Executive Chef vs. Sous Chef: Where Does AI Fit?

[MUSIC] You're listening to the Duo Podcast, Underwriting Evolved. Join us as we go behind the scenes with underwriting leaders building the next generation of underwriting solutions. I'm your host, Adam Lorenzen, and today we're joined by-- Taylor Pickett. --and Nick Kozyak. And you are? I'm Vice President and Actuary leading our novel assumptions team. And I'm an associate actuary on that novel assumptions team. In this episode, we'll dive into how the Amazon effect is changing consumer behavior. What if a spreadsheet could tell you more than a blood test? And how pick a ball can actually lead to better policies? So whether you're a seasoned underwriter or just curious about the future of insurance, let's evolve together as we explore-- --excelerated underwriting. Love it. All right. Let's get started. [MUSIC] So Taylor, Nick, thank you so much for joining us today. This is awesome for the-- like second time we've gotten to do a Duo Duo Podcast, which is awesome for you our second Duo, which is Duo times two. How do you feel about that? Pretty excited, honestly. Oh, that's amazing. Yeah, that's-- Yeah, that's-- --okay, excited. So Taylor, a little birdie told me that you've been here for quite a while. Guilty. Yes. Yeah. So how did this all start? Yeah, so I actually started in our administration department when I was 16. Wow. Which was awesome to get to see where all of that data comes from. That now I and many of our other actuaries use a great place to start. And wow, RGA has been a spectacular journey since then. That's amazing. So what happened after that? Several years ago, about 2016, I was an actuary working on our pricing team and doing what I call real pricing. So I was building access models, which is the software that's used for a lot of the actuary modeling that determines what price to put on the insurance coverage that we offer. And that's my boss at the time approached me and said, "Hey, there's this new opportunity coming up to work with one of our subject matter experts on projects around the future of underwriting and how we evaluate and approach those from an actuarial perspective." Whoa. And sounded pretty exciting and I talked with her about it. And really interesting opportunity to work on some new tools that RGA was helping to bring it to the market and also the product and underwriting ramifications that those would have. And it has been history since then, nine years that I've been part of that team or the different forms as it's evolved. And it has been beyond exciting to be on the cutting edge of what the industry is doing and to be able to work so closely with our underwriters as part of that transformation. And it's very exciting to see how far the industry has come even in those nine years that I've been part of this. And it's incredible to think about how much potential we still have to make the future even more successful. I mean, that's fantastic. I think it's interesting though, what were we doing before real pricing? Was it just was it was a joke? What was it? No, so I should have been more specific. So I was doing real pricing at that time. And I still tell people that I work in pricing because I do work very closely with our pricing actuaries, but I'm not building the models and solving for the rates anymore, but our team is often an input to that process. So insurance is a unique product if you think about it that if Apple makes an iPhone, they know exactly probably down to the penny. What the cost of that iPhone is for them to produce. What is the cost of a life insurance contract that we sell? We don't know. It depends on a lot of things when how long is this person going to survive? Will they keep the policy that long? Will they lapse? What will happen? So because of those unknowns, we need assumptions. And that's where our team can apply. I can't say. Okay. All right. And then Nick, what about you? How'd you end up here? Yeah, I think the first role that I had here at RGA really set the tone. I learned that I really enjoyed the data side of things and sort of looking into how the different aspects of a person's either credit history or medical history really plays into how we view their mortality. After that, there was really a transition that I found in my career that I really wanted to get more in touch with the clients, get more in touch with the industry. And this role really opened up the opportunity for that. And I was lucky enough to be able to join this team about a year ago. Oh, wow. Okay. Great. So we have two innovators here to talk about like the core of like actuarial science. Let's talk about just that a little bit like from your perspective, how is what you do different than what an underwriter does? We've interviewed a lot of underwriting leaders, data analytics, AI. Can you just give a brief overview of that? Yep. Absolutely. Happy to. So St. Louis, we have the Ted Drew's Christmas trees every year, right? Famous ice cream. Oh, it's miss trees. Yeah. They get their ice cream. They suffer through that in the cold because it tastes so good and they pick out their Christmas tree, right? So imagine the underwriters going through the Christmas tree farm and they're looking at every branch, every needle and evaluating individual trees, right? And the underwriters get those trees organized. This group of trees are similar. That group of trees is similar. And then the actuaries come behind and look at these carefully curated sections of trees. And we say, okay, how do we evaluate this section? What do we do with it? What's the number we put on it? I see. And that's kind of how we look together under the actuaries look at big groups. The underwriters look at a very individual level. And when those two are in sync, it opens up some really great possibilities for the business. And they're out of sync. It creates challenges. Yeah. That sounds pretty very collaborative, but also part of the ecosystem that's needed. So Ted Drew's, by the way, for listeners is in St. Louis. It's a family-owned frozen custard company that was founded by a man in 1929 by the name of Ted Drew's. They every year at Christmas, they throw up a Christmas tree lot and you can get your ice cream, just like Taylor said, and also get a tree while you're there. Nick, being newer, I guess, to RGA, like what are some things that you are seeing develop that, you know, not having the history for, but like seeing forward, being in this innovative space. What are you recognizing from your position? Yeah, I think as of recently, the vast amount of data that's been become available and the techniques to process that data have really opened the door to revolutionizing the way that we evaluate risks in the life insurance industry. Okay. Historically, it's been very detailed. We've been looking at blood urine, measured build, measured blood pressure. It takes weeks to really collect all that data and analyze it and in other writers very involved in looking at those individual risks. But as we get further and further into this data world and how we're able to, in real time, look at a person's medical history from the last, let's say, seven years or so, it opens the door to be able to make those decisions, evaluate those risks without the need for that intrusive measuring of blood. And it's also just a quicker response time. So, wow. I think that's the big push that I see that there's a much more of an emphasis on getting these decisions done faster and still at an acceptable risk level. That's cool. And what do you see as the benefit of that? Like why is faster better, I guess? Faster is better from the consumer perspective because you're catching them when they've already made that decision that they want to purchase. So, the quicker response allows you to make that sale while they're still in the mindset of they want to buy this life insurance policy. It's also important because I think a lot of people don't know the cost of life insurance at this stage. Yeah. And to think about going through an eight-week process just to figure out what that price could be, that doesn't seem worth it. But if you can just go and fill out an application and find out in a day or two what your price would be, that might unlock a whole new market that we've been under ensuring because they just don't know what the price is. They just assume it's way too high for them to afford. Also, in a way, it's like going from like a custom design shop to just pre-made things on the shelf in a way where I'm walking into the store, I see what I like, the price is there, I can grab it, versus, hey, give us a quote, we'll come back to you eight weeks, our guys will look at it. Taylor, does that resonate with you? It does. I think that in today's world, consumer expectations are so different than they were 30 or 40 years ago. So 30, 40 years ago, if we wanted to buy something, we went to the shopping mall. Yeah. We didn't want to buy something, we went to the shopping mall, it was something to do on a Saturday, right? And now I can sit on my couch and click a button on my phone and I used to say in two days, but now even faster than that. Amazon has something sitting on my front porch for me. So when that is a process that the consumer has gotten accustomed to, it can probably feel pretty incongruous to say, wait a minute, you mean I have to go give blood for this and I have to wait four weeks to get a decision on whether it's going to work out and what price I'm going to pay. And that that could feel pretty out of step with consumer expectations in our day and age. And I think Nick raised an important point to that there's an opportunity to maybe dispel some myths or some misconceptions about the price of life insurance coverage. So lemma publishes their insurance parameter. And consistently it shows that Americans overestimate the cost of life insurance by a pretty significant margin. Is there any reason rationale behind that? I think it's maybe just lack of familiarity with the product. I think that for a lot of people it may not be something that they think about until something happens in their life that they feel a need for it or they become interested and then they might start looking into it. So I think it's more just what insurance products are we accustomed to? People have car insurance starting from what 16, 17 years old. They may pay for insurance on their cell phone. They have health insurance. They have. home insurance or renters insurance, and a lot of those premiums look very different in terms of how much am I paying to the insurance company every month for an amount of coverage that I have. That relationship looks very different for those products than it does for life insurance coverage. So the Amazon effect is really this idea that people are getting used to the idea of speed and more transparent or clear pricing. Are you noticing this like generationally? Is there a different way buying if you're 50 versus if you're 30 like Nick do you have any thoughts about that? I don't know if there's necessarily a preference that I've seen show itself, but the big thing about accelerated underwriting and how we're tackling that now is that we are targeting a lot of the younger age individuals and those are the people that are getting the most benefit from accelerated underwriting at this stage. Okay. That's cool. Explain the benefit. It comes back to that speed of response. It's that ease of going through the process. It really does just make the whole life insurance buying process so much simpler so much easier. I haven't really thought of this, but getting these out is more like like I'm thinking Amazon effect, right? It's like a website. We think of websites. We think of kind of this non-human interaction to buy things. Like I could buy a boat. I could probably buy a plane now, right? And not have to talk to somebody. But what is the impact of this accelerated underwriting on like an agent who's out there and who's used to this, you know, potential three to eight week turnaround time for for a policy? I think that there are tremendous benefits for agents. Interestingly, agent adoption has maybe not been as fast or strong or universal as I might have expected, which is which has been interesting. So benefits to the agent, they get through the underwriting process faster. They have a policy and their customers hands more quickly that can have benefits in terms of acceptance of the policy like we talked about when the person's in the frame of mind to buy your life insurance. If we can get the offer to them at that time, oh, there's a benefit in terms of take-up rates. Okay. Yeah. So that's very beneficial for the agents. Something that surprised me is hearing from the agent perspective that there could sometimes be resistance from giving up control of the process. So the agent may have been accustomed to they would put in the order for whatever lab tests if medical records were needed from the person's doctor, they would put in that request, they would ask the questions on the application, you know, what's your family history, what's your personal history, all those sorts of things. And I've heard that at some organizations, agents have been a bit reluctant to give up that measure of control that they had in the process. Something I hadn't thought about until several years ago, I heard a lady from a marketing organization, a direct carrier speaking at a conference. And she mentioned that one of the reasons that their agents weren't thrilled about tele interviews, where from a third-party company an interviewer would call the customer to ask them all of these questions, you know, to have your head cancer, have your head heart disease, all of these conditions, you're asking about some really sensitive information on that life insurance application. And that their agents were nervous about giving up that control because, you know, they have put in the effort and the investment and time to sell this person on life insurance, explain the benefits to them. And if a sensitive question is not asked in an empathetic understanding way, that could totally upset the apple cart on that sale. And that's a very relational type of sale. It's not a quick transaction. It's literally, tell me about you, tell me about your life, why are you doing this? And then now we're kind of chopping it up and seeing these efficiency gains, but it's coming a little bit of that cost of that personal relationship. Life insurance is a very personal product. We're asking people to think about their own mortality. That's pretty sobering. Yeah. You know, I don't think any of us wake up in the morning and think, today's the day I'm going to plan for the time that I die. Yeah. That's, that doesn't exactly get you out of bed with a spring in your step. No. So having that empathy and that personal touch, I think, is really important. So thinking about what are the elements of the process that really lend themselves to more automation and to using technology to help us at the same time. It's important to understand what are the unique capabilities and skills that a human offers in this process and making sure that those remain in place. I think that gives us the best of both worlds as we move forward. And so it's interesting with the example I mentioned about the interviews, I think from my perspective and maybe from others in the industry, we were thinking, Hey, you should love this. We're making this process much faster for you. It's easier. You know, you get faster sales. This should be great. And those aspects are, I think it's also important for us to be open to feedback on how can we do this in a way that still preserves those personal touches where they're most valuable. To, in my opinion, that's something that RGA really focuses on, even here within our company walls, if you will, that we focus on the people. Having that personal touch shouldn't be lost for the operational efficiency. And that's something that I think is something to keep in mind. It reminds me also of AI. So AI is this big behemoth of an idea that's being put on us. And it really is causing us to re-evaluate how we work. And so are there any things that you guys are seeing that that AI is impacting your area? Yeah, I would say that AI should first off just be looked at as a tool to help us not to replace us, especially in the underwriting context where human intuition and knowledge is so important. We've seen a lot of cases where AI just can't necessarily pick up the nuances that a human would be able to pick up on. And as it evolves, it might change. But in current state, we do see a lot of opportunities to use AI in the way of summarizing things. So electronic health records are a perfect example where we have all this unstructured data, notes, handwritten notes, and things like that. AI is extremely powerful in the way that it could take all of that pages worth of unstructured information and boil it down to just a few bullet points, pull out the key aspects of a person's medical history. Even answer questions that an underwriter might ask of that person's medical history. And I think that's what really opens the door to AI assisted underwriting where we still have human underwriters leading the charge, but they have an extra tool in their pocket. I see. And is it impacting the actuary space as well? I think that it's beginning to. It's interesting. I think it's important, as Nick said, to think about this as a tool that can help us. And there are a lot of things that it can do. But to prioritize, where can this really make an impact for us today? And I think that some of that is exactly what you said, Nick, those tasks that don't require as much judgment. They don't require as much intuition. Okay. And domain expertise, the summarization, taking away some of those more menial tasks that really don't require that higher level thinking. That may be the low hanging fruit for us to explore first. And you know, as I've talked with the last few years, our underwriters about some of the ways that these new technologies may be coming to bear to help them work more efficiently and work faster. I've kind of thought of it in terms of if you look at the professional kitchen in a restaurant, you've got the executive chef that's tasting everything for flavor. He's making sure the seasoning is right. But you also have a sous chef that's doing the slicing, dicing, grading the cheese, all of those things. That's where I see AI really being beneficial right now. I think that we, the professionals, are still the executive chef, but AI is a strong up and comeer in the department of the sous chef. Oh, that's a great analogy. I love that. It opens up these kinds of questions. I think that that now when we're going faster and we're changing the relationship with agents and really just with purchasing in general, what do you see as the future as it stands right now with the technologies and the processes? What are you seeing as like kind of like three to five years out? Yeah, as much as I want to say that the accelerated underwriting might take over in the next three to five years, just because of this speed and efficiency, I don't see that being the actual case. I know there's a lot of people out there that still want to go through the blood testing, the urine testing. They want to get the most accurate answer to their risk. But I think that the future for us holds a lot of opportunity in the way that we look at risks, the way that we supplement those blood tests, urine tests, or the absence of them, the data that's available. I think that's really the hallmark of what's going to be changing over the next three to five years. Okay. Okay. Taylor, what about you? Yeah, something that comes to mind for me is as long as I have been in the industry. We've been told electronic health records are just around the corner. Yeah. And gosh, that's been the longest hallway, right? It's been over a decade now. They've been around the corner way over a decade. But I do think that we are absolutely approaching that corner and technology is the key in getting us there faster. So I think we all recognize, certainly if you talk to our friends in underwriting, there's so much value in medical records to understand the treatment plan that a person may be going through the conditions that they're that their physician or their care team are trying to address or maybe even prevent. And that is available to us in electronic health records. Right now, the process is still fairly manual in reviewing those. It requires an underwriter's time. Oh, yeah. And that can slow down the process. It also makes it more challenging to scale, right? Because we're using humans there. So as tools advance to be able to turn that valuable unstructured information in the electronic health records into structured data, yeah, that opens the door to automation. And so I think when we see that happen, we know that electronic health records will continue to become more widely available as more and more networks across the country start to have that data available in a digital format and I think that the key to taking that next step is going to be when we get the technology that turns all of that into structured data that's machine readable that we can automate that opens the door to scale and even greater efficiencies. That's true. And we just actually did our interview with Eric Westis, who is one of our data scientists and he was explaining that we have internal tools that grab structured data, try to pull out impairments, try to pull out key data points about that and the process of going through that is fascinating, but it is changing how underwriters review and approach and even the data that we personally collect, it's making me think though that this volume of data that's getting ready to hit us is a bit unprecedented, but it feels like that will slow things down. Is that a way of thinking about that? Maybe the amount of medical information that carriers are sending us is increasing by 10x every year, right? Or 10% I should say. And that I think would slow things down in terms of assessment, what were you thought about that? Yeah, I think that that's where technology really keeps up with that pace. Technology is what enables us to first off look at the data faster, we're able to pull from databases faster nowadays. We're able to use AI to pair that information down, pick out the most important bits of information. Okay. So I think that as the data grows, so will the technology and we'll always find a way to make that approachable to for our purposes? Okay, that's very comforting. Taylor, what are your thoughts on that? Yeah, I agree with what Nick said. The other piece that I would add is when we think about the value of a fast decision, I think it's natural for us to think about getting to yes faster. Clearly that has tremendous value. I would say that when we think about an insurance context and an underwriting context, there's also value in making a no happen faster. So that's not necessarily saying that our answer is no more often, but it's saying if it's going to be an eventual no, can we pull that forward in the process and have it happen faster with fewer human eyes and fewer man hours on it? So as we start to get more medical data coming into us, particularly if it's structured, we can have processes in place to look for heavy hitters, big ticket items to say, oh, you know what, stop the presses, this isn't going to be a good fit for this product offering. If we can get that faster, it helps us prioritize our underwriters time on the cases that may have a chance of getting to yes and get faster decisions on those as well. So as we speed up, clear cut knows, it gives our underwriters more time to get to cases faster that might be a yes. So even if they still require manual review, if they get to our underwriters' desks more quickly, our underwriters aren't going through as many of the cases that aren't going to be a good fit, even those manual offers start to become faster. If we're going to leave people with kind of some lasting thoughts, right, we'd like to give some personal advice, something that's that's a little pointed. If there's anybody out there listening that is an actuary or in that field in adjacent, like what advice do you have for them right now? So I'm going to take it from a little bit more of a personal interaction side. I know that Nick is so expert and up to speed on all things going on with technology. So green pastures there for sure, I'll leave that to him. But my advice would be make friends with your underwriters. They are awesome people and nothing bad will happen. There's such a benefit to have actuaries and underwriters talking more closely. So I've heard the anecdote shared before that Elton John, and I forget the name of the lyricist that he worked with on a lot of his music. But that apparently some of their larger hits, they would work on independently. And the lyricist would literally write the lyrics and mail them to Elton John. And Elton John would sit down with the piano and work out music to them. Oh, that's awesome. Now that was clearly very successful. It worked really well. But how much more could they have done if they were in the same room together working on some of this? So I think in the past, actuaries and underwriters may have had a little bit more of that separation. It's sort of forced that issue. But I think as we've started to work together more, we've realized, wow, there are a lot of benefits here. And I will say for my team, some of our closest collaborators are on the underwriting R&D team. I know in a recent episode you spoke with Fizzelia, the executive director of underwriting research and development. We work with that team super frequently. And having those varied perspectives as we address some of these challenges and opportunities at the industry faces, it gets us to better solutions. So make friends with your underwriters. You'll learn something. You'll make some awesome new friends. And I think it'll help all of us achieve greater things. That's great advice. We had a underwriter summit and we brought all of our underwriters in, at least most of them, and we went and played pickleball. And I can tell you, they're pretty good at pickleball number one. I don't know how. Maybe that's an underwriter thing. But also the advice that they were giving even from a non-technical perspective was very illuminating and helped me really understand their perspective. And so I would second that advice. Nick, what about you? Yeah, I think Taylor set it up pretty well. My eyes always on the future. So I would definitely advise an actuary right now just to look to the future. If the more you understand about what's coming, the better you can be prepared for that next step. And the skills that are needed in the actual world are going to be changing. They're already changing. I think a lot of early career you might have spent time writing macros or VBA code. But now we're talking more about maybe fine-tuning AI models or building AI models. So I think the opportunities are going to change. But the knowledge that you generate and accumulate through school and through the exam process is still going to be very relevant. It's just how you use that knowledge and what skills you're using alongside it that's going to be changing. So just keep your eye on the future and keep anticipating what's coming next. I love that. And if people have questions about how to improve their maybe actuary operations or innovation, what would you offer to that? The biggest thing in my mind is ask yourself if you are getting all the value you can out of the data coming from your underwriting shop and your underwriting operations. So underwriting is a much more varied path today than it used to be. If you think about the way underwriting happened in the mid-90s and even before that, pretty much every policy over $100,000 sold in the US had blood testing at the time of underwriting. There was a pretty uniform life insurance lab panel that was run some variation from one carrier to another. But that was pretty uniform. We may add more to that at higher face amounts, but that was almost across the board until accelerated underwriting started to come into industry about 10 years ago. And now you have some policies where issues of blood testing at the time of underwriting, some weren't, some had this data, some had that data. As it becomes less uniform, it's important for us to understand the underwriting that happened with each policy so that we can later see the impact of that when we look back at the performance of the business. Oh, I see. So it's kind of like when you see these renovation shows on HGTV and I always talk with people about, you know, what is the number you multiply the budget by that they give you? It's at least two or three, I think. You're not going to go to Home Depot and find things at price or contractor to do it, but I digress. And people are always so crestfallen, understandably, when the contractor comes and says, we found this electrical issue, we're going to have to rewire the house. Nobody's excited to spend money on that kind of thing. And I realize it's also probably unexciting at first blush to say we're going to spend money making sure we have this great coding and data capture coming out of our underwriting department. However, as unhappy as those homeowners are to be spending money on that rewiring, they're very happy when they walk into their bedroom and turn the light switch on and they don't get shocked. True. So if you do this now, you will reap the benefits from it in the future. And I think it creates a virtuous cycle where the more we learn about how these things are performing, it helps us be more targeted and more strategic in the ways that we refine and improve them going forward. Okay. I like that. Do you have any specific advice for someone in your shoes? Yeah, I think the thing I want to point out is to look at the industry, see what other actual teams are doing, what other actuaries are doing and take some notes from how they're approaching different challenges that might come their way or what skills they're building. I think that there's a lot of insights that we can glean from just seeing what others are doing and then ignore that and think about what you could be doing. Maybe think outside the box and don't be forced to stay within the confines of what the industry is doing. Oh, that's cool. That's good advice. So, look at what everybody's doing, but also sort of like look at the white space of where they're not playing in and go and investigate that. Yeah, that's where innovation happens is take that next step. I like that. Okay, well, this has been fantastic. Thank you both so much for being here today and we'll let you know, you know, when it goes live, so you can hear yourself on the podcast. Awesome. Thanks. Thank you. Well, that's all for today. Thanks so much for tuning in and being a part of our community. Hopefully you had a great time and maybe learn something new. Remember, underwriting is evolving and so are we. So if you have any suggestions or ideas for future topics, drop us a line through our website or LinkedIn channel in the show notes. We'd love to hear from you. Until next time, keep evolving, keep learning and keep automating. We'll catch you in the next episode of the Duo Podcast. Take care.