seed investors at reserve were insurance carriers themselves who really believed that the market needed a tech enabled alternative to the legacy TPA world. So that was four years ago you came together. How is it different from what you thought it was going to be? That's better put, after four years of doing this, what wisdom have you acquired? What have you learned that is going to stand you in good stead going forward? We've grown really quickly. We're over 700 people now in just shy of four years. And I think there's a lot of lessons along the way that many entrepreneurs have when they're building businesses. The first is scale, pressure tests, all of your operational process. It actually exposes very quickly what technology matters and what does it and where you should be investing your time. And the market also shifted along the way. In the beginning AI technology was like a really nice to have. It was the early days actually pre the release of GPT. We were working on our own variants of AI. But now the kind of speed of that technology has just really quickened to a pace that is inspiring, but also makes you rethink on a pretty consistent basis everything that you're doing. And one of the things that I think we've figured out along the way is using AI in the claims process. There's some flashy applications that kind of everyone wants to build. But a lot of the real value that we found is like the quiet and consistent stuff. It's AI that's reading claim files every day and flagging anything that needs attention. Whether it's like coverage gaps, misdeed lines, reserves that look off. Is this kind of continuous monitoring? It's not necessarily magical, but it's exceptionally powerful. Along the way we've also I think learned to be skeptical of our own enthusiasm at times. AI is moving fast. It's really tempting to chase everything. But we've had to build I think discipline and ruthlessly prioritize the ones that we think are going to add real value to the process instead of the ones that just like sound good on a spreadsheet or on a PowerPoint. And I actually think being an end-to-end service provider is almost one of the best ways to do this because you eat your own food in the sense that you're the consumer of your own technology in many respects. So you really figure out is this helping the team or is this not? I've said this before, but we wanted to invent last year called boring, but brilliant just to highlight all that really cool stuff that did boring, but vital things rather than what we normally do, which is exactly the opposite. We're just talking about unbelievably cool stuff that really probably doesn't make a difference at all. It's just in the nature of being an innovator, I think you always want to look at the shiny stuff, not the boring, independent, but with this new tool set, which as you rightly say is changing unbelievably fast, you must be building something because you've always been a sort of market leader in this stuff. If you put what you know now and you put the tech that's available, you can probably do something incredibly cool. Are you doing anything cool that you feel able to share in the public domain? Yeah, we're doing a bunch of different things, but I think the cool element about it is how it all works together seamlessly. So if I were to zoom out, there's a couple of things that we obsess about every single day. And then, how can we bring efficiency to the claims experience? This is like a classic thing that cheap claims officers have chased for years. You monitor cycle times, you look at key loads, you look at kind of small improvements, wherever you can find it. That stuff matters. We still chase it. And AI is really just really accelerated what we can do there. But the second thing that we're really starting to chase right now is efficiency alone, it's not enough. The new standard has to be efficiency and quality together. And when you think about it, like that's just a fundamentally different technology challenge. You can automate a lot of claims in a very risky way. And you need to be able to bring efficiency, not blow up indemnity spend, not deliver kind of poor experience along the way. The other stuff that we've been working on is very infrastructural. Our ambition is to solve some of the really key problems that industries had for a really long time. And one of those is migrating data from systems and ingesting data. So reserve has technology that helps us lift and shift claims files out of one system and into another. Insurance carriers, I think a lot of them operate more claim systems than maybe they're willing to admit publicly because they have these kind of legacy systems that showed up from when you bought another company. And being able to pair it down and to move claims into a single system has just been a long standing challenge. And even in the TBA space, MGA's are carriers who maybe have switched TBAs a couple of times for kind of whatever reason. They might have data sitting inside of various companies. that they've never been able to centralize together. And we spend a lot of time thinking about this, and we've migrated technology from various platforms, from other legacy TBAs, from internal systems that people had built to even guide wire. So we're spending a lot of time thinking about that, and then also thinking about how we can ingest people's data, whether it's border over ports, or other data feeds that they might have and normalize it. So you get a one-stop shop to come look at all of your claims information. - It's long been an aspiration, as you say. We're busy planning an MGA event. For September, we're starting to ask around what are the big themes everybody wants to talk about. And time again, the siloing of data, the fact that they can't get to everything from one place, the fact that there's no sense of any real time or digital or dynamic data sets that inform from one system to another comes up time and time again. Look, I hear all that, and I'm on my reply, which is all terribly techy, but converting that into what it really means for a claims handler, what differences it makes? Some big US legacy insurer with 20 claim systems from a massive acquisition spree in the last 25 years. Tell me what they now have that they didn't have that would fundamentally change the way they can do their job. - If I were to think back to where a lot of the really exciting innovation and insurance has happened in the last five to seven years, I would say the tooling that underwriters and actuaries have today is just dramatically different. And they're able to look at more information, use it intelligently to adjust where their market appetite is gonna be, adjust where the price points are gonna be. But one of the things that's never really caught up to their capabilities is being able to feed your actual claims data into it. And I think what's happening right now is actually a really exciting time in the market is the claims files just have an exceptional amount of really powerful information for insurance carriers. When you're repairing something, whatever the claim was caused by, you get a lot of information about what's the cost of a windshield? If it's a livestock book of business, like what is the cost of a cow? You also have a lot of information about what caused the actual thing to happen or how do injuries develop over time in under certain conditions. And now you can tap into this really rich data set that's existed for a really long period of time that there was just never really the tooling to map the claims information in a wide scale to what the underwriters and actuaries need and want. But now it all feels like it's unlocking at the right time in just an accelerating pace. - I can't see who wouldn't want that. But what does real adoption look like? In other words, I'll people told you that I'm saying, "I, he you've got this great thing, how quickly can I have it?" Or is there still healthy skepticism? Because I've been talking about this quite a long time and it's been an aspiration. And then I wonder whether there's skepticism, whether the enthusiasm overrides the skepticism? - I think day to day, I don't hear people spending as much time being skeptical. Four years ago, I think there was a lot more, right? That was pre-GBT. But now consumers are actually interacting with this technology data. I hate people's kids are using it. And as you start to use it yourself in your personal life and find these applications, you begin to understand that it's real. It can do really powerful things. So I think that kind of skepticism is diminishing. But what I would say is still happening are people are still in the strategizing phase. Like how am I going to use it? How is it going to transform my company? And what I think we're about to see is a real separation of people who are still strategizing and people who are executing on it. And they're executing on the second version, the third version, the fourth version of what they wanted. And that is going to kind of start to separate the winners and the losers, I think, in the market. And you can see really big announcements coming out from major insurance carriers. In December, Chubb announced that they want 85% of all of their underwriting and claims workflows to be automated. They're trying to shave one point five points off their combined ratio and really investing in this kind of technology. So if people are still in the PowerPoint phase or talking to consultants, the clock is ticking. In that sort of strategizing industry doesn't like a good PowerPoint. There's a classic arguments that we've heard, or the classic benefits case we've heard so much, like operational efficiency, giving you all underwriters and claims handlers more time, the ability to develop more business and underwrite better and so on. But if we're starting to look ahead, do you think that this spawns new commercial models? Have we got to start thinking beyond operational efficiency and enhanced underwriting to more fundamental changes in the way insurance model works? I think so. I think you'll see it in a few places. One is just the kind of speed people make changes to what they're doing. The feedback loop of having real time claims data can really speed and quick and a bunch of cycles. You can understand your severity trends faster, your frequency patterns faster, your emerging loss characteristics faster. And if you're feeding that back directly to your underwriting decisions, that gets you closer and closer to dynamic pricing, faster loss, pick adjustments, portfolio level insights, for kind of my part of the ecosystem in a TPA space. I absolutely do think it's going to change and is changing the commercial model for a lot of legacy TPAs or legacy long term TPA arrangements. You pay your TPA by time and expense, like you pay attorneys or accountants. But if you think about it, claims never age well. They always get more expensive, the longer they are open. And when you're paying time and expense, your interest and the TPA's interests are not aligned. Their economic incentive is for it to take longer. Your economic incentive is to get it handled more quickly. Efficiency actually hurts the margin profile of anybody who's operating in this cost plus time and expense way. And I think what's happening now is really exposing that. And so you can argue, I think, whether or not AI really collapses the cost of TPA services. But I do think it just fundamentally changes how people want to pay for it. They don't want to pay for inefficiency. And that is going to be a pretty big shift. And the same thing applies across professional services more generally and software development. You can see anything that was time and materials is going to need to rethink. How about your own commercial model? You've spent a long time now. The four years have been spent working out what's good and what's bad, what makes a difference, what doesn't make a difference. But you're in, that's the space you're in now. Is there an argument that says the more that becomes available, the more confidence people have in it that they won't need a TPA at all because they'll say, we can do this stuff ourselves now. We have to do this copy of all that model stuff that is built. Yeah, I think it's the right question to ask. But the way I think about it just as a builder of companies is just because you can build something doesn't mean you should. So as an example, Cloud Code can build you a CRM. That doesn't necessarily mean that the vast of use of your time is ripping and replacing Salesforce. I think the same logic works here. Cloud large carriers and MGAs build claims technology in-house. Absolutely. It's a pain. There's different licensing. It's different platforms. You need to hire a different talent. And it may not be where they think they're going to maximize their value as a company. Should you be building more products? Should you be thinking about your risk allocation differently and more intelligently? Or should you be working on this? The other thing is there's like a practical reality for large carriers that hasn't changed. They're running on claims infrastructure. That's probably 20 to 30 years old. It's really brittle systems. And you can't exactly bolt on modern technology. And think you're going to get outstanding results. The tech debt alone is sometimes a year's long process. And then lastly, I would say there's still cultural element that the entire industry is going to have to work through. Claims operations are deeply workflow driven and getting any sort of large organization to change how they operate. It's just genuinely hard. People have been doing the same thing for decades. And you need to be able to change their patterns. change where they go to look for things in different systems. That's just very challenging. And AI can be really transformative, but this kind of gets back to how are you going to use it? Where are you going to invest your time? What makes the most sense for you in your business? I think I'll do another one. This is a question as much as an observation. But you're free of legacy. I don't think you're going to have a legacy that's only for years old. And then you've got a culture which enables you to explore endlessly, what's out there. Add it if you think it helps. You can be nimble, the tech is easy to use, it's easy to implement, connect. You just have a flexibility and the ability to move, upgrade, improve relentlessly. Isn't that going to be an absolute fundamental requirement in a world that is going to change faster and faster? I think absolutely. It's the kind of culture that comes from building. And when you work in a startup, we're 700 people now. I don't know how much longer we can keep calling ourselves a startup. But when you have the startup mentality of question everything, build fast, test, iterate, feedback loops, those are the characteristics of what makes great companies. And actually when you think about the companies that everybody admires today, they still culturally have those elements. Amazon has that kind of thinking and mentality. Apple has that thinking and mentality. And just the best companies in the world still operate with a lot of the principles that your classic VC backed startup has. So that raises another question, which is if you really want to do this, you get to a point where scale matters and you need a lot of R&D. And you can see it happening in a gender difference, in the gentics based amount of money that the biggest companies in the world, or most valuable companies in the world, I'll put it into this. Are there threats long term with those sort of budgets? My answer to my question, this comes up regularly in insurance. Amazon's going to come and eat our lunch and they never really have. Do you sit there and think, where does this all end? I need to keep an eye on what big tech companies are going to do in this space with what they're investing in. I think everybody, even ourselves, needs to constantly ask themselves, are we building the right things right now? We think about it all the time. You never want to get to a point of complacency about what you're doing. And so, yeah, we do think about it. But I think being AI native, that's a buzzword these days, being AI native isn't just about having AI tools. It's having the kind of right culture. It's the architecture of your data. It's the design of your workflows. It's the testing and iteration and all of that has to be built. I think with AI and mine from the start and some of the legacy people in my space, they're exceptionally smart and talented people, but it's really hard to change the architecture that your data lives in or the workflows that people have followed for two decades. And I think where reserve is at is we're not in a technology transformation phase. We're in a human capital transformation phase. We're thinking about how our adjusters should be interacting with these tools. How does it change the metrics that we track? How does it change? How claims move through the process? And that's a different set of problems than thinking about is my email infrastructure connected to my claims platform. It's just a different set of challenges. I hope you haven't scared people by all this chat about the speed of AI. That's a good way to finish. For all those people who do get spooked by the idea that we're now in some space race that the half world's going to get left behind. Have you got any calming and comforting thoughts that we can end this podcast with? One of the things we talk to our team members about a lot is if you're talking to a TPI. If you're just the average person and you happen to be talking to reserve, something fundamentally bad happened in your day. You had a claims experience. You were in a car accident. Your small business was flooded. There's still this very kind of human element to what we are all doing. And if we at reserve can reduce the cognitive load of our adjusters, the administrative burden of them to help them focus on the uniquely human elements, being empathetic, setting expectations, having a smart negotiation strategy. I think that's going to be positive. The other place where I think this is going to be just at a macro level really helpful is insurance is expensive. It's gotten more and more expensive. It's expensive for consumers. It's expensive for small businesses. And if reserve or the industry at large can improve the claims experience, control cost, and make insurance a little more affordable for everyone, I think that's a mission we're staying focused on. And something to be really proud about what we're doing and how this technology is being employed. It's one of the things that really motivates me and others here at reserve. Well said, for a long time we've been able to pass the cost of our inefficiency onto the policyholder and it's time to stop. And of course we have now the ability to fundamentally become more efficient and to stop spending so much money with the policyholders that give us in return for providing a fundamentally better service. That should be all of our missions really. Well this has been a real pleasure. Thank you very much for joining us. We're next time to CJ could tell him that he's now what I call legacy. We've now found our podcaster from reserve. I've been an admirer of everything you've done over these last four years. I think you're one of those companies that is a market maker in the sense that you are showing everybody the way and there's a responsibility that comes with that but you're doing it very well. So great to hear the story. Thanks for joining us and keep up the good work. Thanks for having. Well if you've made it this far and I'm pretty sure you found that as interesting as I did. The Instech podcast comes out every Sunday morning where we spot like the latest news, leading voices and freshest updates across insurance that you need to know about. If you would like to take part in these conversations head to www.instech.co to find out how you can join our network and be a part of the insurance intelligence for the curious.