Webinar replay: How carriers are leveraging large language models (LLMs) and automation to drive better decisions
Watch Now
  Everest Group IDP
             PEAK Matrix® 2022  
Indico Named as Major Contender and Star Performer in Everest Group's PEAK Matrix® for Intelligent Document Processing (IDP)
Access the Report

Interview with Automation Center of Excellence expert Steven Weiss

Christopher M. Wells, Ph. D., Indico VP of Research and Development, talks with Steven Weiss, former Senior Vice President, and Chief Underwriting Officer at Munich Re Specialty Group Insurance Service, now on garden leave, in episode 8 of Unstructured Unlocked. Tune in to discover how enterprise data and automation leaders are solving their most complex unstructured data challenges.

Listen to the full podcast here: Unstructured Unlocked episode 8 with Steven Weiss

 

Christopher Wells: Hi, and welcome to another episode of Unstructured Unlocked. I’m your host, Chris Wells, VP of R and D at Indico Data. I am more than pleased to be joined today by Steven Weiss, former Senior Vice President and Chief Underwriting Officer at Munich Re Specialty Group Insurance Services and now on garden leave. Looking for that next opportunity next year. So, Steven, welcome.

Steven Weiss: Thank you, Chris. And thanks for the little pitch there for me. I appreciate that.

CW: Of course. Well, thanks for joining the podcast. As we get started here, why don’t you tell us a little bit about your background and how you came to be a chief underwriting officer.

SW: Sure. It’s been a journey, and you know, I spent five years active in the US Navy and three years in the reserves. And I knew how to work in leadership capability. I then went into consulting for the energy and marine insurance industry. Spent 15 years doing that. Both doing marine survey work and loss adjusting. Culminating for five years at my own company back in 99 to 2005 or 2004. One of my largest clients, Liberty International Underwriters at the time, hired me in-house to run their risk engineering team in 2004. I hired seven surveyors across the country across, actually across North and South America.

And then I started taking over more and more underwriting duties. Eventually opened up two offices for Liberty in Latin America in San Paulo and Miami for the Marine, for the marine groups down there. Moved from Liberty to Aspen to run two product lines. I ran the General Cargo and the project cargo product lines over in Aspen. And then in 2018 started with Mina Cree as Chief Underwriting Officer for Ocean and Inland Marine. So I had under, under my purview was all the various inland marine lines, which is like 52 different lines of business, including construction contractors, equipment, stuff like that. And then I also ran the Ocean team, which mostly focused on cargo boats, cargo and logistics, and stock throughput and things like that. Project cargo as well. Have been there, was there for just over by the time I finish, I’ll be there just over five years. Finished my gardening leave here in in May of 2023. And you know, over that period of time, I’ve learned by doing, I’ve also gone back and gotten my charter property casualty underwriter, my Associate Marine Insurance Management and then went on to get my master’s in insurance from Boston University, which is a really good program. If any C P C U wants to get a master’s degree, it saves you four courses, if you through Boston University. So there you go.

CW: Ocean Team is a great name for a team, by the way. I love that. And then two, what is a risk engineering team? I’m not an insurance guy, so, yeah. Okay.

SW: That’s a good question.

CW: So I wanna understand what that is.

SW: So many insurance companies, especially the larger ones that are focused on property and casualty with high severity possibilities. It could be frequency as well, but usually a high severity. They hire teams internally to basically mitigate their risk. So I worked on the Ocean team where, you know, we had guys go out and, and load chips and, and look at cargo and, and you know, basically give the lo part of it was lost control. And then part of it was actually doing the engineering, working with the various groups in order to make their risk more acceptable as well as give them the usually they ended up with a better deal from us if they worked with us on the engineering side because we were more, more in control of the risk. It’s really a risk mitigation tool as well as they tool for underwriters to better understand the risk.

CW: Interesting. So they, they let you a little bit deeper under the hood and you know, the risk better, and so the pricing is easier.

SW: That’s what I did in my consulting business was really the risk engineering side of things. Okay. And so it was perfect having an in-house tool like that for an insurance company really gives them a leg up on understanding comp large and complex risks and Exactly right. We go out and do a lot of site visits, a lot of under the hood l looking spend time with the clients in and around there. Whatever it might be. You know, everything from looking at warehouses to project cargo load on ships to construction looking at construction projects. Anything that had a severity or a potential was an excuse, an unusual risk, it was something out of the ordinary.

CW: Okay. Interesting. Final thought on your intro. Thank you for your service in the Navy. Appreciate,

SW: That’s great. Thanks Chris.

CW: Yeah, of course. Most of the audience for unstructured unlocked comes from sort of the automation, artificial intelligence world come from a lot of different verticals, including insurance, but not exclusively. So talk to us a little bit about what it is that a chief underwriting officer does. Like, what are the responsibilities? What’s the organization look like, that kind of stuff? Okay.

SW: So the Chief Underwriting Officer, basically anything that is underwriting directed falls under his purview, his or her purview, cuz there’s, there’s a lot of females that are chief underwriting officers as well. And what it, what it really means is you set the standard and the guidance for your teams as to how they’re going to assess and look at, look at risks and, and then underwrite them from the standpoint of both pricing as well as terms and conditions and the policy. You know, you, you’re the one that assigns different responsibilities to different underwriters. You assign them authority levels; you take away their authority levels if they don’t do what they’re supposed to be doing, you know, so, so really, he’s the gatekeeper between the capacity provider, meaning that the insurance company that actually gives us, cuz we are, we are an m a managing general agency an mga and I, as an mga, we had capacity from our parent company, MUN agree.

But we had to, as the chief underwriting officer, my job was to work with the capacity providers to make sure that they were comfortable with what we did. And if, if we wanted to do something more or something different, then I’d be the one that would be negotiating with the capacity providers. So, okay. Also, if you placed re, you know, reinsurance, we helped to place reinsurance. We were very heavily involved in all aspects, from the underwritings side to the administrative side, and the operations side. We were in the process of a big automation upgrade in our systems. And that was another part that fell under my, you know, I worked with the chief operating officer to make sure that that went well.

CW: Wow. That’s a big umbrella. What, what are, what are some of you know, sub-teams that rolled up to you in that role?

SW: So I’ve had at Munich, it was specifically all the underwriters. So I had, I had a team of 14 different underwriters across the country. And we worked in business in Canada, North America, and Latin America. There was a team of seven or six, sorry, six inland marine underwriters that were focused on construction contractors, equipment electronic data and all these different, all these different things that fall within the inland marine scope of business for an insurer. And then I had a team of eight ocean underwriters that focused on cargo logistics stock throughput, all the different things that basically greased the wheels to keep the supply chain working.

CW: Yeah, okay. Right

SW: On. Involved in supply chain management. We did a lot of supply chain management.

CW: Okay. Interesting. And those you know, that inland Marine team, are they, are they sort of, do they break down along lines of sort of specialties in terms of risk? Or is it geographic, or, or how is that

SW: Organized? It was, it was a little bit of we hadn’t re we had ’em regionally done geographically just because that’s where they lived, but if they, you know, but some of them had a specialty in motor truck cargo, some had a specialty in construction, some in contractors equipment. So they worked together as a team in order to share brains. If they got a risk in there, they were, they, they weren’t as familiar with.

CW: Okay. That’s pretty interesting. Who said insurance was boring? They were wrong.

SW: I’ve never said that. But then again, I work in specialty lines and there’s a you know, it’s, there’s a much greater diversity of what we do than I’m sure in some people’s business.

CW: Okay. Yep, that makes sense. Speaking of, you know, speaking of interesting things, what, what’s like the most exciting part of this job? Like, what got you out of bed in the morning when you were doing this day to day?

SW: Well, I mean, there were, there were two things really. One was the people, you know, not only clients, but also my my staff. You know, I have, I have a lot of very good friends that are also clients. And and then the second part was how do you make something that is a little bit esoteric, i e insurance into something that is, is concrete to your client, and then build them the best program along with your broker partners to really cover the risk that you’re trying to cover. You know, it was really that service part plus the plus the the, the people part. Okay. Those two really were kept me going.

CW: Yeah. Okay. All right. That’s the good. How about the bad? What, what are some of the key challenges that are facing the underwriting, you know, vertical or industry today? What’s hard right now?

SW: A couple, you know, you get back on the flip side, we have a people issue in the insurance industry because we don’t have enough. Okay. We have a very, there’s a very we, we, we missed out in about the 30 to 45 year range. We have, we’re, we’re well below where we need to be because as people my age, and I’ll say just something above that 45, I won’t say exactly how old, but something above that, there’s gonna be a lot of baby boomers that retire. Yeah. And, you know, I’m the last year of the baby boom, I was born the last year of the baby boom. So, you know, 10 years from now, I’m gonna need somebody to take my place and these people are gonna be 20 years younger than me. Yeah. because of the, of the, the, that gap in the insurance space.

So what one of the things that I’ve focused on over the years is training, and not only training my own folks, but also as we mentioned in our, in our our in intro call, I work with different colleges to help them, help them have their people find jobs, as well as go in there and teach some courses that are kind of outside of the box for most of their instructors. You know, one of the, I’m the chairman of the board of a group called the Houston Marine Energy Insurance Conference. And that conference puts on a yearly, or that group puts on a yearly conference in September or October, where we focus on issues in the industry. But the biggest thing that we do is we bring in students from eight different colleges to get them introduced to folks that are in the industry, as well as give them a chance to hear about what’s going on in the industry. And we had over 500, or just about 500 people, at the conference back in October. And we’ve just finalized all of our donations to the schools. We give school scholarships. We’ll be giving away about $120,000 this year from Oh, the money that we made from the conference.

CW: Wow. So, it sounds like you’re having pretty good success trying to rebuild this pipeline.

SW: Yeah. I think the numbers somewhere up around a hundred thousand short people in the industry by, you know, by 2025 or 2027. So it’s gonna take a while, and it’s gonna take a lot of work to make it happen. But the great thing is, I tell the young folks is, you know, you’re gonna have an accelerated career because you’re gonna have to step into higher positions than you would typically at your age because of the lack of people that are older than you, or more experienced than you.

CW:

That’s a pretty exciting pitch.

SW: Honestly, I would think it’d be very exciting. I’m not sure where you fall in, in the spectrum, but there are a lot of millennials, you know, that wanna be the vice president when they walk in the door. I know you’re a vice president, but I tell them is that you know, that’s something you earn, you know, and you gotta spend, you gotta spend some time. You probably spend less time than I did to get that title, but you have to spend enough time in there, to know the business. And you don’t wanna, you don’t want to get ahead of your, your skillset because if you get ahead of your skillset, you’re gonna, you’re gonna flounder and it, it could cause you to have some issues with your career going forward.

CW: Yeah. That’s good advice. I am technically a millennial, although I, identify as the greatest generation

SW: Also. There you go. My son is also a millennial, so I understand it.

CW: Yeah. Very good. It’s interesting. I was, you know, I was in academia up until about 10 years ago, and back then all of the granting agencies were like, we need more people in stem. And I don’t think they were thinking of underwriting, but of course, to be a good underwriter, I assume you have to have a good command of, you know, mathematical principles and certainly

SW: Math, logical math. Yeah. Certainly math. You gotta be able to think logically. And that stem really drives you in engineering too, and that’s where your risk engineering group comes in. Yeah. You know, and a lot of the lines that I work in need people with some technical knowledge of the actual industry.

CW: Yeah. So, I love it when folks like yourself give back to academia, because academics, they’re not an industry. Right. And so, taking that industry knowledge back,

SW: Some people would say that’s not true anymore, but, you know,

CW: It’s changing. Yeah.

SW: You know, there, there are certain colleges that seem like they’re, they’re on, on an industry focus, but yeah. No, I think that you gotta pay it forward that we have. I still remember professors from college that didn’t put me down this path specifically, but they taught me how to learn, and then they gave me the tools to be successful wherever I went.

CW: Yeah. Awesome. Yeah. Pay it forward. I love it. Even in, even in specialty insurance, you gotta pay for it.

SW: That’s absolute. You paid it forward. Yeah.

CW: No, that’s cool. So we, we talked a little bit about some of the skills to get started, and you talked about the need to sort of, you know, you gotta get some time under your belt, get, don’t get out over your skills. What are some of the hallmarks of just a good underwriter and then the hallmarks of a good chief underwriter?

SW: Well, so a good underwriter has, has some sort of technical skills, not necessarily underwriting skills, but I’ll use one of the guys that I, that I, that I hired back about two years ago. He was at, at sea. He was sailing for one of the big US shipping companies. But, he was also gonna be getting married and didn’t necessarily want to spend his whole time away from home. And, you know, I get, I get that cause I’ve traveled so much in my life and spent, you know, some years at sea with the US Navy. And so he came in, he, you know, basically volunteered to come in as an intern. He says I wanna learn the business. And I said, all right, you got three months between shifts. Come on in. And he worked hard and did a great job.

And then at the end of the internship, I didn’t let him go. I hired him didn’t really, didn’t really have a slot, but I hired him because he was a hard worker. He learned fast, he put in the time to learn it. He asked good questions. And it was never shy about asking questions. You know, people think that if they ask a question, they’d look stupid. And it’s totally the opposite. You know, if you don’t ask a question that’s on your mind, then you’re gonna look stupid because you can’t, you can’t get your job done. You know, and, and, and one of the things that I start most of my presentations with, you know, I expect you to ask questions. Interrupt me, raise your hand, talk to me. I said, let’s have a conversation. Like you mentioned before, let’s have a conversation here. I don’t wanna just sit there and talk to you, talk to the students or whoever it is for, you know, two hours or whatever. Let, let’s have a conversation here. Cause it goes much more quickly.

CW: Yeah, yeah. Absolutely. And yeah, and you know, that rich dialogue helps you get below the surface level, you know? Yes. Between, between the lines of the textbook. Right. And that’s, that’s important.

SW: Totally. You know, and, and it’s funny because one of the courses that I I teach, one, I teach one chapter of is the where’s my book?

CW: Oh, they’re real books. I thought it was just a background. That’s a great, yeah.

SW: They’re real books. Yeah. It’s assessing commercial liability risk. Okay. And so chapter eight is exploring marine and aviation coverage. And so I don’t really, you know, I don’t teach the book. I just talk about what I do. Yeah. And you know, and that gives it, and, and great, I have great pictures, all that sort of stuff. So that’s the part that I think it makes, it, makes it alive, makes it exciting. That’s awesome. So when you go from an underwriter to a chief underwriting officer, you know, there’s some, obviously some growth in there. And, you know, you, you, you first manage a team, then you might manage a product, and then you’re gonna go on to, to, to manage a group of products. I think it’s, you know, again, it’s two things. One, you gotta have a skillset, an underwriting mindset.

Yeah. And the second thing is, is you gotta, you gotta hire good people because there’s no way you can run a line of business that has, you know, 25 or 30 different lines in it, or a, a business has that many lines in it by yourself. Yeah. You know, you, you’re not gonna have the skillset across all of that. What you have is a overall understanding of what the goals are of the, so, so you gotta switch from a tactical to a strategic. Ok. Now you gotta go from, you know, your day-to-day to, okay, what’s our, what’s our, what’s our long-term goals here? And you know, so you gotta shift from, you know, the day-to-day mindset to a more long-term mindset. You know, where, where do you need to be in three years, five years? And then build your team in and around that as well as, you know, work with your capacity providers, work with your colleagues, work with the folks that, you know, drive you to success or help drive you to success, to make a make a, you know, make a business out of it, you know, and, and treat it as your own business and treat the money as your own.

You know, you wanna, you want to consistently perform, you wanna be you wanna make sure you stay ahead of what the market’s doing as well as you know, you wanna make sure you’re staying ahead of any claims trending that might be coming through.

CW: Okay. Interesting. All right.

SW: And, and you gotta learn how to, you gotta learn how to work with data, not so much. Yeah. I’m not gonna, you know, I’m not the one that’s doing pivot tables and spreadsheets, but, you know, I need people that first of all can do that. And I need to ask the questions of them or give them the KPIs that I wanna look at, or the key performance indicators that I wanna see so that we can, we can take this huge volume of information we get in and make it into something that we can actually track and see how things are going. I think I’m trying to get over your AI side.

CW:  I was gonna say that that’s an excellent transition. I’ll send you a check later to make that one easy for me. <Laugh>. no, I, you know, I think, I think we’ve done a good job here of establishing the fact that you know your stuff. And so let’s let’s dig in. Let’s talk a little bit about underwriting processes. And specifically, I’m interested in, a lot of people are talking about this nowadays. I’m really interested in the submissions intake process, and what that looks like.

SW: Well, depends on obviously how you are doing your business. So if you’re, yeah. You, so, so let, let, let’s start with the one that everybody’s familiar with. You’re auto your homeowners. Okay. You basically, do it usually doing it directly with an insurance company. You’re doing it online, you’re entering the data yourself. Yep. I’m not in that business. The business that I am in is, we are driven by agents and brokers. So, you know, our, there’s an intermediary between us and the client. The client brings the business to this, to this in an insurance intermediary. And it could be, you know, there’s, the big ones are like Marsh, Aon, Willis, but then you have also have mm-hmm. <Affirmative> AM wins Ryan Turner, you have, and then, then thousands of smaller ones. You know, we had, we had probably 150 to 200 appointed agents that we worked with on a 65 million book of business.

 And all of these folks, or the whole, the whole goal is, is they need to give you as much information as they can that is risk specific. And then our process is relatively manual. You know, they send us the risk, and then we input it into our system. Okay. And then, then we use our system to assess the risk as well as here underwriting knowledge and background. You know, we had, we were just, you know, the the, the last two, two and a half years, there’s been a, there was a, we were in, in the middle of a, a, moving to a new policy administration system, which would be, you know, intake underwriting claims. Yeah. and all the, AI data, all that sort of stuff in the, you know, one, one of the, one of the well-known ones.

And <laugh> it, it didn’t go, or it hasn’t gone quite as well as they told us it would. And, what I mean by that is the salespeople from an organization tell you they can, you know, they can take you to the moon. Yeah. And then we actually, then we actually sign the contract and start working with the data, the data folks, you know, we’re, we’re now trying to get to the top of the next hill, much less the moon. And so it was a long, a long process over a two-year process to get from start to finish. And then, then you see, you know, you, you get your, you do all your you do your road mapping, you do all your sprints, you do all these things, and then you end up with a, a minimum viable product. Then you spend the next 10 years making that minimum viable product into something that actually is a tool that can work for you. Yeah. And you know, and I don’t know if there’s an easier way to do that, but that’s, that was, that was less than I, I was less than impressed by how that whole process went. Not because they weren’t hardworking people and didn’t, you know, and, and it, it hasn’t gone, you know, we, we haven’t made huge progress. It was just that we’ve had to build so many pieces of it that we thought was gonna come straight off the show.

CW:

Yeah. Okay. How much of that, I’m, I’m curious about this cuz I, you know, I talked to folks in a lot of different industries and I, I hear similar stories with their adventures, their adventures and enterprise tech <laugh>. How, how much of that is lack of understanding on the vendor side of the process and how much of it is maybe, you know, bespoke processes within the organization that are probably due to change?

SW: Well, I think that there’s certainly both of those things certainly play in. Yeah. You know, and I, I think the biggest hickey that, that I saw was is that the sales team, not of their heads, but, and they didn’t really understand what we were asking for. Okay. Cause you know, we do need a bespoke product. I mean, it’s not something, you know, we’re not, like I said, we’re not selling 50,000 auto policies a month. We’re selling 200 large policies a year on the Ocean cargo side or maybe 500, you know, to get to a 40 million number. So they’re, they’re large policies and they’re very complicated and very, yeah. And so it’s been, so there’s, I think it’s, I think it’s both sides. I think that it’s a, it’s a vendor misunderstanding. Maybe, maybe we didn’t explain it well enough or give them enough detail. And then, of course, it is bespoke or, or the products themselves are difficult, the wrong word is difficult, but may challenging to take from you know, from where we are now to the, from the manual process into an automated process.

CW: Yeah. Okay. So that’s interesting. All right. To the vendors out there selling to specialty insurance, make sure you’re paying attention. This stuff isn’t easy. 

SW: Absolutely. And, you know, and, and, and I think that the other piece that I would say is, again, bringing in your, bring in some of the people that are gonna actually work with the client to Yeah. Develop the product. Don’t just have a sales team that then never touches the product again. Yeah.

CW: You know, that’s, yeah.

SW: That’s, that’s the disconnect. That that’s the major disconnect. Right?

CW: Yeah. It’s funny, you know, you, you go to a car dealership and you buy a car and you’re pretty sure that the person across the table has driven a car before, but in enterprise sales, you know, you don’t have the same sort of thing, which is bizarre to me. Except at Indico, our salespeople actually know how our product works.

SW: Well, you know, and there, and there are a lot of companies that do, especially, you know, all these, all the new InsureTechs, they’re getting better and better. A lot of these guys come out of the industry, they have a fantastic background, and they’re trying to make it smoother, you know, and, and hopefully in my next position or what, you know, the next, the next place I fall, we either have a system that works really well or they have the, in the investment income or the, they can invest into something one of these InsureTechs or, you know, some, some other, some other you know, including some AI stuff, including all the different pieces that Yeah. Yeah. I think would really, you know, you know, it’s as, it’s as simple as, you know, if you could smooth the intake process from, you know, not having to do all the manual side, which I know that, I know there’s many people out there that are starting to do that now, especially if you use an Accord form or something like that for your submission.

But, you know, oftentimes our submissions come in, they’re not, they’re not on a form, you know, they’re, they’re not standardized. So it makes it more challenging. And so, you know, I think that there are plenty of folks out there that can do the standard stuff. There are plenty of InsureTechs out there. It’s when you get a little bit more complicated than it’s still, it’s still, there’s still a lot of manual work to be done. You know, AI’s not the point now that it can, take an on the formatted document and figure out what, and where to put everything. You know, we’re not quite there yet.

CW: Yeah. Yeah. You, you know, you get those loss runs or SOVs that are on, you know, scratched on a scan of a napkin. Right. <laugh>, and you gotta deal with it. 

SW: Yeah. Or, or the whole submission comes in, you know, just as a Word document or a series of emails.

CW: Yeah, absolutely. So, speaking of that, let’s, let’s walk through the entire process if you don’t mind. So, sure. I gather most of this submission process starts with an email, assuming I assume there are some attachments to it. Is it anything else? Does stuff come in by mail? Do you get faxes? You know, what else

SW: Is there? No, no faxes or mail much anymore. It’s, that’s, say it’s really emailed 99% email and then maybe a little bit of texting to follow up when you’re just on specific questions. Okay. but all the, all the documents, you know, all the documents will come in in a PDF or a Word format or Excel format, you know, and we do ask for the native environment so that we can then, you know, if it’s s o v, we wanna see it as an Excel. So then we can manipulate it simply into, cuz we have to run, we have to run it through catastrophe modeling, aggregation modeling. We have to do all these, all this modeling on it. And if it’s a PDF and it’s a thousand lines long, it’s gonna be a total bear to work with. So we do ask for the, you know, for the native, either, whether it be Excel or, you know, even Google Docs, whatever, whatever it might be. Yeah. Something, something out there that gives us the capability of easily manipulating the data.

CW: Interesting.

SW: That’s a starting point there. Yeah. Yeah. And so then that, that submission comes in, and like I said, it’s usually from an intermediary broker. And then, you know, initially, there’ll be a few questions back and forth, you know, what, what do you want us to do You want any other lines you want us to look at? And then we’ll get it to our, typically to an underwriting assistant who will then take the data that’s there and then structure it into the system. Okay. You know, they’ll pull out the insured’s name, they’ll put out, you know, eventually, this will be more broker driven. We’ll have a broker portal to do all this. Okay. Right now it’s, it’s all done internally.

CW: That was gonna be my follow-up question. 

SW: Yeah. And, we’ll get to that. But, you know, at the company where I work, it’s all driven by the underwriters. It comes in, the under own assistant gets it in there, and then it goes to modeling for the the catastrophe and the aggregation modeling, and then it comes back to the underwriter, and then the underwriter is then runs it through the rating engine so that we can then rate it up with all the different pieces that are involved to that rating engine. And then the rating engine then spits out a quote, and this is the new policy administration system, you know, we’ve got it to the point where we can go from, you know in input through the quote, through the binder, through the policy for the most part, although we have a lot of bespoke policies that are either broker, broker driven, or, you know, so they might, they may just end up being a pdf d in a document file, and this system will take it all the way through policy issuance.

If it’s a policy that is more forms driven, that are standard forms like AIs or ISO or something like that and you know, then so the quote, the quote goes out, then we negotiate back and forth. And there’s actually quite a bit of pre-negotiations done, especially on larger risks before the court even goes out. You know, what are your expectations? You know, losses, we see losses at this level, you know, that’s gonna demand an increase for next year. You know, what’s, how much pay can the client take? You know, we have inflation now running it, you know, between seven and 9%, you know. Yep. And there are claims inflation, you know, there’s, there’s gonna be an increase. How much are we gonna be able to, you know, and if it’s a long term client, you wanna, you know, wanna work with them to make sure that you’re not you know, compromised in their future or, and you don’t necessarily wanna lose it, the client either if the client is, has consistently been making you money.

CW: Yeah. Fascinating. All right. There’s, there’s a lot to chew on there. Let me start by asking, you know, in the context of this podcast, we talk about unstructured data and obviously, an email attach is unstructured. Yeah. if I could wave my AI magic wand and you know, sort of knock off one chunk of the process you described, where’s the biggest bang for the buck?

SW: I would say that if, if a system could take a pdf, let’s say you, you’re gonna see in three or four different formats, pdf, excel. Okay. Word. And you know, maybe, maybe pictures on that. If the system could take those documents and digest the documents and put that information into, the underwriting system, thereby eliminate that first step. You know, basically what the broker would do is he’d send it to a, a, a specific mailbox, and then a specific mailbox would then take all these documents and spit ’em into the, and, you know, spits a pretty, that’s a pretty low tech word, but, you know, put, put ’em into the <laugh>, put ’em into the right places, and then end up with a submission that’s been now input without any, any hands having to touch it, you know, and unstructured data. That’s a challenge. I know that because the fields don’t, don’t match up.

CW: Well and you know, it’s important to get these details right. Would Yes. You know, would an underwriter slash a broker ever trust this process to be truly automated? Or are there things that you would just have to, you know, at least glance at before you would say, go ahead?

SW: Well, you know, here’s what I would say. We say, go ahead, but then we gotta go back and check. Yeah. You know, it doesn’t, it doesn’t go past the submission phase until the underwriter’s taken that look at what has come into the system.

CW: Okay.

SW: And if it looks like garbage, it probably is okay. You know, because, because he’s gonna held the documents right here, and if, if the insured is b c Incorporated and the this address, and then in the system it says G K U company that is at this address, then we know there’s something wrong with it. I don’t know how you’re gonna solve that, cuz the unstructured data thing is a huge, that’s huge hurdle for us. Yeah. to, to get over.

CW: But it sounds like if you had, you know, I’m just sort of dreaming here if you had one screen, which is the original document, and then you’ve got the stuff from the document next to the stuff that’s in the system Yep. And, and the underwriter just has to go through and sort of look for anomalies then, then certainly a lot of time

SW: Then that should be saved a long time. That could be, the underwriting assistant could do that too, you know? Yeah. Because, because they’re the ones who would do the data entry anyway. So now they just, they just go ahead and review the data and then then the, or the assistant underwriter and then, and then the data’s been reviewed in maybe we’ve saved four or five hours. And if it, and if the, if the system can also recognize and say, oh, this is a schedule of values for, you know, this particular cargo program. These are all the warehouses. Okay, now I can look at all this and I see, you know, I see the addresses, I see the amounts, I see the locations, and then I can just go ahead and reach out to r m s or, or a i r or whomever you’re using for your, for your CAT tool.

Yeah. And then you, you can bring that information back in and you can eliminate maybe the step where it has to go to a separate person who has to, has to analyze the data and change it into a different format. But I think one key to all this is to not so much except structured data, except structured data with certain parameters or unstructured data. And what I mean by that is your s o v has to have boom, boom, boom, boom, boom, and then we auto automatically upload it. Or your bordero when it comes in, if you, let’s say you’re, let’s say you’re the capacity provider for an insurance for mga, when that bordero comes in, it has to be in this format. So then we can just upload it or the API can recognize it. So that’s not truly unstructured data, but it’s it’s, it’s a good use of a, you know, a known capability in, in the, in the system. But yeah, you’re exactly right. Have that two screens, boom, boom, boom, review it. If it’s 99% accurate, you know, then you say, okay, that just elevated my trust. One more step up. Yeah.

CW: Okay. So it sounds like, sounds like the wishlist is ai, you know, bionic arm for the underwriting assistant. And then on the other side I’m trying to think how to phrase this, some way of rationalizing the various formats in, you know, there are core concepts you need to get to, right? Yeah. Rationalizing the various formats down into those core concepts. That would really streamline the process.

SW: Well, yeah, it would, it would make it so that, you know, if, if the, if the AI could recognize that, okay, these two lines are flipped from what I need ’em to be Yeah. To go to, to go into a, into our aggregation system, I could just flip them and then, you know, then a quick review by the underwriter, this is, or the risk, the the risk engineer, does this look good? And then it’s like, okay, well you fixed all the columns and yeah. Now we can send it up in there. You know, that kind of thing might be simpler than truly unstructured data. It’s just taking Yeah. Data that’s in the wrong format and fixing the format could be Right. Could be the first step. And then the second step would be more the totally unstructured data. Like Yeah. Scanning 20 emails and, and then the system’s gotta figure out what everything is from those 20 emails.

CW: Yeah, absolutely. And this is interesting. All right, let’s, let’s come back down from the clouds and dreaming for a second. My, my experience, I, you know, in financial services I worked with and for insurance companies periodically. And my experience is that digital transformation and insurance for a long time has been kind of like the effort. Like people say it a lot more than they’re actually doing it. So

SW: Especially legacy companies. I agree.

CW: Especially legacy companies. So talk to me a little bit about where you’ve seen success with digital transformation and automation, you know, sort of in the all of the underwriting flows, submissions, you know, claims, all of that. So

SW: There are certain systems out there that make the underwriter’s job simpler than that. It does’ take away some of the, you know, fixing of the data. But what they do is they they, they take away most of the legwork. Okay. And there are a couple of MGAs out there. There’s probably more than a couple, but there are some MGAs out there that were able to start with a, like are you familiar with like Lloyd’s lab and these other yeah. Accelerators that are working with these InsureTechs, you

CW: Know, some Yeah. Great stuff.

SW: Some of the folks that have come out of these Lloyd’s labs have developed systems with underwriting input that is, let’s say they’re 75% of the way there. Okay. you know, they, they, they don’t ha they can’t do the unstructured, but, you know, they, they can do most everything else except for that one piece, you know, and, and it’s, it’s simple input and it’s, and, and then, but it’s also hugely focused on tracking that data and being able to produce numbers and KPIs and things like that out of that data. And I think a lot of insurance companies struggle with that because every one of their product lines has different fields that they use. Okay. Yep. Same thing. So that’s where, that was one of the focuses we had building this p a s system was that everybody has to use the field consistently. You know, so if it’s, if you know, even though, if you call it something different, you know, cuz I, I have, I’ll have, I have a, I’ll have throughput, you know, on cargo, I’ll have, you know, a hundred million worth of cargo throughput. That’s the same thing as revenue for cyber risk. 

CW: Ah, Okay.

SW: Yeah. You just have to make sure you’re thinking logically so that the system, so you don’t, so you don’t have to build four different systems. You can build one system that does four different things.

CW: Yeah. Yeah. Absolutely. And so you would say you’ve, you would say that’s a win is like rationalizing those things across different lines?

SW: Totally. Totally. Yeah. You know, and, and, and a lot of the, a lot of the Ensure techs that have been very successful have been very niche focused which makes it a little simpler for them to, you know, develop a system. But, you know, not, not to take away from them, but the, the, the work that they put in there to, to make it, to make it simple has been a, has been amazing as what they’ve been able to do. Okay. and I mean, I’m not real familiar with exactly what you guys do, but I will say that people are, people are becoming more and more aware or cognizant of what it takes to, or, or you know, what, what you can and can’t do within these systems right now. Yeah. Yeah. And I think that, and I think that that’s becoming, you know, what you can and can’t do now is 50% less than it was 10 years ago.

CW: Absolutely.

SW: And I think there’s gonna be that, that progression. And I think then eventually, Chris, you’re gonna be able to just throw in 50 faxes or 50 emails, no faxes, 50 emails and boom, you know, you’ll be able to spit out and have, have a quote within an hour. You know, I, I think that that’s down the road, but I think it is achievable and, you know, hopefully the folks that are listening to this call, you know, you know, give me a call and we talk about it and next time they do something, you know, they, they think about this next step or they listen and, and, and they, they figure out that, okay, what he was saying here was, let’s make sure this field is consistent across all these 10 different lines of business. Yeah. And if we could do that, that’s gonna save 10% of the time or of the cost of building something. And I also, I also think that really good pa a s providers are,

CW: Could you, could you define pa a s real quick for us?

SW: Policy, policy, administrative systems. Thank you. These are the guys that are developing the systems that, you know, do the submission quote, bind policy issuance, and then all the tracking claims on the backside and, and accounting usually off to the side as well. But the, the guys that are really doing this are starting to do it well, are taking everything they’ve learned every single time they talk to somebody and building it into the next system. Okay. And so each time their product gets a little better. Now, what I would suggest to those folks in the world to build these systems is if you were, if you don’t have a system that’s close to what you’re delivering, work with your client at a reduced rate to help build one because ah, you will also learn a lot of good things from doing that business and don’t expect your client to pay for all of that.

CW: Yeah. Subsidized research, right?

SW: Yeah, exactly. I mean, we’re, we’re happy to work with people that wanna work with us, but you know, if, if you come in and tell us you can do this, then you can’t <laugh> you don’t, don’t say, well, you gotta pay for it anyway. You know, so.

CW: Yeah. No, that’s, that’s crazy. And very short-sighted. Yeah. So I, lemme see there, there are a few things in there that you’re talking about. One I think there’s an AI arms race coming in the insurance industry, but one of the things that you pointed out is that there are, there are big wins to be made just in terms of having consistent data and processes, right?

SW: Yes.

CW: Yeah. I think one of the other things you said just now that’s really important is that there’s no one there’s no one platform, there’s no one company that understands really the full breadth of, of what’s going on. And so we need to work together to learn how to solve this problem.

SW: Yeah. And, and, and, and I think that so what’s happening now is these inure techs are getting bought and, and and bolted onto an existing company. Yeah. And the, the, the, the biggest capability that I, that I see in some, in, in some of the better InsureTechs is, is that their systems are flexible. They can input, they can have APIs into the, into that system from many different sources. Yes. and and I think that people, you know, the people that are developing these systems need to find out what all those sources could be so that they’re, they’re ahead of the curve. You know, you wanna be able to bring in the catastrophe modeling, the aggregation model. You wanna be able to bring in forms, you know, cuz they’re like ISO and ais s there’s hundreds of forms in each of those organizations.

But if you can bring them in and, and keep them up to date and, and take that away from a manual thing, that’s huge. And those, you know, those two things, the ISO and the a i s have been sort of figured out by several companies. But then you know, it’s, it’s, it’s then the little things, you know, because what we’re trying, what, where I think at automation and AI work really well is in things that are that it’s taking away process or Yeah. Putting process, automating process versus having to do it with, with a human. And I mean, the <laugh>, when I, when I started, when I got outta college long time ago you know, we didn’t even, well, first of all, we didn’t have cell phones and the first cell phone I had was a Motorola brick which was about, about, you know, this big, big guy.

Yeah. Yeah. And you know, we, I got my first laptop when I was, you know, probably five years into five years out of the Navy, you know, and in the Navy, all the, we could do with our, with our, with the computers was either type in a, in what? Pre predated a word program or, or do you know, there, there was a lot of ordering you could do online. You know, there was, that, those systems, like a Citrix system was, was working back then. But I mean, and I mean, in college we had, we had word processes is what we had. We didn’t have, you know, we couldn’t, we couldn’t look up, call up to Google and find out all this stuff. We, we had to go look it up ourselves and, you know, and so I mean the, the mere fact that we have that now and there’s so much information out there has made our lives simpler yet more difficult because now we’re expecting to be on call more, you know, more Yeah.

Overwhelming amount of information. You can’t, so structuring the information in your system, you know, and then the part that sometimes gets looked over or, or skipped over is that we have all this data, now we’ve gotta filter it into something that’s valuable Yes. To management, to the chief underwriting officer. And you know, you either have to have a person who’s really good with you know, pivot tables and things like that because Yeah, we have all this because our, the, the, the previous system was mostly on structured data that my op my head of operations, she was able to figure out how to structure it and was mostly because she understood the system from inside out. Yeah. And then she’s able to pull it into a spreadsheet and then pivot it and, and get the information that we wanted. And, you know, we didn’t have anything to compare it against, but it was way more accurate than doing manual.

And you know, so being able to go into the legacy system, pull out all that data and structure is, would, would, would be a leg up. But even, but even being able to go forward and properly use the data then, which is I think is where a lot of the InsureTechs have excelled at, is Yeah. Being able to pull in these or bring out this information so you can see trends early so you can, you know, you can, you can find out things that are happening before you visually or, or, you know, before they’ve actually happened, maybe even,

CW: Yeah. There, there, there’s a hierarchy, right? There’s data, there’s information, and then there’s insight, right. That’s, you really want to get to that insight level of things.

SW: Yes. And, but that’s where I think all anybody that’s doing data right now needs to make sure that that’s their goal Yeah. Is to be able to provide insight to the people that are using that data.

CW: Yeah, absolutely. Well, we’re, we’re we’re coming up on the top of the hour and you’ve throughout this, you’ve sprinkled some little gems along the way and I, I kind of want to try to get you to collect that all in one place. So if you, if you could force underwriting leaders to do anything right now to prepare for the future, what would it be?

SW: Well, I, I think we’re gonna go to two things. Okay. First thing is make sure you grow and build people.

CW: Okay.

SW: Because people are gonna be essential to the industry forever. And then the second thing is, is find a, a provider of systems or pla let’s let’s call ’em platforms that is cloud-based that is very much focused on future state, not past. And that brings the data from the data state to the insight state. Yeah. And, you know, and, and, and I, and I think that, well, and I, and I also, I also mentioned make sure you have your implementers involved with the sales team. And I, I mean, I think that’s, I think that’s super key because if the implementer is there and, and he, he hears something that’s, no, we can, and he knows we can’t do that. Now he can even make sure that that gets transmitted back or he interrupts at that time and says, okay, no, we, we can’t do that. We can do this, but we can’t do that. You know, I, I think that would’ve saved much of my frustration with our current provider.

CW: Yeah. I see that a lot where folks are afraid to say no on the sales side and they forget that you could say no, but Right. And that’s where those implementers become or that’s where they save the day.

SW: That’s not something that I know, but let me check and see if we can do that. Exactly. And if you can’t do that, that seems like a great thing that we can be working on with you.

CW: Yeah.

SW: Exactly. We’ll work on you with that at half our normal rate or something like that. You know, there’s, there’s ways to work around all these things that would make the that make the, they make the experience much more palatable to the people that are paying the bills.

CW: Yeah. Yeah. So the advice to the vendor side is you know, be honest don’t overpromise and unli underdeliver. And I think on the, on the underwriter side, you’re saying, you know, look for vendors that are not just selling you, you know, something that they think is, is fully baked, but look, look for partnerships correct. To get to that end state.

SW: Because it’s gonna be a long-term partnership. If you’re spending 15 or 20 million bucks on a system, that’s not gonna be something you use and discarded the next year. I mean, unless it was totally misread. But I mean, it’s gonna be something you’re gonna be living with. So Yeah. You know, you’re gonna be partnering with this same organization for many years. And so, you know, you, you need to take that, take that mind, that from the mind, that mindset needs to go in there. And from the seller’s mindset, then you can think about, Hey, we’re gonna be working with these guys for 10 years. Yeah. We better do this. Right. And not, like you said, over promise is under-delivered. Cause that’s gonna leave a sour taste in the mouth right off the bat. Yeah.

CW: Words of wisdom. All right. Well, this has been another episode of Unstructured Unlocked. I’ve been speaking with Steven Weiss, who is on garden leave from his role as a Chief underwriting officer and senior Vice President. Steven, you’ve, you’ve told us a lot of great stuff and given us some ideas for you know how to go forward and I, I really enjoyed it. Thank you so much.

SW: Thanks, Chris. Appreciate the time today.

CW: Yeah, of course. Take care.

 

Check out the full Unstructured Unlocked podcast on your favorite platform, including:

Subscribe to our LinkedIn newsletter.

Unstructured Unlocked podcast

March 27, 2024 | E43

Unstructured Unlocked episode 43 with Sunil Rao, Chief Executive Officer at Tribble

podcast episode artwork
March 13, 2024 | E42

Unstructured Unlocked episode 42 with Arthur Borden, VP of Digital Business Systems & Architecture for Everest and Alex Taylor, Global Head of Emerging Technology for QBE Ventures

podcast episode artwork
February 28, 2024 | E41

Unstructured Unlocked episode 41 with Charles Morris, Chief Data Scientist for Financial Services at Microsoft

podcast episode artwork

Get started with Indico

Schedule
1-1 Demo

Resources

Blog

Gain insights from experts in automation, data, machine learning, and digital transformation.

Unstructured Unlocked

Enterprise leaders discuss how to unlock value from unstructured data.

YouTube Channel

Check out our YouTube channel to see clips from our podcast and more.