The Not Unreasonable Podcast Podcast Artwork Image
The Not Unreasonable Podcast
Why Building Systems Is Hard With Bill Jenkins
May 30, 2018 David Wright

My guest this week is Bill Jenkins. Bill is a technology specialist in the insurance industry. I like to joke that the customer satisfaction rate for insurer systems is 0. But does that need to be the case? I've finally had the chance to ask these questions of an out and out expert. Bill has headed up internal technology projects at insurers, he's run the technology at brokers. He's been a consultant. A Board member. An industry standards advocate. If there is a puzzle in insurance technology, Bill has probably thought about it and here he is today to help us all better understand why we struggle with technology in insurance. Are you an actuary? Someone you know? Check out the Not Unprofessional Project at notunprofessional.com, content dedicated to Continuing Education Credits for Actuaries, especially Professionalism credits, for the price of a CAS webinar. CE On Your Commute! Subscribe to the Not Unreasonable Podcast in iTunes, stitcher, or by rss feed. Sign up for the mailing list at notunreasonable.com/signup. See older show notes at notunreasonable.com/podcast.

Episode Transcript

Speaker 1:0:11Welcome to the and reasonable podcast. I'm your host, David Wright and actuary reinsurance broker. This is a show of interviews with people who have something to teach us about managing our businesses and ourselves. There's a lot to learn out there, folks, so let's get to work.

Speaker 2:0:32The show is brought to you by beach, right work and have worked my entire career. We are a global reinsurance intermediary and we pride ourselves on creative thinking, deep analysis and client service. Those three qualities are actually intimately related because you can't spend the energy digging deep into a problem unless you really care about the client. We find that when you really do understand the problem, the solution becomes obvious even if it seems a little bit unorthodox to the outside world. At first, the nature of insurance is to take solutions we know and trust and try to force them onto the problems of the day. We just don't settle for that. You can see further at beach, GP.com if you're an actuary. You liked me, probably dread the professionalism continuing education requirement. I think that the best time to satisfy this isn't podcast time. While on your commute, why walking your dog while mowing your lawn while doing the dishes head over to not unprofessional.com were for the price of a cow's Webinar. You can get content dedicated to continuing education for actuaries, especially professionalism. See, that's not unprofessional.com. Thanks for listening and thanks for supporting the not unreasonable podcast.

Speaker 2:1:46My guest today is bill Jenkins. Bill is the founder of Agile Insurance analytics, the consultancy specializing in insurance data solutions. Bill is had a variety of technical and management roles over 35 year career, including a cio of a broker and carrier and board member of various organizations, notably including being along am best's first outside directors. Bill, welcome to the show. Thank you David. Pleasure band here. So you've had. You've had an enormous amount of experience managing the technology side of the insurance business. And what's interesting to me about your experiences that you have both on the carrier side and on the broker side. Right. And one of the things that I think is, is, is underappreciated in the insurance business is actually how much technology there is in it because so much of it as in a human side and we think of ourselves, obviously we are humans interacting with other humans. Right? I'm wondering if you'd give an overview of kind of the technological chain of insurance from kind of the first interaction and insured might have with a risk, you know, all the way to, who knows where it ends, the claims payment, I don't know. Okay. Well, there's a variety of ways you can attack this. David,

Speaker 3:2:45uh, obviously, um, a personal lines, a type of insurance would be a different approach then would a commercial lines policy. Um, so let, let me talk about probably the biggest market is the personal lines market and uh, the, uh, initial, uh, input that would be, uh, taking place in that process would be through an agent independent agent. And I'm leaving out the direct writers right now as well as a exclusive agents. And uh, today, as you know, most individuals are connected to the Internet. And I think I saw a statistic that indicated that over 90 percent of the people who, uh, who are, uh, with a mobile devices are connected to the Internet and 98 percent of those do their shopping on the Internet, at least gather information.

Speaker 2:3:48Oh, I mean, I, I totally buy that and I do almost all of my shopping online, almost all of it, right? So, um, I'm interjecting here a little bit, but Canada originally and an Amazon.ca is a pale imitation of Amazon.com. Right? So the, the Canadian version of it, the amount of online shopping you get done and Canada's is dramatically less. It's, it's, it's a whole different game down here. I was certainly phenomenal. Certainly what you can do is just incredible. So that, that requires

Speaker 3:4:15leslie is for the agent to have something to attract the potential insured, the policy holder. And in order to do that, obviously they have to have a presence on the Internet and they should have a webpage that is attractive to a individuals and gathers their attention. Right? So as part of that, also they need to have a degree of robustness, uh, to make sure that they can accommodate and mobile phones and the mobile phone also would be used by the agents, producers, the individuals who go out and sell insurance to do quotes. And so therefore they had to have a pretty robust presence on the Internet. Uh, the third thing that an agent really should have, it's basically a crm, a customer relationship management system so they can sort and categorize their, a potential customers and obviously they want to have some sort of social presence because people today mostly have a profile on a social presence so they can gather information that way. The next step in the process would be for the individual who is shopping for insurance to basically a fill out some basic information and it's rudimentary information because today there's a lot of technology that allows a prefill of applications and it's the accord at. That's basically using personal lines.

Speaker 2:5:43Yeah. How long has it been caught up in around is, is that like an internal thing that started back in the midst of time? Uh, I was only a cord board. No kidding. Uh, yeah, I was chairman of the board for two years. Okay. And here's where the days, uh, those where uh,

Speaker 3:5:58she, uh, back in the nineties. So was it a new thing then or was that a court? A court basically was, uh, initially as it started out with a former company and really what transpired was iso use to develop the forms for the insurance industry and accord took that over. Okay, interesting. You know, I think a isl had different ideas as to what their business plan on a look like. She willingly walked away from it. They did, we didn't want to do this. They did ISO by the way. At the time of nonprofit. Yes. And as the court is in nonprofit too, I didn't know that. Okay. They're both nonprofits. Both. Both have surpluses away, but none of that. Who owns accord then accord is owned by their membership. Okay. Much like ISO. ISO was. Yes. Yeah. So it's still owned by insurers that being the membership?

Speaker 3:6:49That's correct. That's correct. Any non insurance part of that? No. So no agents either. I mean, not okay, there are agents, members than me take that back. What I meant was vendors, their vendor vendor agnostic, so they say, so it's a standards organization. It's kind of like this kind of like the other ISO, right, which is the international standards organization. But where I grew, I grew up in this kind of manufacturing part of the world outside of Detroit in Canada. And I remember seeing banners outside of these facilities that ISO nine. I mean there's a lot of zeroes in there, 9,001 certified or whatever it is. And so that's just saying we comply with this and this I'm aware of that. I'm aware of that. I got involved later on in my career around the standards especially they said related to data quality and ISO 9,000 was a big issue.

Speaker 3:7:37They had a thing at the time they called the Baldrige Award and there were certain criteria, uh, that was required that would afford a accompany because it was a cross industry, uh, the uh, possibility becoming a Baldrige Warner. And uh, so, uh, the ISO 9,000 was basically the, the premise of all of that. So, so according to standard accords, the standard standard information, is it the same everywhere? I think I've seen an accord APP before, but I'm a reinsurance guy so beautifully. This is interference with their heads. That's right. So it's the same for like a commercial risk or personal risk. They would have a different fuels. They would have a different forum and there are forums as you know also that accompany a policies and they have the charge also developing new forms for the industry. Oh, interesting. Okay. Trying to standardize. Interesting. You would pay for the forums?

Speaker 3:8:36Yep. You would pay for the use of the standards which they have backing away from, but I am with the understanding that they are now wanting to start to enforce that charging for charging for it. Okay. Right. Now the other interesting thing to me was that people don't realize the numbers of members that a court has and as you know, they're worldwide. Okay. Uh, they cross industry from the auto insurance industry from the perspective of life insurance a property. And casualty and insurance and bonds and, and a lot of different forms, a lot of different forms, but there aren't a lot of members and a thrust has been, how do you make those standards more pervasive because larger carriers as an example, Safeco and my day, uh, use their own standards. So a tier one carriers began to use around own standards because they can pretty much dictate to the agent and the policy holder what they wanted and the data they wanted to collect. But that has changed somewhat, but the, that continues to be a, an issue for them.

Speaker 2:9:51There must've been emerged, some kind of conversion process between news will get an accord forum or maybe maybe the agent's only ever done with travelers. And so now they have to convert it to an acord format. Does that happen a lot?

Speaker 3:10:03I don't think it happens today as much because I think it's pervasive. The acord forms pretty much dominate, so now, yeah, pretty much dominant. They've been able to achieve that, which has been a great achievement for him. Yeah, that's interesting. So pretty interesting.

Speaker 2:10:16Okay. So we're, we're, we're, we're, we have the story down to filling out a form and that that's electronically done these days or how many,

Speaker 3:10:22that's pretty much done electronically and as you know, if you shop online, you do it electronically, uh, and, and your own on your own pc also. So as the carrier a obtains a year application and has rudimentary information, they can then do certain things with it. They can, they can either go out and, uh, augmented data that you have provided them as a policy holder applicant or they can and slash or they can go out and they can, uh, find out, uh, as part of that prefill your history, your loss history, all the data that you had primarily furnished your past insurance carrier and fill the application in using that process. There are people out there that can figure this out. Yes, there are. There are companies today and I think what has transpired and nexus lexus is one of the companies, and I think iss does this also. It's a prefill and they go to companies and his companies that they want to collaborate with the other companies to pull their data kind of interesting so you don't have to fill in all the coverages that you had before that's already, that's already in this pool of data that they can extract from and fill in the, uh, the application with and uh, I think just pretty, pretty much new over the last several years has had done that well, but it's a voluntary type of process to carriers are doing so. So, so we have three

Speaker 2:11:50existence so far. We haven't even written a risk yet. So there's a web presence,

Speaker 3:11:54right? And it's probably a multitude of systems. There's a crm process of just activity in the sales process. And then there's this accord app kind of series of systems. That's correct. That's correct. Are you gonna write a policy coming soon? Well, what happens then is a car. Most carriers will have a um, a web front end. Yup. And the carrier was submit the data that they have collected to the carrier and the carrier. We'll go ahead and do a quick rating and pricing with the data that they've collected thus far. The Acorda Corp Right, and the prefilled data that they have received. Last history though, which is not a new cord up. That is correct. What isn't an accord up? Maybe we should pause again on a court. Ups Court App is I am an address, it's a template for collecting data and a lot of carriers have built out and agents and agency management system have built out their front front screens or input screens.

Speaker 3:12:51The follow that APP makes sense. Yeah, of course. So with like the data once. That's correct. So the idea then is that that data is then provided to the carrier. They'll do a rudimentary pricing and rating on the data, what the data they have. Then they will bind the coverage by sending you the applicant, any agent, a copy that you're bound, you have a policy with them at a certain time, at a certain rate with these certain coverages. So that's a lot. So there's a, there's a, there's a system integration that's going on here. So there's a, uh, some kind of web form for mobile phone form where you can enter in your information that gets sent to a carrier or some kind of carrier code and bind carrier, carrier, carrier, um, and then, and then it comes back, right? Additional information is then collected.

Speaker 3:13:48Okay. And then the questions and more questions. And in the extraction, uh, an input of additional external data like mbrs clue reports, what's her last reports that are more robust, go back further. Underwriter will make a determination now in most cases today, and personalized system that's all automated. People would call that a artificial intelligence or I would call it the old expert underwriting. And that's how we used it back in the 70 [inaudible]. So just a matter of a number of rules, organic intelligence or intelligence intelligence. So what transpires then is, uh, the system will make a determination whether it will pass that data, that information through that you qualified, they get this, they get this policy, or it will do a referral which will have an underwriter to take a look at it, or we'll do a, a reject. We don't want that. It doesn't fit our appetite, our underwriting appetite. So how many integrations

Speaker 2:14:52do we have here now? So there's a, there's a web form, maybe it's integrated with the crm, integrated with your core up perhaps I guess, and then that that is to quote and bind system is some kind of record front end and then all these vendors that are pinging, you know, Lexis nexis and external data we're bringing in. So who manages all these integrations? Like who, who is there a person that says, oh, we're going to upload the assistant to the contract people. Okay, that's okay at the broker. I guess at this point.

Speaker 3:15:20And with the carrier pretty much. Okay. Right. She had the proper basically says, here's the data I have. Here's my connection to you. It's all yours.

Speaker 2:15:31Why not at the broker? It seems like this is all customer service related kind of thing. I mean, someone's underwriting

Speaker 3:15:35ease of doing business and that's going. The carriers, that's a carer is mindset. You know, they, they rely on the broker or the agent, agent Barbara, two different animals, you know, they rely on the agent to bring nothing in business. Yup. All right. And, uh, if they don't make it easy to do business with them, yeah. The agent goes with someone else. So the whole idea is, you know, we'll do all that work for you because you've heard of things called direct bill. Agent used to do all the billing. They don't do that anymore, especially in personal lines. And the carrier does all that work for them.

Speaker 2:16:11I gonna kind of on this integration thing just because I think that this might be something that's underappreciated about how hard it is to talk to you. What, what is that? It's hard. And how hard is it?

Speaker 3:16:20Well, uh, let me, uh, I think, uh, uh, I'll comment on that based on the fact that, uh, to do the mapping of the data, making sure that you understand the receiver of the information and how that format or to interface exactly what the data looks like, what the data looks like, what format it is, a, is it in field 63 years and fields 50 over here, and then you'll have to do the mapping at 63 maps in the fifth date that the data is there that you need to do your underwriting and ignored the raining. Uh, that's always an issue. What do you do if it's missing? What do you do? If it's missing a w, uh, working on, uh, a broker agent side for a number of years. It became pretty, uh, interesting to me that when an agent is seeking business, a lot of times the policy holder won't give them all the data they want. To what degree? If you're trying to get that business, do you strong arm that potential policy holder from giving you that data?

Speaker 2:17:27Yeah. Because that's bad. Easier doing business. That's correct. That's correct, besides right from both sides. So it's kind of, it's kind of a um, so what, what does, what does the technology manager at a broker do, if all this ease of doing business technology related sits with the carrier? What's the broker's role with and understand there's different, there's different tiers of brokers and categories of brokers, uh, were using a general term

Speaker 3:17:50broker in this, are you interested in? Um, it depends on how big the brokerage and what they can afford to have and that also gets into not only the staff to maintain it and support it, it also gets into how many systems they'll have their [inaudible] and uh, you know, a lot. A lot of agents don't have crm, a customer relationship management systems. They still do that manually. Presumably that's going away though. I mean I feel like there's a lot of consolidation that's going on in those businesses using so many different aggregators. Are companies recently bought by a company by a lot of producers and agents? Absolutely. And I would think that there's a, there's a scaling up that's going on probably in part driven by the need to have the systems. Do you agree with that? Is that. I agree, I agree. And you know, each system is kind of interesting because generally speaking, uh, and I guess I guess I guess it's a good time to revert back to the older days, uh, where our systems were really designed to be operational in nature as through data's as fast as you can and what that meant was data that is used today to do underwriting and to learn more about your customer was never captured.

Speaker 3:19:04Okay. It was don't write it down on its way through. It was either aggregated or it was just a overwritten and, and gone, you know. And that's the big issue with legacy systems today. Okay. So, so make a long story short, it takes a carrier because I just did some research on this that's in a invest the best review addition to that. I interviewed about 25 companies, about third party data and one of the biggest stumbling blocks they had and using third party data was the time and effort it took to do the data mapping. And part of that is the integration we're talking about. That's the integration, right? And it took on average anywhere between six and nine months. Wow. The extracted data that's wanted to merge it in, find a place to, to capture, not only capture it, but store it silk abuse.

Speaker 3:20:07So I mean, I heard, you know, you always think a here, these quite remarkable references to long lead times of her systems deployment, right? Day to day, what's going on? Why does it take so long? What are people doing everyday? It's just that many problems to solve or you know, what slows it down while there's a lot of issues, obviously if you look, if you look at, uh, and I, I think it's gotten better since I used to run it departments could, uh, but over 80 percent of the resources and the it department or used to maintain systems. And so what does that tell you? It tells you that these old systems broke easily and were bugs in them that, uh, required, uh, effort to address and fix a. There were regulatory changes that came in that required system changes that was all part of that 80 percent. I'm adding new data modules to assist them. Part of that. Uh, so what did that leave you with that left you with about 20 percent to do what we call discretionary spend on new systems?

Speaker 2:21:18How so? What would, what would be me coming at this in a different way to please it would be a difference between call it the top tier and call it middle or bottom tier organization in terms of the ability to execute on, on technology projects like w what makes the best better than the rest?

Speaker 3:21:39I think that's what the big movement around a replacement of legacy systems has been all about. The component type structure or systems that you can plug parts in and out, more integrations, integrations, but it's easier to do, easier to integrate, easier to doe. And uh, as things went forward, um, standards were becoming more and more pervasive. Yup. Uh, so therefore the data that was being captured and stored was almost the same. Part of the problem today is that vendors go out and they don't have a standard data model or do they have a standard, a data dictionary or business glossary. So what happens is if you have 10 vendors as an example, each one of those vendors will have not only their own data model, but to also have their own data dictionary, which would then cause or data models and data dictionaries. Okay? Data model shows the relationships between data and it's broken into components, customer table and like a transaction people or something, right?

Speaker 3:22:50But it's, it's a way to store data pretty much and how data relates to each other and how it goes from one road to another. And it could pick up all right and how you would measure it. Um, a data model tells you how to do that. And it's way of also tracing the data through the organization. So a data dictionary or business glossary would say, we call as an example, uh, on, on premium this, this is how we calculate on, on premise. Everybody speaks the same line we've named. It's a standardized way of doing it, right? What you have is if you have 10 vendors, each one of those has their own way of categorizing data there an anway of naming conventions and therefore if you are, if you're trying to integrate vendor product one, the vendor product to get to go through that and matching because they don't talk the same language

Speaker 2:23:47which, and you get to learn about that and then you've got to deploy that. No one one, as I was kind of thinking a little bit about this, this conversation and in prepared preparation for it, I was, and I sometimes come back to this, this idea in my own head of, of the metaphor for technology as being another kind of infrastructure transportation infrastructure. Right? Right. And so one of the things that really fascinates me and kind of a dark way is this idea of legacy systems and it's, it's such a burden on the industry and I don't know that firsthand as everybody tells me. So I assume that you know that they're all, they're all, they'll know what they're talking about and that's true, but you don't feel as much of a burden of legacy systems in, in transportation infrastructure. Now I'll pause there to say people complain a lot about the transportation infrastructure.

Speaker 2:24:30United States, it's deteriorating the engineering report or whatever it is. But I think quite a lot of that is there's some of that is going to be safety related, but some of that's just going to be aesthetics. Like I, I don't know if you, how often you come into New York, but from you're based in Philadelphia? Yes. And if you drove in towards the Holland Tunnel in New Jersey there, when I first moved here eight years ago, that was like something out of like, I dunno, deep third. Like I felt like I was driving through Baghdad. It was terrific and they've upgraded that I think as a result of the financial crisis, they had that big, you know, spending plan when they're. And they have pediatricians but it just looks better now. I mean it's still drive on the road. I mean, I don't know how much, how much better it is.

Speaker 2:25:09It kind of actually, that was a legacy system that didn't really affect performance. That's correct. That's correct. One of the things that come to mind as you were talking there is the idea of standards and so I know a little bit about this where in the 18th century or 19th century, 19th century, 18 hundreds, the rail lines had different gauge tracks and so training couldn't go from one track to another. You literally couldn't couple them, but that change and they have standard gauge tracks now in the United States and so maybe legacy systems did exist back then, but they just sorted all that 100 years ago, so we don't even know. That's. That's a great analogy analogy. So legacy. So we just have this maturation phase and that is actually defined by standards really. It's not. It's not that all apps are nicer. It's just like, oh, it's the stuff.

Speaker 3:25:57See, that's correct, but the unfortunate part about it is, and I love that analogy. If you're going from one gauge track to another gauge track, think about how you're going to do that and how long is it going to take you to figure that out and get that done so that the train can make the pivot across. That's the problem with these legacy systems.

Speaker 2:26:14Yeah. So we're like 150 years behind the rail technology. Yeah. That's all it is. That's correct. Interesting. Right. Um, let's, uh, let's, let's put the story back up again. So we are, so brokers brokers are managing all these integrations. Actually, that's that kind of something that they do or the broker technology manager.

Speaker 3:26:33Well, um, it's going to be the, um, the insurance carrier that basically does all that. Okay. Okay. So what does the broker technology manager didn't do? Not a lot. Okay. Well they'll with uh, you know, they'll put in new systems. They inner office systems that the broker will, you as the agent will use crm that crm, uh, you know, they'll do, um, you know, uh, making sure upgrades are done on the systems that the Internet is working, that kind of thing. Get your email, you can get your email. So that's basically what they do and make sure that, what pages out there, you know, is that changing necessarily carriers or stole, you know, carrying and most of that burden if I say it.

Speaker 2:27:12So is that, um, one, another, another theme that I like to think about it sometimes is it's sort of this grant, I don't know, I guess evolution of technology and how there's a implicit power struggle maybe between carriers and brokers, because for carriers there's this disintermediation which always looms over the broker, right? In the carrier. If they just start going direct, why don't they just want to just do that? And do you really need the broker anymore? Do you need to produce or anything

Speaker 3:27:35more? That's been in an argument and discussion that's been going on for the last 15 years. Yeah. What's going to happen with the agent, especially with the new millennium generation coming in who want to do you want to have access, direct access to the character? I want to do business directly with the carrier, but it all evolves back to making sure that they get the advice that they think they should be getting. And that's the agent that fulfills that role as a human interaction. That's a human interaction that's going to be an individual who has the expertise that's going to tell David Wright that he needs to have this coverage because of his circumstance that, uh, you should get. You should look into an umbrella policy as an example. Uh, what does that entail and what are the deductibles, what can you afford, that type of thing. Agent gets that kind of advice. So,

Speaker 2:28:30so, alright, so the agents are not really all that involved in the kind of a system, the grant and the carriers are, are owning all that. Um, so let's talk about carrier system.

Speaker 3:28:42Okay. So the carrier gets data, as we mentioned earlier, from the agent or directly from the potential policy holder a, they'll do a very quick, uh, underwriting and rating process that allows you to make up your mind whether or not you want to have a in that character, provide that policy for you. If you so decide side you want to do that, then the PPO policy will be bound and it would also be generated. And then they would set up the billing system and the other systems that they might have for statistical reporting to ISO or an ai stage be separate systems. Yeah.

Speaker 2:29:27That. So there's a quote on quote and bind system, right? Distinct from a, from a, from the

Speaker 3:29:33admin system, which they call a, um, uh, the, the, uh, it's the, uh, standard record that they then use to create their financials. And they also used to, uh, comply with regulatory requirements and report that data. Yeah. So,

Speaker 2:29:54well, one of the things that, so as soon as the reinsurance arbitrary, right? A reinsurance broker actuary, we interact with the output of different systems and we don't know what systems they are. They just say, here's, here's some data, right? What I've observed over my career is that premium and claims pay data is the best because that's my story for that is because it's audited, somebody is using it for something other than managing your exposures and those people care a lot about where the money is and so that's real good quality something.

Speaker 3:30:24Carriers did not do a good job. I'm sure eventually their families, when they go back to collect and a reinsurer sand, we don't match your records and you're saying you want us to pay you for this, and a number of carriers have gone belly up

Speaker 2:30:41because there's. There's very much A. I would put it this way, there's very much some natural controls in the world that exists to make sure you're paid data's correct. Right, right, right. I would say on a relative basis that's better. I mean, of course they're gonna be problems relative to cold exposure data. Right, so what's the limit on a policy? Did I buy a facultative of reinsurance on this, for example, or, or a variety of other things which aren't audited, which aren't necessarily cashflow related until there's a claim and so our crap or crappier I guess, than the, than the pay data in. Another thing to have mentor of mine likes to say is that all data's dirty until it's used for something. And if it goes unused then you know. So let's say the policy administration system data must have a lot of some messes in there until it interacts with this called the billing system or the auditing system on the right track there. What do you think?

Speaker 3:31:34It depends on what, you know, it depends. There's, there's another variables involved with that. David, uh, one would be, um, what kind of a quality control process company uses? Um, I at one time, uh, as part of that actuarial function that I had had statistical reporting and where the edits that the carrier had, our carrier, that company I worked for had, were not as severe as the ones that we reported data too at that point in time was isl. So what happened was that in order for us to turn the data around so it would be accepted by an isl and if it's not accepted then you get fine to cost the company money and got report cards. Oh yeah. And finding money. I still find suit. Yeah. I still find you. And uh, what we had was Rosa clerks who imputed data, right? So as an example, they would plug in data that would get through the edits, didn't necessarily mean that it was accurate data, it was just valid data set.

Speaker 3:32:46Maybe dwell on that for a second that this is accurate and valid data. Accurate data is accurate, you know, if you're 25 years in reality it is. Yeah, it is. And has integrity. It's complete, it's timely. So it meets all that criteria validated mean it's acceptable, right. She just accepted it's an integer, it's an integer. So, or not. So you would find, you would fine as the Edis got more and more sophisticated for entry into these other systems, uh, that, that people who are imputing the data would find out that their birthday as an example, was acceptable. So will you would find rows upon rows, rows, or fields upon fields of data with this individual's birthday. So obviously valid. Got Her the edit but not accurate. So I'll give you another classic example. Uh, last carrier I work for, uh, found out by running certain edits that all of the farm equipment that was being insured by the company, we're all Mercedes Benz 500 sedans doesn't sound right.

Speaker 3:34:04It doesn't pass the smell test anyway. So anyway, so the data is important from a number of perspectives. I know that you're aware of that, but it's trying to standardize the data within the carrier that becomes a big issue. Standardized, standardized. What do you mean by that? Well, making sure that everybody sees a data the same way that all of the dashboards that a company may have, you know, the management information systems, everybody's looking at the same thing and it means the same thing to everybody across the board in today's world. That's an anomaly. Yeah. Interesting. Yeah. So how about claims? So claims to me is always thought and it's kind of like alternate universe is like a separate path, right? Right. It doesn't connect to the policy of your billing, I guess. Yeah. Yeah. Well, you know, to finish off the jury upside, you know, then you have the stat reporting and I think I, one, one of the things I wanted to mention is that some carriers have multiple and duplicate systems.

Speaker 3:35:11So, uh, I worked for a medium large, medium sized carrier. We had $8 million systems. Whoa. So think about trying to tie acquisitions, partly acquisitions, partly because it's one system than a address. One problem with the other system did. Oh my God. So they decided that they needed the, uh, this additional functionality that the old system didn't provide. They just wanted another one where we had three bop systems. Okay. Uh, so, uh, I think, uh, I was listening to a talk that the chief technical officer at the Hartford was giving and he said every year they go through the examination and review to determine if they should replace all their legacy systems. They had over 330 system and he said the replace all these, they, they projected out would be, you know, 50 years or so that do that and the cost would have been astronomic. Right. It's cheaper to just live with it. Said they just had systems.

Speaker 2:36:14Well, so what, so was there a, is there anything that's new about that process? So I'm wondering, was there a time when it was harder? I guess maybe this is kind of coming to the maturation of technology. There was a harder to integrate or harder to replace back in the day, whatever that means. And so you are more likely to buy a new system then then replaces the system. Is that true?

Speaker 3:36:33It begs the question that you had a, I think when we initially talked about, um, sideload siloed thinking and how underwriters and other, uh, insurance technical folks, um, see the run with fiefdoms. Yes. And they don't look at things, they don't look at things across the board from an enterprise standpoint. They look at it in silence and we constantly see in the practice that we have a predictive analytics falling into that situation where an underwriting department here will have a predictive analytics solution that could be used by other functions within the organization. Are there, the other parts of the organization don't even know that that predictive analytics solution exists and is part of that, you know, they have their own data models so they have their own language, so to speak. Uh, and so it goes on and on. You may have a, as an example, a 10 predictive analytics solutions across the enterprise and where it could be used across for the enterprise that is not being used because of a silo mentality.

Speaker 2:37:53And so in a siloed world is more likely to add a new system because they don't realize the solution. It might be an easier way. That's correct. Coordinating organizational resources to make a big splash. That's correct. This is hard to do. It is hard. Yeah. Yeah. Um, so one thing that's clear about systems it is to me is that they're low for everybody. I talked to you, right? So I might have the world that I tend to go run in are the, you know, call it insurance companies as the management is the buyer. It's the actuaries, underwriters who aren't, who aren't technology people, you know, whatever that means. Uh, and I wonder what that means. But um, you know, they, they really don't appreciate, I think how hard it is. That is one of the things I want to get into this call just this conversation was, you know, like how, how is the, how in what ways is this hard?

Speaker 2:38:42And I think the way that it manifests itself, most clearly in. I went to undergraduate business program when I was in university and they were, there was a, I forget which class it was saying, it's probably a technology class. So some sort of that talked about how it projects that are being assisted in his project creation project are always overtime and over budget. That's correct. Right. That's one of the things that I've kind of pondered a little bit since then is to what degree that's a technology problem. We went to project management problem. I want to read you. I'm going to read out kind of a lengthy quote, have you ever heard of the book thinking fast and slow by Daniel Kahneman? So he's a, he's a, he's a psychologist who won the Nobel Prize in economics. I'm an energetic, a little passage from the book of his, which I think is really interesting and I want to get your reaction to it.

Speaker 2:39:26Bearing in mind, to what degree is this a project budget problem versus it's no problem. So here, here's the quote unquote. In the 19 seventies I this thing, Dan Kahneman convinced some officials in the Israeli military of education, the need for a curriculum to teach judgment and decision making in high schools. The team I assembled the design the curriculum and write a textbook for it included several experienced teachers, so psychology students seem more Fox who was then dean of Hebrew University School of Education and an expert in curriculum development. After meeting every Friday afternoon for about a year, we had constructed a detailed outline of the syllabus, written a couple of chapters and run a few sample lessons. I asked everyone to write down their estimate of how long it would take us to finish. The average was two years and there was. He made a point of it being clustered tightly around two years, which I didn't copy down.

Speaker 2:40:12Then it turned to see more or a curriculum expert and asked whether he could think of other teams similar to ours that have developed curriculum from scratch. Seymour said he could think of quite a few and it turned out that he was familiar with the details of several. I asked him to think of the teams and the teams and where they were. They were at the same point of the processes we were and how much longer did it take them to finish their textbook project? He fell silent when he finally spoke. It seemed to me that he was blushing embarrassed by his own answer. You know, I never realized this before, but in fact, not all the teams at a stage of completion comparable to ours ever did complete it at all. About 40 percent fail. Okay. Of those who finished, how long did it take them?

Speaker 2:40:49Well, I can't think of any group that finished in less than seven years. Seymour said, nor any more than, nor any that took more than 10, I guess a graph, but a straw. And I said, when you compare our skills and resources to those of the other groups, how good are we, would you think? How would you rank us in comparison to the others? He said, well, we're below average, but not by much. In the end, the book project took, it took eight years. And so it's interesting about this whole bunch of things are interesting about this one, that this was a long project. Um, and that they were kidding themselves. This, by the way, this is Daniel Kahneman who literally wrote the papers on this very kind of on this bias. And, and he himself still fell victim to it. And it was amazing was that this guy was in the group, knew the answer himself, and he was amongst the people who average up to two years. He didn't

Speaker 3:41:34say it's going to take eight years. You're kidding yourself. Right? So there's this inherent problem that we have as humans and predicting how long complicated things are going to take. What do you think? Uh, I think it's cultural issue. Yeah. I think it's a cultural issue. And I can give you a classic example of why I say that I was thrown into a cio, cio role for a carrier, a big carrier, and the thrust was that I was being given nine months or a year rather to build a new system. Okay. I knew it was a bob system, right? Just being a pricing quote system. This would be a, a policy admin system, system of record, right? As well as the quoting and the rating piece of it, the pricing of it. So what I was able to do, and at the time you may recall at the book in search of excellence, I have heard of that yet I haven't read it.

Speaker 3:42:40Uh, there was a concept of teamwork used in that outlined in that bulk caught a skunk works. And so my mindset was because having been through these wars many, many times that David Wright would have not only the, uh, the job to a responsibilities to build out his piece of the system and work as part of the project, but he had his day job. So the mindset was there had to be 100 percent focus and an individual's attention span to do this project within the set parameters. So what we did was, and by the way, we also brought in a new tool to use and it's an Israeli tool called sapiens, which was a business roles that company, not by oracle or something like that. Yeah, it was bought by. Um, no, they just bought somebody. I've heard of the company. Yeah. It's Israel. They built, they built two things of note using this tool.

Speaker 3:43:45They built the Israeli air defense system and they also built the Atlanta. I'm a Olympic Games, a reservation systems. So it was a different type of a emergency management kind of thing or. No, it's basically a coding process, but they use rules a instead of a code instead of a cobol code, they use these different. It's a different language. However, uh, what we did was we see quester and a totally different building. The it folks we needed as well. We thought we needed as well as user community in the subject matter experts. Not only did they build the system and nine all working together and accumulating and cross feeding data. We also use the project management discipline that we call black hat white hat. Black hat was a hired gun project manager who comes in and his or her and only charge was to make sure that the specifications for the system were done and that was they. We're going to be followed for the requirements of the system. And that, uh, the timeframe that was set would be adhere to a, the white hat was an internal project manager who basically made sure the right people were on a project to do the work and also did all the reporting to the senior management and navigated the political waters.

Speaker 3:45:22So with, with those new processes in place, we were able to build a system of Bob System and by the way we use function count analysis and brought somebody in from the outside who was an expert in this, the layout, the complexity of the system and what they indicated that it was a, a very high complex system because we're building our own rating engine and as such we built the entire system in nine months. So now things can be done. Well, let me, let me just give you the antithesis of that. The next time around we went with an internal project manager. Kept same skunkworks 22 months and the same difficult, same function point accumulation, right? For the next system and the difference project management. Right, right. And what is, what makes a good project manager? Well, first of all, uh, the, the problem with an internal project manager, and I argue this all the time, even when I sit on boards and people are having project problems, uh, a project manager from an internal may know all the project management disciplines, but they pretty much don't have the personal characteristics to do the work, right?

Speaker 3:46:34You have to be a pit bull, put the black hat on, but it's black Haddaway, uh, and your, and you go native too quickly. I mean, you're native too quickly, so therefore your scope creep because scope leap and you're fitting more and more in the project and that doomed to failure. Right? So that's the big problem as it was one of the big problems that stick to, right? They said he wants her to one of the big problems when the problems come to your mind. Oh yeah. Um, data first data when it comes to projects is pretty much put on the back burner. What do you mean by that? Well, uh, what needs to transpire as you have to have a, a data model to make sure that everything standard, when you call something, you also want to make sure that you're capturing the data that you really need, that the system is, has the capability of capturing the data that you require and her many, many times they don't detect that until they get into the last phases of the project testing and then you got to rip it all up and I got to go back and re engineer and all of that type of thing.

Speaker 3:47:38Whereas if they'd done it upfront as part of the planning phase of the project, they would have known that. So why is it easy not to do that? So obviously this is a common pitfall and a big one, but what, what, what, where, where is the point at which this happens? What'd you get distracted by that prevents you from doing that in the project? I guess. I guess it's a fact that the project manager and the people in the project don't have the wherewithal and knowledge about it. Right? And it becomes an educational issue. They don't have the discipline to stay on something that's correct. Were you distracted by whatever the micro issue is there and went off and there was always a fire that's Yada, Yada Yada, and you do it. And there are two Achilles heels, uh, when it comes to building projects, especially, you know, large complex projects once project management, then we just alluded to the other one is a testing process and testing generally gets short shifted because it usually comes at the end of the project and as a result of that, uh, it's cut time is cut because they want to make the project deadlines.

Speaker 3:48:42Yeah. Yup. And then therefore the other, the other family and is, and what I've seen across the board is the lack of putting measures up front about what this project is supposed to achieve. A, I call them project Kpis, key performance indicators. Very rarely. And that gets the culture of the insurance industry too, by the way, David, uh, you know, what is this project going to achieve for us? Why are we doing it specifically? And then at the end of this project that I achieved those measures that tells you whether or not the project was successful. And number two, is it done at all? Yeah. Because some of these projects continue to go and go and go, you know, the, the comment you made there about, um, about not having the discipline to actually keep on task really resonates with me. I think that that's a very general thing called focus.

Speaker 3:49:41Focus. Focus. Yeah. I, um, I remember sitting on a panel and this topic was being discussed and I said it's a fairly focused, you know, you can't do a two number one priorities. There's either one or is not. Right. Well, what's interesting me about it, and it was kind of mulling over there with why, you know, I guess maybe it's just an inherent blindspot for people to just get or the opposite of a blind spot. And this is just get distracted by things then to have a hard time kind of resetting and went back and starting again. Part of it is accountability back to the culture. You know, I cut my teeth in the banking world and they had project management down pretty well and there were severe ramifications if you're part of the project or if you have a project manager wasn't getting done, whereas insurance more laissez faire.

Speaker 2:50:37Yeah. Is that true? That's not the case. So where does that come from? Well, I don't like to be held accountable, you know? Uh, that's, that's a great question. Uh, I think it's just part of the culture that they grew up in. One of the things that, one of the fundamental differences about insurance culture and other cultures, which is something that I think might surprise people who aren't to insurance is it so intensely relationship driven. And so I think it's harder to hold people accountable if you have this sort of close relationship with them, you know, captured linguists to what you used for. And, and I think the root cause of that is just that risk evaluation and risk management is an almost inconceivably complex thing to do, but you can't, you can't foresee all the risks that are in the world. And so you have a relationship with your clients and your other counterparties that, that kind of see you through those kinds of things. And so relationships are really inherent in this management business and maybe because relationships are stronger than the technologies,

Speaker 3:51:33it's a risk adverse culture. Yeah. And as you know, uh, and you may have read that, uh, you know, there's been this big push around innovation and then short tech and all that, that's a major disruption to this industry. And those who can afford it, I think we'll, we'll do okay because they're willing to take the rest. They have enough surplus that if they lose money, it's not going to kill them or smaller carriers and mid tier carriers are going to have a very difficult time adjusting to that. We will have to merge or bought or I did a, I get a master's degree paper on a strategic planning and insurance industry. And one of the things I uncovered was one of the reasons that they're adverse to do strategic planning as you could be flushed with money tonight. But if the wind blows Tamara, you don't have that money anymore because you've got to pay it out in claims.

Speaker 2:52:29Yup. So you might not have the money to pay for a big it project or whatever it is.

Speaker 3:52:35That's right. That's right. So they've a very, a johnny come lately as it related to the use of technology. Yeah. They were like one in four flat, one who was a payback. So he got rid of all the payroll clerks because the system can do it. Well that's all fine and well, but we're not gonna venture outside of that. And it's a lemming mentality. The mindset is, you know, if this carrier is the one that, well maybe I better look at it.

Speaker 2:52:59Yeah. You know that maybe that's an interesting segue to kind of move a little bit off of technology into a sort of more recent part of your career. And in particular, the thing that caught my eye the most in one way in which we connect that initially it was your involvement with a invest. Yeah. In your other board of a and best out of that happened. I mean, it doesn't see it. What was the connection there?

Speaker 3:53:20Well, obviously I grew up in the insurance industry and I guess over the years, just through Moses, I gained some knowledge about the, uh, the business. But, uh, I have, uh, a long, uh, a long relationship with am best, uh, years ago they had a technology conference called effusion and as part of that conference they would give out awards for innovative use of technology and uh, I was uh, on and um, I had a number of speaking opportunities with them, but then I became a judge for their technology or words. And then it got into, as you said earlier, the relationships and knowing people and then them assessing you as your ability to do things. And I guess, uh, along with that, uh, you know, the varied experiences I've had a looked attractive to them, especially as it related to, uh, you know, things around technology and human resources. And I think the thing. So I've had, I had the relationship, I was recommended and I've been sitting on a board now it's a four year term and I'm into my third year. So anyway, but it's interesting because they are going through a big change.

Speaker 2:54:38They are. And, and the reason why I brought it up and the reason for the segway was, was that you were saying how the insurance industry moves collectively and there, there are a lot of ways. I think there are institutions whose explicit purpose in some ways is to make sure that people are the kind of, the average of the insurance industry is, is, is a, I'll call it, call it solvent level, right? They invest regulators, they have similar roles, right? Yeah. It's their kind of enforcing a kind of a similarity amongst carriers least in terms of a lot of different performance characteristics, which I think winds up drawing a lot of behavior from them. And it's interesting that you came across him in a technology from a technology judging you know, side of things. And I'm wondering, you know, what, what, um, to what degree, what they think about, about the system's problem and the technology problem and insurance and what is, what, how are you, how are you bringing that perspective to ams? And they must have obviously felt like there was something to it.

Speaker 3:55:36Well, that's, that's a great, that's a great comment because I'm a, as you probably are aware of, their whole thrust over the years has been determined the solvency of a solvency strength companies. And uh, you know, with my perspective, uh, and by the way, there's two outside directors, myself and another gentleman. Okay. And I, he and I align pretty strongly along the lines that, uh, when uh, and best looks at how a company is managed, they, they really just a little backdrop as they have four pillars of evaluation. One is the balance sheet, balance sheet strength, and they run it through what they call them, the car model. And that's just the classic model that runs a number of scenarios and says, well, if this happens, how is this going to affect the balance sheet so that, that evaluates the balance sheet. Then they, then they have a, some practices around the revenue stream.

Speaker 3:56:38Cashflow was, I would call it, but it's really a income and expense. They want to make sure that the balance sheet strength is sustained over a period of time. Then they look at the third pillar, which would be a business profile, you know, where they do business, the type of business they've, right, you know, that type of thing to see if there's any threats there. And then they have an enterprise risk management, a pillar that they manage and that gets into, you know, how, how are you managing, what are, what are the biggest threats you have, what is the value to the organization. So the argument that I have brought to the table is a, I am best really starts, didn't have to look at things like technology and data to determine whether or not a company's going to be able to sustain itself going forward.

Speaker 3:57:29Because as you know, in the older days you looked at three things to differentiate yourself. Price, product and service. Well, it's really dated today because that's what really differentiates a carrier. The data they collect and how they use it because each carrier uses it differently. And, uh, you know, whether or not they're managing and controlling the data properly. And alls you need to do is look at solvency two that has taken place over in, uh, you, you, uh, to see what they're doing around data and what the impacts have been a to see how valuable that is. So

Speaker 2:58:06what, what, what do you think that, what do you think a invests ease or you know, as a director, what do you see as the opportunity for am best if there is one right now because your regulatory sort of the, that they're not regulated by the rating agency world is a funny one. It's a funny business. The weird business where you also tend to have these dominant players in certain regions. So, and best is so dominant United States but not elsewhere. SNP somewhere else. And you have of other, other organizations. So are lurking in the shadows, right?

Speaker 3:58:40I don't know how do, how do I think about that? They have, uh, they have, um, you're absolutely right in your remarks and they have a concerted effort to grow themselves. Global end yet. But the public felt that for a long time. They have trouble breaking in. Yeah, they have. And, uh, you know, part part of the board's charge will be to determine, you know, how they grow or if they grow. Uh, so that's one of the table and it's something that is going to be vetted out. Uh, they are beginning to look at a mortgage loan credits. They started that process, new product against products. So, uh, you know, they're looking at different ways now to do that, but it's something that needs to be vetted out further. And why did they change the model? Was that a board level item as well? Yeah, yeah, yeah.

Speaker 3:59:34The board sets on something they call Mac and that's the methodology advisory committee and any changes to the, uh, uh, methodology for evaluating carriers, uh, goes through that, that group, and then ultimately has to be approved by the board. A car model was basically undertaken to a, make the process more transparent. That's up for debate, right? Uh, do you use a much more powerful modeling? Uh, but primarily what the results are going to be basically the same from a, what's the balance sheet looks like, but they're also using this, the castic models to project out different types of scenarios that they're going to be turning over to the character that this is what it would look like if it was a one in 100 year, one on one, one in 500 year, that type of thing. So it must just gives them more powerful, more information, more data, more data.

Speaker 2:60:40One idea that, that, um, another thing that we've talked about before is because I think a invested is just radically under appreciated as an organization. I think any company that can be so dominant in the marketplace like they are in the United States has to be, has to be respected for, for being that effective. And there's, you know, there's obviously these network effects where once everybody we went to your, you know, in entrenched, it's Kinda hard to dislodge you, but it's still not screwing up is a pretty good achievement over a 100 year time span. Um, you know, do, do you think afs is under appreciated the organization?

Speaker 3:61:13I don't think so. No. No, I don't think so. I think the states are pretty powerful. Yeah. I mean if you want to get a c level executives attention, that's all you need to talk about is an best because they're so sensitive to the ratings they get. Yes, yes. However, if you go across the pond to the EU or UK and Brexit, by the way is a big issue that we're dealing with now because uh, a lot of those laws have not been passed a,

Speaker 2:61:43it looks as if we're going to have to deal with it as thanks, but it's not as appreciated or over there as it is here. I don't think they have alternatives which are alternatives. That's correct. Yeah, that's correct. I'll start running out of time. I'm wondering if, if maybe you could, you could comment a little bit on, on what you think of as the future. I mean, you wrote an article recently on, on the survey, you've done that where, where is, where is technology going into insurance and treat insuretech is all over the place and you know, there's hype there, but I think there's a lot of reality there too. Maybe not as much as the height, but that doesn't mean it's not going to be transformational over time. What do you think's going on? What's going to happen next?

Speaker 3:62:19Right. I'm

Speaker 2:62:20more and more data's going to become available for, uh, for companies to address, uh, you know, I think, I think what we're going to say is, um, quite frankly, those companies who aren't staying up with technology change and data change, are gonna end up being either going out of business or not? Um, that's my cut on it. Uh, I think the insurance carriers are beginning to understand that, uh, you know, technology is more and more part of their lives. They're going to be relying more heavily on artificial intelligence and cognitive cognitive intelligence studies like the things so, you know, robots and things like that, uh, are basically, uh, when horizon. I think so, and I think attesting to that is the fact that they have invested. The insurance companies have invested over $2,000,000,000 in short texts. Yeah, I heard that the protection this year, there's 2 billion last year or something in this unit they think is going to go to 10. Doesn't surprise me. I made it. Is White hot. Well, it's a way of um, bringing innovation. Yeah. Yeah. So it'll be interesting to see how the cultures clash. I right. I agree. I agree. It will, it will. Who's going to, which culture is going to win. I mean, insurance culture is exist for a reason. Uh, and, and the technology culture, I mean. Right. Stay tuned. Yeah, exactly. It's my guest today is bill Jenkins. Bill, thank you very much for joining me. Thank you. Pleasure talking to you.

See All Episodes