Andrew Kramer
This episode of Life Accelerated is brought to you by Equisoft, a leading global provider of end to end cloud based solutions with deep domain expertise in the life insurance industry. To Learn more, visit equisoft.com don't fear the future. Lean into it. Embrace it. It will change the way we write and underwrite business in the future. It could tremendously change the customer experience and the efficiency within the things that we do.
Olivier LaFontaine
I'm Olivier LaFontaine and this is Life Accelerated, a podcast for life insurers who want to achieve digital transformation. When we think about the life insurance industry, underwriting and risk management are at the center of some of the most important advances happening today. In this episode, we're joined by Andrew Kramer, an industry veteran with decades of experience in underwriting, risk management, and life insurance distribution. Andrew's journey spans both reinsurance and chief underwriting roles, giving him a unique perspective on how the industry is evolving. We'll discuss the challenges of underwriting and how M Financial is reshaping the insurance market. And particularly in the high net worth segment, we'll talk about the groundbreaking work being done to streamline the application and new business processes by working and improving standards to reduce duplicate work. So let's get into it. Hi Andy.
Olivier LaFontaine
Welcome to the show. I hope you're doing good.
Andrew Kramer
I am doing well, thank you very much. Glad to be here.
Olivier LaFontaine
Great. What I like to do on the show is start a little bit off topic to get warmed up and maybe start with a bit of a personal note. And I'm told that you've recently traveled to Japan, so you want to kick us off by talking a little bit on how that was. And did you enjoy it?
Andrew Kramer
Sure, yeah. My wife's goal was to run all six world major marathons. She ran Tokyo a couple weeks ago, so I joined her. Obviously we did a two week vacation and really enjoyed ourselves. Did a lot of sightseeing after the race.
Olivier LaFontaine
Good culture shock also, I'm sure at the same time, while a very modern country, rather different than the US obviously. So did you enjoy that?
Andrew Kramer
Yes, it was. The culture was tremendous. Totally different than I had expected. I mean it's, it's clearly a very populated country, but people are all very respectful of their own personal space and things like that. So I really did enjoy that. You would get on a subway and it'd be perfectly quiet. You'd get on in an elevator and it would be perfectly quiet. Nobody was interrupting in the time we were there.
Andrew Kramer
We were there for two weeks. We only Heard two sirens, a fire truck and an ambulance. We didn't hear all the car honking and stuff like that. It was quite a different experience than, you know, from traveling to New York City and other major metro areas in the US.
Olivier LaFontaine
I've never been myself, but we have done business with a partner that used one of our tools in the past. And so we had interactions with team over there. And it's always very polite, very respectful. There was a lot of Olivier San and Jose San when exchanging with them. So really interesting approach to doing business. Anyhow, with that we can move on to the thick of things. First of all, I obviously I'd like to hear a bit of your background. You've done many things in your career in different roles around underwriting and risk management.
Olivier LaFontaine
So you want to talk a little bit about your path to where you are today and what the role is.
Andrew Kramer
Yeah, it did meander early in my career. I started as a Life Reinsurance underwriter for seven or eight years, then moved into a direct underwriting role with a carrier. Went back into Reinsurance, a company that I was with in the early 2000s, got out of the business and I could see storm clouds on the horizon in the Life reinsurance marketplace back then. If you think back then, there weren't many reinsurers that survived. There was a lot of aggregation and a lot of outside European reinsurers came in and purchased those reinsurers that, you know, the U.S. reinsures. So at that point I could see it was going to get very challenging. So I used the skills that I had developed and landed a role in other parts of financial services.
Andrew Kramer
I worked. I went to bank of America in their insurance services operation and became the risk executive of the insurance services operation. I did that. And that particular entity, by the time I left it did about $1.1 billion of net income from its insurance operations. We had lender placed property. We had personal lines that we manufactured and offered ourselves. We had debt cancellation, which was an addendum to a loan agreement that functions like insurance. And then we sold other people's insurance.
Andrew Kramer
So it was just a fascinating experience. I used that opportunity to get more into the risk management side of the business. I work for an asset management company, but then I decided I wanted to get closer to the the actual business instead into a staff function. I wanted to get into a line function. So I got back into underwriting and I went to New York Life in their direct to consumer operation and did that for three years. And then I moved to M Financial where I've been working in the high net worth and ultra high net worth financial planning space. So it's been kind of a whirlwind. I've done a lot of different things but you know, everything seemed to have built upon each other, which was kind of fun.
Olivier LaFontaine
And this is great. A very unique background. So can't wait to hear some of your insights around the industry and in particular around underwriting and the life insurance industry. But before we go there, can you give us a little bit of background on M Financial? Because I think it has a very, the company has a very unique business model which I think you speak about very interestingly. So you want to go through that a little bit and give us some background on what the company does.
Andrew Kramer
Sure. M Financial is a life insurance distributor. We focus on the high net worth and ultra high net worth. We represent six or seven companies. We are comprised of 140 independent agencies, if you will. Some are large, some are small, some might be one or two producers and three or four back office staff, but some might be hundreds of people. So it really does vary. We're located across the country.
Andrew Kramer
M is actually owned by our distributors. We were created by a very innovative and thoughtful group of them that founded the company and now we're across the country. We are unique in that we share in the risk with our carriers. We created our own reinsurance company so we take a piece of every risk we write. And because of that carriers, because we have a reinsurance company, we know the mortality and the persistency. And our compensation program with our producers is essentially our dividend from our reinsurance company. So if they write bad business, it will affect their dividend. So we write really good business and the business sticks around.
Andrew Kramer
It's got really good persistency. And because of that we partner with some of the best carriers in the industry. They create proprietary products for us, we take half of the risk and it's a win, win, win all the way around.
Olivier LaFontaine
So you said this in our, in our preparation. You know, I asked the question why do you not do the life or the policy administration part of it? And your answer to that was that it's because it requires a lot of the regulations and, or regulatory compliance and things like that. And I thought, oh, that's so interesting because you know, you're doing almost everything that's necessary in the business. Taking the risk, evaluating the risks, selling, working with agencies, but you don't have to deal with the regulators. So I thought that was very smart and good way to be doing business in the industry.
Andrew Kramer
Well, it was the same thing we realized in the banking world when I was in the banking side of the business. If you are selling or reinsuring the business, you're really only regulated by one essential regulator. If you're a carrier, you're regulated by all 50 state departments of insurance. So it's a much more complex regulatory environment than it is if you plan playing in a reinsurance role or in a sales role. So you don't need the heavy compliance infrastructure that you would as a carrier.
Olivier LaFontaine
Yeah, that makes a lot of sense. Okay, so back to your current role and you know, some of the things that you're, you're working on and, or have worked on recently. So I think you, we talked about the Accord standard life application, for example. You want to talk about that a little bit and I think you're very passionate about that project. So why don't you go through that a little bit?
Andrew Kramer
Yeah, I'll kind of give you the background and how. I was introduced to ACORD back at that time in 2020 when I first came to M, my boss was actually on the board of directors at Accord and he came from the PNC side of the business. And in the PNC world, both in London market and the international PNC, they've really adopted the Accord standards. And almost every communication they send between companies or between parties within the vertical hierarchy of the organizations, from carriers to reinsurers, it's all prescribed and standardized with the Accord standards. They get great efficiency. If you buy a Pullman auto policy, chances are you're using an Accord application. And they've standardized a lot of the communications between every step of the food chain in the P and C world. In the life world, we haven't adopted that.
Andrew Kramer
And I saw great inefficiency. So my boss at that time back in 2020 said, Talk to these folks at Accord and see how we can drive adoption to gain efficiency within the life insurance value chain. So I quickly realized there were two gaps that I saw that they could help us out with. One of them was the standard application. If you buy home or auto covered, chances are you're using the Accord standard application. And the advantage of that is it becomes a very clean user experience and very homogenous. And carriers can then focus on competing on other things other than their customer experience. And as a distributor or producer, you can focus on.
Andrew Kramer
It becomes easier to sell many other carriers products. Now you could argue that it makes it more of a commodity, but I view it as spend your money and focus on driving greater value to the customer rather than trying to create customer experiences that may be more expensive. But anyway, to make a long story short, I realized with the advent of and the pandemic created a push towards accelerated underwriting across the industry. A lot of carriers had it, but a lot of them had it in pilot mode or they didn't have it completely resourced. But once the pandemic hit, it became very critical to have an accelerated underwriting program. Because if you think back to the depths of the pandemic in mid-2020, when there were all the quarantines and stuff like that, you couldn't get an examiner. Wasn't what was the term used during the pandemic? I want to say critical, critical resource. But you couldn't get examiners to go out to do exams.
Andrew Kramer
So the only way you could write business was non medicine, which was accelerated underwriting. And my thought was, let's standardize this so that we can have a cleaner data exchange across every member of the food chain within the life insurance industry. So we really started working on the standard application at that time and we also realized that there was, if you think back to that time period, it was very difficult to get medical records, attending physician statements. And the reason that it was is because if you think about every doctor's office, they had high vacancy rates but you know, they couldn't get in to do their job because a lot of them, you know, many people were sick. But in addition, anybody that was in a back office role was kind of pushed to the front office to serve patients and manage the front office staff and stuff like that. They didn't have the resources to actually send somebody back to the photocopier to copy the medical records. So electronic health records went from, in this industry, electronic health records went from pilot to photograph, full blown implementation overnight, which is unheard of in this industry. Otherwise we'd still be probably sitting here and every carrier would still be in pilot mode with electronic health records.
Andrew Kramer
But now they've really adopted them. So it forced the adoption of electronic health records. And at that point we quickly embraced it, actually before the pandemic. And I realized when you order electronic health records, you can access not only the PDF that the underwriters would read, but you could actually access the XML or the JSON code. And I started looking at that and I talked to some friends in the industry that were knowledgeable in that area and they said, yeah, that's the healthcare data standards. And as we started digging into it, we realized that those healthcare data standards Were really robust. They weren't created for underwriting. They were created for the needs of the medical care and medical reimbursement or insurance industry, the healthcare insurance industry.
Andrew Kramer
So there's a lot of extraneous information in those files and a lot of redundancy that life underwriters don't need. So we spun up that project to try to narrow down the redundancy in those records and drive more value added data that could be processed more efficiently at the point of sale to get to potentially point of sale issue by using medical records. So those are, those pieces are all starting to come into place right now. Where we stand right now. Let's go back to the life application. It's a project that we've worked on for probably two and a half years. And we pulled in the interstate insurance compact and they were along the whole way. They endorsed it from day one.
Andrew Kramer
And they really helped us drive it to the point now where it's actually approved by the interstate insurance compact. And they're the primary regulator for the life insurance industry in I think 47 of the 50 states, said 48 jurisdictions with Puerto Rico.
Olivier LaFontaine
One thing that I'm curious about, and I was holding up, something was burning my lips. How do you manage the fact that there's differences in the way that different insurance companies will ask questions, like tobacco, for example, or there's some medical questions that are obviously different from one insurance company, which is what to some degree they consider their proprietary sauce, or at least it's the opposite of the commodity, or you mentioned that they don't want it to become a commodity. So those are the types of questions you need to have. So how did you deal with that in the application?
Andrew Kramer
We did receive some resistance from a number of carriers. You know, one carrier told me that they felt their application was a competitive advantage. And I had to chuckle because what I've observed when I lay out in the last five years, most of the carriers, at least the carriers that we deal with, revise their applications and their exams. They all look the same. It's 90% or more consistent across everybody. So you know, how many ways are there to ask, do you smoke? And when was the last time you smoked? Or do you have coronary artery disease? Have you had a heart attack? And then the reflexive question, the reflexive. When we laid all of these things out, there were more similarities than differences. And we tried to adopt the key innovative approaches of the cutting edge applications, the ones that were most recently adopted.
Andrew Kramer
If you think about what's happened with the format of the life insurance application over the last 10 years, the questions used to be very broad. And please provide details below. So they would ask, do you have or have you had any of these following things? Coronary artery disease, heart attack, congestive heart failure, you know, the laundry list. If yes, provide details below. And that was it. And then all of those details were in a text format, which is very difficult to use technology to interpret. Where everybody is essentially gone is, have you had any one of the following? And it was a bullet list of coronary artery disease, congestive heart failure, valvular heart problems, heart attack, et cetera. And then if yes, provide details below it.
Andrew Kramer
But you would actually have an affirmative check mark by each of those items. And then the details were your typical date of onset, treatment and doctor's name. So that's what we've done. There are a couple of other things that we added, a couple other layers in certain impairments, like with diabetes. If you have diabetes, then the question is, what type of diabetes? Type 1, type 2, or glucose intolerance? And then what's the degree of control? And stuff like that? But yeah, we defined all of that and the reflexive questions for all of those application questions. And we also then defined and got approved all of the hazardous avocation questionnaires. Historically, carriers really only approved their aviation questionnaire and their scuba diving questionnaire. But we did rock climbing, backcountry skiing, hang gliding, ballooning.
Andrew Kramer
You know, we did all of the hazardous applications from all of the underwriting manuals that we had access to. So it's a really robust set of forms if carriers want to adopt it, and we have them approved by the Interstate Insurance Compact. What's different now? This isn't the first time we as an industry have tried a standard application. In 2004, I think they created one and it was paper based. The challenge is there was no flexibility. It was take it or leave it. And we've addressed that flexibility. Because the second thing that was a challenge was we didn't have one regulator.
Andrew Kramer
We had all 50 state departments of insurance, plus D.C. and Puerto Rico. Now we have the Interstate Insurance Compact. In the past, if we got the. If we got the application completed, every compliance department of every insurance company said that'll never work. So that people didn't even try. Nobody adopted it. What's different now is we involve the Interstate Insurance Compact and they are now the form approval and policy approval entity for 47 states, plus Puerto Rico, I think.
Andrew Kramer
So we've got their approval so that eliminates that friction. And we also enabled carriers to deviate from that standard. For example, we created more of that super set of questions. And if you don't want to use a question, you can strike through it when you file your form and you don't have to use it, then you can just reformat your form, you know, show them what the final layout will be of that physical rendering. Most of the carriers are building an electronic application first, and that's what the strategy we took. It's an electronic first submission. Each question will probably be on its own screen to walk through or a couple questions per screen with all the reflexive questions. We feel with this approach, it's going to be much easier for carriers to only use the questions they need.
Andrew Kramer
They can strike through the ones they don't. And we created optionality in things like the nicotine question, the smoking question. Some people say, do you use nicotine in any form? Others say, do you just smoke cigarettes? Or variations thereof. And then there's a different look back period, 12 months, 24 months, 36 months, or have you ever smoked? So we've provided all that optionality. We've put brackets around the sections that people don't need to use. They can just strike through the stuff that they don't want to use, use the stuff they do want to use and give them all the flexibility that we can. The only caveat is you can't add any new questions that aren't approved. But what we're asking people to do is take a look at this application.
Andrew Kramer
If you are interested in adopting it, we'll give you a lead role in the governance process to update the application and we'll get new questions added that people want to use. But what we're trying to do is drive for more standardization so that it's easier to send the data to everybody in the life insurance, food, ch from distributor to carrier to reinsurers. And if you can get all that in a standard format, it makes it much easier for everybody to analyze because, you know, the what's going to differentiate the winners and the losers is their ability to analyze the data that they have.
Olivier LaFontaine
I guess you probably also use that standard app then. And you said M Financial will negotiate products or I guess get the design products specifically for M Financial with the carriers. So I assume that standard application is part of that discussion, right? You're doing it on your application so that you get some level of standardization for your agents.
Andrew Kramer
We haven't implemented that yet because we've only got the Interstate Insurance Compact approval Last November. So we at M haven't implemented anything. We have to work with our carriers if we want to implement all of those. And we're starting those conversations now. If you talk to any of the major distributors, they all see this as a huge win. Because when you have multiple different formats of applications, there are very subtle differences that are easy to trip up a producer. And when you miss a question on an application, let's say your replacement question's on the left hand side of the page instead of the right hand side of the page and you forget to fill it out, that becomes a NIGO, not in good order. And that application then gets delayed and underwriting and you got to get it resolved before the application can proceed.
Andrew Kramer
So we think that the early adopters, we feel will have an advantage at least in the independent distributor marketplace. If you've got a career agency that or career distribution system there, there's probably not the upside that you would have in the independent distribution system.
Olivier LaFontaine
And I guess if you combine that with the other project, the electronic health record proposal, or I guess the electronic health record system, then I guess you could probably accelerate quite a bit the underwriting process with those two things combined. Because less NIGO, less errors in the application combined with less time waiting from documentation from doctors in other places, you probably can shave off quite a few days to the typical whatever 50 days it is nowadays for underwriting. Right.
Andrew Kramer
That's our hope. And the electronic health records data standards don't just apply to electronic health records. You know, it'd be great if we could get. So if, when you order like a health information exchange document or an electronic medical record, it'd be great if we could get it right from the vendor in the cord format, de dupe it, remove all the redundancy, et cetera, all the, all the data that an underwriter doesn't really care about. But the other use case for it would be for the medical records summary services that a lot of carriers use. If those medical record summary services could take the PDF medical records and then summarize them and create an extract in that format. Now you could feed the data directly into your underwriting system, triage all the cases, approve all the easy ones, decline all the obvious declines that are tripped, and then re reduce that subset of cases that your underwriters have to look at. That's what we're trying to do.
Andrew Kramer
We've defined the data standards for the top 200 medical conditions. If you then. So for example, there's probably hundreds of potential cardiac issues, but we've probably got 30 or 40 of them that we've defined if we can. And then for everything else, you just put other cardiac condition undefined and route those out to the underwriters. Cause they're low frequency and potential high severity. We should be able to get huge efficiencies in the underwriting space immediately. And then over time we'll continue to build out those data standards for the next 200 medical conditions. And that's where we are right now.
Andrew Kramer
If we can get people to adopt this, then we'll bring people back and scope out Gen 2. Right now it was just Gen 1 with the top 200 medical conditions.
Olivier LaFontaine
And have you been able to deploy that or what's the status of that project at this point?
Andrew Kramer
Well, right now we had our kickoff meeting last week, actually with all of the people that were active and engaged in the project. We started with the project, we opened it up to whomever wanted to participate, whether they're ACORD members or not. We did the what I would call the technical definitions from an underwriting perspective over two to three years to get the final data structure, which were basically all the top two under medical conditions and the risk factors for them. So for example, to define those risk factors, we looked at all the underwriting manuals that were out in the market, all the reinsurance manuals manuals that everybody had access to. And we had members from each of those reinsurance organizations on the team. And we defined that super set of risk factors that pretty much everybody uses when they're looking at a medical condition. So for example, if you've got somebody with diabetes, what the underwriter to evaluate that case, the underwriter wants to know how old are you now? How long have you had it? So what was your age of onset? And then what is the degree of control, which is measured by your blood glucose or your hemoglobin A1C, for example. And then do you have any complications? Kidney problems, retinopathy, peripheral vascular disease, coronary artery disease, etc? Pretty much everybody looks at those same things.
Andrew Kramer
So we just defined all of those in a JSON format. So now if you can either take that JSON or XML file that you get from the electronic health records vendors and either have them or if you've got a team to extract the information that really matters in that format, it should be able to feed directly into your underwriting system to be able to triage some of those cases. Now, the interesting thing is that I do agree with this wholeheartedly. One of the guys on the team was a medical doctor and he used to say, your structured data provides the verbs and nouns, but the textual data provides the adjectives and adverbs. And I do believe that. But with those nouns and verbs you can identify your knockouts and either decline cases or potentially put him on a fast track to approve. But on the complex cases, you need the adjectives and adverbs to get the whole story because underwriters like the story of the case.
Olivier LaFontaine
Right. And I think that's something that comes back once in a while. But the underwriter job and I guess analysis is going to remain. But what we're really trying to do is sort of filter out the things that are easy or can be jet underwritten or maybe sent to a junior underwriting queue and so that we can route the more complicated cases to the more complicated senior folks. So I think that's, that's amazing. And yeah, I have to imagine that coming up with standards and trying to lead the way in terms of as an industry initiative, that's always long and complicated. So kudos for trying to make that happen and make the industry better.
Andrew Kramer
We had our kickoff meeting last week or two weeks ago with everybody that worked on the project. So we defined their medical conditions and risk factors for the top 200. Then a cord took that information and they drafted the data dictionary and de duplicated it and you know, standardized it all and wrote it out and you find all the JSON code from there. And then so we had the first kickoff meeting last week with everybody that worked on the project and I was very pleasantly surprised at how much interest there was. We created a separate membership with ACORD for the data vendors and the some of the technology vendors that really don't have the need to the forms and everything else. It's just the electronic health record piece created a skinny down Accord membership so they can access all those data standards. Because I think that's going to be key to adoption. Because I think a lot of the vendors, whether it be the data providers, APS summary services and the technology vendors supporting the processes, realize standardization is the only way to go.
Andrew Kramer
It's going to drop the price significantly in the cost of implementing everything. So if we can get them to really embrace and endorse this and basically create, I guess you could say, a plug and play version with these standards, it's going to drop the implementation cost for the carriers significantly for sure.
Olivier LaFontaine
And as a system vendor, obviously we really have a vested interest in seeing that happen because that just makes the overall process easier and better. And you know, we can standardize, we can take advantage of those standards to automate certain tasks in the system. So looking forward to that, I can't help but to think about, you know, when you describe health records, summary services and things like that and analyzing documents and lots and lots of information where only a little bit of information is necessary to make a decision. So there has to be a play for a generative AI in that. I'm sure you must have thought about that a little bit. And so what are your thoughts on that and where that could play in the future to support the process?
Andrew Kramer
I think that's going to be the game changer in the coming years. And I think the horizon is much shorter than we all think because it's. If you think about the change that occurs in the life insurance world, it usually takes a very long time. But I think generative AI is one of these things that's going to go like lightning in the next couple of years. We've done a pilot at M to see if we could indeed extract the Accord electronic health record data standards out of the structured data. In our proof of concept, we just did two medical conditions, diabetes and coronary artery disease. It was amazing how much of that data we could extract out. And we only did it one week for diabetes and one week for coronary artery disease in one iteration.
Andrew Kramer
We didn't do any multiple iterations. It was just a proof of concept and it was extremely powerful. We also then used that to try to summarize the case and it was very effective. So I do think generative AI is going to be probably the key enabler to drive all of this stuff. You know, I'm having a lot of conversations with folks active in that market and they're all in agreement that, yeah, we can do that. That's pretty straightforward.
Olivier LaFontaine
Yeah. And then that improves the speed of underwriting again and reduces errors also probably. Do you see any impact on that sort of changing the line perhaps as to what is high net worth versus maybe affluent mass market or whatever you would call it? Because now we could perhaps do this faster. We could perhaps do broader analysis at a lower cost. So does that provide more cost effective products to the, perhaps the lower end of the market in your mind, or that's not really a thing.
Andrew Kramer
Yes. If I kind of look in my crystal ball and I'm not as close as my counterparts in carriers that are actually making these decisions, but I could see a day when cases are triaged by the technology and the key medical conditions are flagged and listed and Then those that can be resolved by the decision engines are resolved. Like if somebody says they have hypercholesterolemia and they can extract the cholesterol and the cholesterol HDL and the ratio and they can resolve it and say no issues, move on to the next one. But then all of those medical conditions they can't resolve get kicked out to an underwriter. I think it's going to change the way an underwriter looks at a case. If you think back to the traditional way an underwriter looks at a case, they like to read the APS or the medical history like you would read an encyclopedia from A to Z and understand the source story, what's the story, what's the evolution of this patient or this customer's medical history. And that's very labor intensive, right? Particularly if you've got a 350 or heck, even a thousand page set of medical records, that's very labor intensive. It can take a seasoned underwriter two hours to four hours to get through the whole thing.
Andrew Kramer
That's not an effective use of people's time. Where I think these technologies are going with the APS summary services and the new tools that they're developing, the system is going to do the initial triage, identify and resolve the issues that it can resolve, and then kick out to the underwriter those medical conditions that are left to be resolved that somebody's got to look at. And they'll probably flag the pages that they need to read or the snippets that they need to read to really assess the risk. And then all they have to do is resolve each of those issues and prove the case. That's where I see it going now. What has to happen to get there? Well, number one, management needs to enable that to happen. Because the first time you tell an underwriter, we only want you to resolve the things that we flag for you. But then you ding an underwriter for missing something that wasn't flagged, then you've got a problem.
Andrew Kramer
So I think the technology really has to evolve so we feel confident we're not missing anything. But then you also then have to have management that's willing to provide the guardrails and the air cover for the underwriters. So I think there's a lot of cultural change that has to occur before we do that. Give you a funny story. We were experimenting with one of the APS summary services and it was a 350 page set of medical records and we sent them to get summarized and the summary unfortunately was like a thousand pages and that's because it created additional summary pages plus all the cross links and stuff like that. It was just a very dense piece of work. But if you looked at the outline and you clicked on only the things you needed to read, you could get through pretty quickly. But we sent the thousand page summary to the underwriter and didn't explain how to do it.
Andrew Kramer
And the underwriter felt compelled to read all thousand pages. So it was instead of a time saver, it would, it took three times as long. So we do have to be thoughtful on how we implement this and provide the underwriters with clear instructions and error cover if they operate within those instructions. That's going to be the, I think the biggest challenge. Build confidence in that whole process because.
Olivier LaFontaine
I think that's one of the challenges in this modern age of a lot of data available for any process. But then sifting through that data can be very time consuming. It can be net negative in the underwriter in this case, but almost any job can be overwhelmed with data. So that's where I think AI can help a little bit with that. But if we switch over to, I guess to the sales side or the advice side of the business. I've heard you talk about the ability perhaps to predict which informal in your world, right. You do a lot of informal cases because those are high net worth cases and so they're expensive and people want to shop around a little bit to see who's going to have the better rates. So you talked about perhaps AI could help with that also a little bit.
Olivier LaFontaine
So do you want to explore that a little bit?
Andrew Kramer
Yeah, that's something that we've been kicking around because unlike a carrier, where a carrier is a little bit constrained by some of the state laws around use of AI and decision assistance, you know, we're not actually making an offer as a distributor. All we're trying to do is predict which carrier is going to be the most effective carrier in the underwriting. My view is I want to avoid sending a case to a carrier if that carrier is not going to be in the running or they don't have a product fit, their underwriting isn't as good. I would rather only send it to the two or three carriers that are the most effective at underwriting that particular type of case. So yeah, we did a pilot, an AI pilot where we tried to predict the carrier's offers and we had in Just Gen 1, we predicted 25% and 29% accuracy of the two carriers that we piloted. The challenge is you need hundreds of thousands of data points to effectively build the models to predict accurately. We don't do that volume at M Financial. So at that point we proved that it can be done and we're not constrained by the regulatory environment like a carrier would be.
Andrew Kramer
So I think some of those new tools and the modeling is extremely effective, but you just need a lot of data to drive it. That's where I hope to go to help reduce the cost of underwriting across the entire value chain so that companies like M that do low volume but high premium, we can be more effective and lower the expense burden on our.
Olivier LaFontaine
Partner carriers by submitting less application that are unlikely to be accepted. Right. And perhaps Jenny, I can help there as well to sort of read we've done here at equisoft a PoC on reading through the underwriting guidelines that the carriers produce. So that could perhaps be another area where if you combine that with your machine learning that you do with the actual decisions, then perhaps eventually we get to a place where the prediction level gets a little higher up, closer to that 80 to 90%. Who knows, right.
Andrew Kramer
We're experimenting with the AI tools. We've created a field underwriting guide to guide our producers so that they don't illustrate everybody preferred best. You know, we would rather have them really give an accurate estimate to the client to set expectations. They don't have to go back and resell it later when they get an adverse action. But I want to use AI as the user interface or as the front end of that so it makes it more efficient for the producer or the back office staff assisting the producer to get to the information they need through an AI interface. And it's an incredible tool for that.
Olivier LaFontaine
Yeah, no, that's so exciting. That's all great stuff and really looking forward to see what comes out of AMP Financial and other businesses looking at how to model risk and reduce the inefficiencies in the life insurance process. So this was a great discussion. Thanks for your your time, Andy. And I think we can wrap it up for today unless there's you have final words that you'd like to say or comment on.
Andrew Kramer
I guess final words would be don't fear the future, lean into it, embrace it. It will change the way we write and underwrite business in the future. It could tremendously change the customer experience and the efficiency within the things that we do.
Olivier LaFontaine
100% agreed. Yes. Thank you very much for your time again and this was a pleasure.
Andrew Kramer
Thank you, Olivier. Take care.
