Unleashing the Power of Data: How Standardization Can Transform CX in Commercial Insurance

Dive into the transformative power of data for carriers and brokers who are shaping the future of the customer experience. Discover how the integration of data and analytics drives client acquisition and retention. Learn how data standardization can benefit wholesale brokers, agents and clients and stabilize pricing and coverage in the commercial space. Don't miss out on this game-changing session – it's time to redefine CX.


Transcription:

Valerie Turpin (00:08):
Good Afternoon everyone. Thank you very much to attend our presentation. So we've heard a lot this morning about the technology, the data, the analytics, and a lot of example were related to personal lines and auto and we are working on commercial insurance, commercial property for the large accounts. So there's a level of complexity with regards to the environment and the macro system that we are working on that is slightly different than some of the example we had this morning and we wanted to discuss about it. And we also wanted to discuss how that can translate to how we can enhance our customer experience and what is the current customer experience with the level of analytics that data that we have to deal with when you deal with very large clients. So I'm going to start for a brief introduction about everybody. Chris, would you want to start?

Chris Carlson (01:14):
Yeah, Chris Carlson. I work for CRC group, A wholesaler. I'm the Chief Operating Officer and the Director of the Property Practice nationally.

John Vedder (01:23):
Hi, my name is Chris Carlson. I am with Bridge Specialty Group, Bridge Specialty Insurance. Brokerage is the division I'm a part of my responsibilities are for all of our open brokerage property and casualty teams.

Leandro DalleMule (01:36):
My name is Leandro DalleMule, I'm the General Manager and Global Head of insurance for Planck. And we're an AI go figure, right? Have you heard about AI today or No, I'm about an AI company in insurance.

Valerie Turpin (01:50):
Thank you. And I'm Valerie Turpin, I'm the Chief Underwriting Officer for Property for Arch Insurance, which is a global PNC companies belonging to art capital. So my first question will be for John. John, how have you integrated data and analytics in your engagement with your customers?

John Vedder (02:10):
Thank you, Valerie. I will start with ultimately for us, the customer experience is absolutely key, especially in the wholesale space, being a broker, one of the items that we always talk about is the experience our customer has on this account, predicates if we're going to get the next opportunity. So we really want to have that be a good experience and be focused in on that. What we look at, when you look and define the customer, or at least the initial customer that you think of as being a wholesaler, you think of the insurance agent themselves, right? Helping them write an account, what that might be. So what we have experienced in this space, as you look at and you can watch, look at the reports that are released by the different states. There are surplus lines. You'll see the increase in the business that's there and it's, it's actually almost all, you really see it in the transactional space.

(03:03):
So we're seeing that happen with agents that don't know what's going on in the insurance or don't know what's going on in the ENS space. They've not experienced that or they have very little experience with it. So what we've found out is that that turns into a real challenge for our more experienced brokers because our focus with our customer experience is to focus on bringing the experience to the opportunities that are on the desk. So our teammates who have the most bandwidth are our least experienced teammates, our experienced teammates, they are filled up and they're completely covered up with all of those actions and other accounts they've got on their desk. So we've developed and worked through an internal referral tool is what we have. So it's based off of a line of business, class of business states and helps those younger teammates figure out, okay, who should I go and talk to? Who can I get some information from to start making that connection so that they then can go to the right markets, go and talk and get the right information in order to make the placement and help with that customer experience, the agent.

Valerie Turpin (03:59):
Thank you John. Next question for Chris. Chris, the ENS market is very competitive. So how do you use analytics to differentiate yourself in this commercial insurance environment?

Chris Carlson (04:14):
Yeah, no, great question. Similar to John's point, right? It's around predictive and prescriptive analytics. So predictive being exactly what John outlined, what should the purchasing process look like? What analytics can we deliver around market behavior, occupancy, statement of values, all types of those things. Secondly, on the prescriptive basis, how can we then be an advisor to them? I think if you think 10 years ago, wholesale was kind of a pejorative term, just throw this risk that nothing can happen in the admitted market. We'll give it to these guys. We're not really experts. I don't even think specialty might've been a word 10 years ago. We might've invented it just to make ourselves sound smarter. But that's really, really the analytics come in. Can we get deeper dive on the risk? Can we talk about actual flood zones? Can we talk about fluvial risk? All the things that are going on now, SCS Parametric really delivering insight and analysis and also not to appear that we're just selling another product just to get another sale and some more commission.

(05:06):
Hey, buy a standalone deductible buyback, buy a parametric. There's really analytics and data behind that. So that's a main way we're using it to differentiate ourselves with our insurer and our retailers. And I think as well with the markets, it's a partnership and I think a lot of independent agents might gasp that their wholesaler views themselves having three customers, their insured, their retailer, and their market. But we're really trying to build a rapport with our markets so they can deliver profitability and be there to pay claims when you need it. So we're using that data as well to try to be a differentiator with our market relationships.

Valerie Turpin (05:38):
Yeah, very good. I'll go back to John. So all the time I'm hearing that our common customers, they want to understand how we benchmark them against their peers. They want to understand how we consider their information for the pricing. They really want to have a better visibility about how can they get access to the capacity and what determine their pricing. And that's not easy. So what are the main roadblocks, roadblocks that you see into delivering those requests to your customers?

John Vedder (06:19):
And I'll speak about this not in relation to the large accounts, the risk management space. This is really about what I was touching on earlier, those transactional accounts, right? A lot of these have become extremely complicated with the market. They don't know what's going on or they've not seen a multilayered policy on some of these ones, which is sometimes surprising for us. But I think it's one, what we have found is a huge portion of our business is about educating the agent and thus from the insured, the information that is needed in order to get the best pricing to know and understand what the true exposure is for the risk that's presented. So for me, from when I look at it, there's really one roadblock, but they kind of two that go together and it's the accuracy of the data that is received. If you think about the process for us as a wholesaler, we ultimately start with our business.

(07:12):
We're waiting for the information from the agent. They get it from the insured. The insured. They may not have been any part of that building when it was built. They may not know anything about what was going on. They may not know the actual straps, the straps that are there for the hurricane, what type are there when were the actual updates. So they work as best they can to get the information. But I'll tell you, and I think it's a good point for us all to think about, the insurance business is not insurance in the vast majority of the case. So they don't know these things. They don't know and understand how this can impact them. So for us, the accuracy of the data, but really it's the continuation of that information that I think about because we're all looking for it. Each one of our markets are asking for that information to put into a modeling system. Once it is a new account for each carrier, they have to gather that information again and verify the information that's come in and all those different steps. I would tell you that is a big hindrance and a roadblock in just the process that's required to verify the accuracy of the data.

Valerie Turpin (08:15):
Absolutely. That's also on the side of the insurance company, something that we struggle with. So we speak about large accounts. So we speak about companies that are multinational footprints, they have multiple buildings across multiple states, and we do rely on getting and gathering information from them to assess the exposures. And it's not standardized because we are on the ENS market, which mean that they are the market where the insured who cannot find their capacity and their needs on the admitted regular market, they go to the last resort. So everything is actually very tailor-made and difficult. So we have to find some ways to compute all this data to have a better assessment, but also to be able to benchmark between different type of risk and the level of accuracy of the data is always questionable. So a lot of issues. So that's why we do have Leandro with us. And that's a good introduction, isn't it? Yeah,

Leandro DalleMule (09:23):
Great. The bar is really, really high.

Valerie Turpin (09:27):
So we collect a lot of data on our side as well. But I think you would agree that today we have the data and they are purely a description. And I read your website and you said that you want to move the insurance company to the generation of actionable underwriting answers, and I do like it very much So how is the market moving to that? Are you pushing fast enough? And also why do you think that we'd not move that fast into this direction?

Leandro DalleMule (10:03):
Well, we're pushing, but we're not an insurer, we're a vendor. I wasn't before becoming a vendor or AIG for many years, so I know the struggle. Just let's do something more interactive. Let's see the difference. I've heard, I've been in many sessions today and everybody seems that every insurance company is doing gen ai. That's where I felt like, but just a raise of hands here, how many of you have used any gen AI ChatGPT and the like? Have you ever used chat DPT? Okay, that's pretty good. Now, how many of those in your companies have a gen AI ready solution that you use in your companies? That's a more likely number. Yeah, the numbers start dropping and now think about get your answer. What is different this time that we believe? If I ask how many times or how many of you actually use big data to build a model ever want? Oh, there's actually, it's a bio sample, but okay, so the number kept dropping, right?

(11:14):
Everybody heard about big data, we heard about blockchain. If I asked blockchain, maybe nobody's going to raise their hand, but even me. So what we see this time that is different is these tools are closer to us. They are enabling the user, us not necessarily a programmer, to actually tap into that technology. So by tapping into that technology, it would allow us to explore that data in different fashion. Because if you were back in the still big data is a thing, believe it or not, but back in the, let's say 10 years ago, yeah, big data, big data. But you need to program, you need to actually know how to access that data and create the machine learning model or just a standard regression to explore that data today. And this is why so many hands were raised when I said, have you ever used this technology? And everybody say, yeah, that is how we believe this and this is why it's different and how we believe that this time the industry I'm talking about seriously, it's my daughter's ring.

(12:26):
She was programming in R earlier today, by the way. She needed some help, believe it or not, speaking of programming. So that's how we believe that this time things are different because the technology is much closer to our hands, to our user hands. Now that being said, and we'll talk more about it later, we're still in insurance and I'm being very cynical when I said that. I heard that apparently everybody, every insurance insurer in here today is using gen AI actively. I'm surprised. We are an AI company. We've been an AI company for seven years before AI was this cool. And we are not that sophisticated. So I'm surprised and skeptical because we do have challenges as an industry and there are many, we'll talk about it later, but I think that's what we believe. It's different this time being still cynical, to be honest. I've been in those right, more than enough to be in that hype and then the hype ends and nobody talks about it. But I think there's hope this time.

Valerie Turpin (13:34):
Very good. Chris, we speak a lot about the quality of the data. I think that when we prepared, you spoke about the single source of truth. So how do you overcome those difficulties? And if you can also give us your opinion about the single source of truth.

Chris Carlson (13:58):
Sure. Overcome the difficulties. I go to insurance conferences and join panels and talk about the difficulties with information and data and analytics. No, I guess for the single source of truth, it is something John alluded to regarding accuracy of information. To me, I spend a lot of time on the underwriting side and all of us in this room, purchase home insurance, car insurance, it's about the facts being in dispute, right? When you purchase car insurance, you don't really think about reporting your car as a four door, black Honda Accord, all these features. The insurance companies have all that data. It's certain it's up to them at that point to make their underwriting assumptions about your performance and trend it and the efficiency of their capital models. If you think about the commercial property space, it's a space of obfuscation, right? Is the building 10 stories?

(14:45):
Is it eight stories? Is it really on this block? What are they doing in there? So a majority of our time and inefficiencies wasted both from the carrier side and the distribution side as well as on the insured side of arguing about facts that should not be in dispute. So I think the first way to kind of turn the industry and AI will be a big part of that is finding a way to take the facts argument right out of it so that we're all dealing with the same fact set. And then it really comes down to talent, efficiency of capital, operational efficiency, things like that. The difficulty in the industry besides me showing up at these events, it's really educating the ultimate insured and retailers in our case about the value of data and analytics as well as kind of the enhanced view and the distribution side.

(15:28):
Our struggle is what's the ROI of us providing a retailer additional information around their risk. How do we quantify that? We tell you your risk is really subpar and is exposed to tremendous amounts of wildfire and now you're in the marketplace and we've told Valerie, Hey, this is a really brutal wildfire risk and we're honest with her, we get an adverse result from Arch. So the incentive on the distribution side is to kind of be hide the truth potentially. So working to educate our teams, we're to educate the insureds about actually presenting the true view of the risk and then transferring that risk at the cost that the market will bear there. John's point was spot on. You could probably draw a graph around age in the industry. I won't say actual age of who's embracing those types of things and who's willing to have those difficult conversations. It is a challenge and markets are just as guilty as that, right? We will lose a deal maybe to another firm, not me or John's. We're upstanding great individual firms that'll just say, Hey, I can knock this down to unknown and I know this carrier will do it or this MGA will do it, and then ultimately we lose revenue for doing the right thing. So we're trying to balance that altruistic view I just laid out with the actual application of it in the marketplace.

John Vedder (16:38):
Well, if you don't mind, I'd like to add something to what you're talking about there. We are talking about the customer experience and ultimately as I mentioned before and Chris alluded to it, is we are all actually beholden to a single customer and that's the insured. So if you start thinking about the experience that Chris is talking about with the distribution teams or being able to manipulate some of the information that's there in order to write the account, think about the challenge and the confusion that occurs for that customer post inspection. When the information comes back and it changes completely and a carry has to get off the account, that's an impact for us as an industry that the single source of truth or that accuracy of data could really help eliminate that in addition to some financial costs that we also incur on every single one of these policies, every single account that comes in the door, especially as we all in the process from a wholesaler or a distribution side, as we start to look at finding third party data, how do we differentiate ourselves? That's a cost that we could end up reducing out of our business and thus also out of the replicate out of the carrier side that's included in the cost of doing business right now that could be eliminated and really ultimately help the insured for them day to day.

Chris Carlson (17:56):
And I think the last part you've mentioned inspection, it's also the claims experience, ultimately the product we're selling and if we're not transparent at the time of the transaction around the exposure, around the valuation of the building since we're talking property, it then results in a really adverse claims experience. The insured's not educated, the retailer kind of passed the buck and everybody's angry at the moment. They need everybody the most and probably the retailer's going to lose the deal. We're going to lose the deal on BOR. Valerie, if she was the carrier would end up losing it, arch loses it and we all lose and we kind fell down in the process. So that education process to kind of take away the unexpected outcomes and those challenges and difficult conversations when ultimately what we sell comes to bear at lost time,

Valerie Turpin (18:38):
I totally share these two because we do use data to assess our exposure and that's how we decide how much capacity we're going to make on a risk. And think about that they are large accounts, so they are multiple carriers putting different capacity on the same risk. So the placement look like the English called them a mud map because they're blocks of multiple carriers on that. And we all have different conditions. So we do use data, we all agree that they are incomplete and they are not necessarily reliable. But as a carrier, I can't really deploy any capacity on this basis. So I need to have a certain level of confidence so I know what my exposure would be because we are speaking about very large exposures and very large limits that we can put on a single deal. So we do use a lot of vendors to help us with the data augmentation.

(19:36):
One of the difficulty we found is that the source of information that these vendors found can be a task assessment, can be multiples, and there's not really a single source of information that will be complete for national and international risk. They may be really good for the state of Connecticut and be very poor for the state of Arizona. And then they are almost a single vendor for each piece of information that you need because otherwise they will have to go to very multiple different type of source of information to collect everything we need. And there are thousand of data for a single location and we are speaking about larger countries, multiple location. So that's very difficult. There are some providers that or aggregators of data, so they give a platform, they give geo geospatial V view and they plug different vendors to it and we are using some of them.

(20:43):
One of the difficulties is that there's no holistic platform. So each vendor provide a partial solution, which means that we still need to have and accumulate different platform, different vendors. So we are really into a kind of struggle to find a single source of truth for our information, but also a complex source of information for us. So those are the difficulties that the very large accounts and the commercial insurance have because our insured not necessarily actually have the information, not the people we talk to. The head of insurance of a large group is not the person who is collecting this information and rely on multiple people at each of their facilities to provide the information and they don't necessarily use the same format or understand what the information will be used for, which mean that they're not necessarily complete.

Chris Carlson (21:40):
I was going to say the only other thing I would add, the ultimate difficulty is in my experience, insurance is the only thing every client expects to pay less for every year. They don't really care about getting it right, they care about paying less. So really educating ultimately the insured on why it's worth to pay maybe a little bit more for that right cover. And again, that kind of falls to us on the distribution side to be able to educate the ultimate end buyer on the benefits, the pros and cons of what we're ultimately selling. But I just struggle with that. We struggle with that.

Valerie Turpin (22:09):
Yeah, definitely. And I didn't speak about the price in my struggle because as a carrier I probably have more exposure if I miscalculate my capacity, then if I miss the premium, I should not say that in front of brokers. But if I meet the premium by 10%, I mean that's not the same financial impact if I miss my capacity by 10% because they are much bigger numbers in those cases. So based on those difficulties that we have explained, we do have some struggle because nothing is standardized. So I'm going to go back to Leandro, our source of solutions.

Chris Carlson (22:55):
You can fix everything

Leandro DalleMule (22:56):
Clearly, clearly.

Valerie Turpin (22:58):
So what exists today that we haven't found yet, but also what do you think will exist using Gen NI and AI and big data that we will be able to rely to solve our problems? And how fast do you think that this solution will be available on the market?

Leandro DalleMule (23:18):
Yeah, all easy questions. So a couple things and I'll remind us, right? We are insurance. So it's part of what I heard. The systems that we work with are still old, right? There are many 40-year-old systems, 40 plus year old systems out there. They're still pumping data. So they're generating data. That's one of the problems. And quick story, before this job when I was working for an insure, I already said who, but let's forget about, it's not a good example, I'm going to give that, but their commercial auto book, I remember talking about data ingestion and the data and so I was like, yes, a new head of that line of business came from another insurer and said, oh, I need a dropdown button here for Garage State. Very simple. You want to know where the car is? And I look at her and I said, yeah, we can't have dropdowns.

(24:16):
She goes, what do you mean? I'm like, yeah, it's a 40-year-old assembler base system that wasn't a concept of a dropdown button back then. And she goes, but how do I know what? I'm like, you don't, right? And so that was just a few years ago. She laughed by the way, she lasted one year, true story and then she laughed. But nowadays, back to your question, what we've seen is, and I'm not talking about plank, I'm generally speaking, is there are solutions, there are cloud-based that necessarily doesn't require a carrier to kind of a decommission a whole system and then go through a full implementation of another system. And then because the problem is there is this spaghetti, right? Most large carriers or old carriers have grown through acquisitions and you see the data, if you see the data flow chart, right? It's this true spaghetti.

(25:14):
You cannot pull one string, otherwise everything falls apart. So what we've seen now is these cloud-based solutions is that are allowing those companies even beyond insurance, but especially insurance and banking, to start developing and actually putting into production the systems without necessarily going to decommissioning every data warehouse they'd ever built or even those 40-year-old assembler, but run it in parallel and then over time, of course retire them, right? It wasn't just a few years ago, I'm talking about five, six years ago. When we have that problem, there's no dropdown. So now you can move to the next level. The second part of your question is how the data is being consumed. And it goes back to my previous comment before, you have a lot of data. We have a lot of data, but then you have to standardize data to canonize it to create a models, and you're rolling your eyes exactly, you've done it.

(26:20):
It was so painful and you had to do it. And then you brigade the model and then the model became stale and oh crap, I need to do it all over again. And then you had to deploy that model and the systems would. So now the benefits of gen ai, which is a very general term, so even a more general AI is in my mind is not so much the ability of giving us the responses or sounding like a human, but it is really the ability to interpret data, unstructured data, because a lot of this data that we think it's structured, it's not really because we don't have a standard, which is the whole subject of this panel. It's not a standardized data set. So that especially gen ai, but AI capability allows you to then more easily put that data in and have those foundational models that those very big companies are building such open AI and Google and meta to eventually create the solutions or the models or the answers you're trying to get from that data. So that's where we see this going.

Valerie Turpin (27:33):
So you touch on the lack of standardization and the fact that AI will be a solution if it's structured data and today it's everything except structured. So who should standardize the data? And I'm turning to my two brokers here because they provide the information to us. Do you see John initiatives on the market about standardization of the data and who do you think that should lead the standardization of the data?

John Vedder (28:05):
Well, the second part's a great question. I don't know if I know quite the answer there, but I will tell you at least the first part of the question there, the standardization. Do we see those initiatives? Absolutely. They're absolutely out. There are a ton of companies that are here that are talking about pulling that information together. I think the challenge that Valerie touched on is something that we do need to be cognizant of is that each one of the companies has a unique data set that has accuracy in a certain specific area. So you still need a collection of them all. The unique part that I see is one is while that focus is heavily on the carrier side, which is critically important to understand the exposure that you're putting at the exposure that you've got now for the balance sheet of the company, you want to make sure got the right information or as accurate as possible.

(28:55):
I would venture to say that even within the model, you're going to have some factor four accuracy, even though you spent a ton of money on a whole bunch of accurate data that's there. My challenge or the piece I would put out there for us is a group is I feel like we're probably missing a focus point and that's actually accuracy or the standardization and availability of this information with that end user, the ultimate customer that insured that's out there whose business is sitting in that building every single day, who sits and looks at it, who actually it does affect those dollars, not only the premium dollars that hit them, but the deductibles that are there are more importantly the impact that occurs post loss. So you start looking at the impact, how quickly could they get back up all of those different items, the impact and the knowledge that is out there that is developed on every single account. You start thinking about how many times you guys do all as a carrier, you're pulling that third party data, but then you still go out and inspect the property, right? You still go out and see it again, right?

Valerie Turpin (29:59):
Yeah. The past buying inspections is a key validation process because we make a lot of assumption and in the ENS market, it's possible to put some warranties. I'm giving this code and this capacity subject that what you told me is real and I'm going to go there to check it and if it's not true, I will give you a notice, but I can exit the risk. Nobody want to get there. It's very uncomfortable for all of us. So we don't want to go there and we find solutions, but it would be so much easier that the pause behind bind surveys is only limited to very, very few cases for very, very specific situations and more to gather an additional level of information but not a validation. Agree.

Chris Carlson (30:49):
I think when you are talking about standardization of data, if you step down from the large accounts, and I think this is what you were alluding to, is the data around those smaller risk and obviously we're pro ENS since we're both wholesalers and Valerie participates in that space has grown. I don't think the world's getting less riskier, but there are those smaller clients that don't want to deal with kind of the post buying inspection process for you to it to be profitable for you guys as a carrier and us as a distribution partner that has to be low touch, no touch business. So I think the AI part and this culmination of data in a way to build capital models so that we don't have to deal with that. We don't have to have this back and forth the source of truth. Let's build commoditized products using gen AI and these massive unstructured data sets. Because if we're talking about improving the customer, the end user's experience, that's really where we got to go and kind of get away from this bespoke underwriting that's always going to exist and there's always going to be that kind of opinion assumption part, but there's a huge segment of the business that's coming into the ENS specialty channel that really needs to be commoditized using large data and assumptions.

Valerie Turpin (31:52):
Absolutely. So I'm checking on the time and I will start with Leandro. So with some of the solution that you are presented, I what could be for commercial insurance, the perfect customer experience of the future,

Leandro DalleMule (32:19):
I'm glad they're going to try to answer that one as well. So you mentioned something right about insurance is maybe the only one that maybe there might be others, but where the customer expects to pay less every year, right? I would add to that before working in insurance, I working with mortgages, subprime mortgages and mortgages in insurance. I mean when I thought about it, and as you think about it, it's not something anybody wants to buy. Let's be clear. And I heard the marketing chief marketing officer of the bank was working with mortgages, said, oh yes, I'm going to make this mortgage experience the buy experience. Great. How can you, the people want to buy a house, they don't want to buy a mortgage, right? A mortgage is a means to an end insurance is the same thing. I want to buy my motorbike. I don't want to insure it.

(33:15):
I have to, but I don't want to. So I think to answer your question when I was thinking about this is I think the perfect customer experience for us is one that the customer doesn't realize is actually or doesn't have to get involved the least that we need to be involved in something we don't want to buy, but we have to. Let's face it, the better for claims similar, let's make this claims process as invisible as possible to rely on that. You mentioned one example about the property expansion inspections post buying. Well, how about if you don't need to have them so you can leverage the data that is out there to just skip that. So if I am insured, I'm not going to experience that. So in my mind, customer experience for us is like no experience. Just make it very easygoing. You've got to buy it. Let's put it out of the way. Go get your car, go get your property, run your business. But don't have to worry about a BOP policy or a property or a flood or a fire. Nobody likes that. So that will be the ideal and AI could help with it. Again, automating a lot of it and more and more in the future.

Valerie Turpin (34:41):
Very good. John, do you want to take it

John Vedder (34:44):
Happy to take it? I think I can expand a little bit on where my mind goes. Deandre was talking about the mortgages, right? I think that's a great place to think about the pain that the insured thinks about for them. They've gone through, they're required to have insurance on these because it is somehow collateralized for that mortgage. They already don't want to pay for it. They already trying to run their business. They want to go put more investment, hire new people, but they've got to go handle the insurance. And if we think about the experience for them currently every single year we're asking for updated information that they may have a new carrier that they're a part of, they're looking to lower their cost as we touched on. So they could be manipulated by an agent or a wholesale, somebody, a part of the distribution channel who's willing to say, Hey, I can get a lower price.

(35:29):
So in order to do that, you got to go gather all this information. It becomes extremely painful for us. I think if I look at the future state, it would almost be an ideal situation that we would know what the cost of an insurance policy is for physical property location based off of just the data points that we know to be true and accurate that are held in a single source of truth. If you think about that, you put in an address, you put a geo code, you know what the physical characteristics of that property are, there may be some minor items that you might need to know about upkeep, but I imagine that's a very small portion. So you'd need to know what type of business is actually being operated in there, but when the straps were put on, when the building was built, when the updates occurred, because you can see that.

(36:14):
And also you have verified that at some point in time, the challenge again goes back to me is that customer experience, that end user, the person actually writes that original check that flows through the whole distribution channel, make it as easy as possible for them to know and understand a product that actually helps them tremendously. Should something ever happen. God forbid it does, but you got the right product in place, you got the right coverage in place and you got a carrier that's behind it that can help them replace their build and replace their business and be able to continue on moving our economy and our society forward.

Chris Carlson (36:49):
Not a lot to say no. I'm probably supposed to disagree and tell you I have a revolutionary idea, but no, you're both right. I think on the way to the invisibility part of the transaction, expanding upon what John said, you're looking for that kind of seamlessness and that comes with the predictive part, right? So if you're the insurance purchaser, usually answer to a board, maybe you're chief counsel, depending on how sophisticated your organization is, they're just like us at the end of the day, which is kind of funny. As you make your way through the industry, they've got people to answer to. They're looking for kind of certainty that you as a distribution partner know your marketplace. The pricing. So the predictive analytics that we're building on the distribution side across most of the major firms is meant to do that on the way to where you think we're going to get, and I kind of agree with you.

(37:31):
So continue to invest in that, a certainty around truth, around wildfire exposure, building exposure, and then ultimately to the end experience where you don't even notice it. I do think that scares a lot of people that, oh, well, it's just going to be machines and AI and me, Valerie and John running a massive company doing hundreds of billions of dollars of premium. We don't really need underwriting talent and we don't need distribution talent. And I completely disagree with that. I do think there'll be some disruption and gone are the days of an A IG, would you set up this massive company and all these underwriting assistants and all this kind of layers of management notes. It's going to be really efficient capital and people are still going to have to distribute it and educate it and look at forms and build capital structures. So I think that's kind of more the nuts and bolts of how we get to the two things they alluded to. But

Valerie Turpin (38:17):
I do like that very much as a commercial property on the writer. They are physical and tangible assets. So we know where they are. We can see them on Google Maps. So if we do have a drone going over and if I can have all my information that can populate my models and everything, which will make the job very easy so I know what the potential of the losses are, what the primary is, then I can focus on something else. I would rather spend some time with you guys to discuss about what the story of the client, I want to know their future investment. I want to know their future growth. I want to know their business changes, which is a conversation we SEL have because we're so focused about what the age of this building, which is really necessary, but unfortunately is not the most important part of the assessment.

(39:09):
We want to project to go a little bit further. And if I do have AI to assist me also very selfishly, what I would like to do is to, because of the better measurement, I would like to have more certainty between what my model told me, what the loss will be and my actual losses are. So we are not in the business of volatility. We are on the business where we want to narrow the gap between model and model losses and actual losses and all of that. Start with the data we ingest in our systems and the more inaccurate data we add on each single risk, they just a accumulate at the end. And it's never good for any company when you have a difference between what your model would be. Just we have a little depression now in the middle of the Atlantic, which is going to be letter B. So right now I've got my cat analytics that I'm called this morning. I said, how much? And they said it's a little bit too early, Valerie, we have to wait a little bit. I said, okay, tell me tomorrow. But I really want to have certainty around those numbers. So I'm hoping that the solution, the standard decision and everything will help us to achieve that, which is also to the benefits of our shareholders and the benefits ultimately to our customers. Thank you very much. I really appreciate.

Leandro DalleMule (40:32):
Thank you.

Chris Carlson (40:32):
Thank you.

John Vedder (40:33):
Thank you.

Valerie Turpin (40:33):
So we do,

(40:38):
We have three minutes, 48 seconds for anybody who has a question. Any question. Okay. Everything was very clear. Thank you so much.

Chris Carlson (40:52):
Thank you.

Valerie Turpin (40:53):
We are still at the conference. Very happy to meet anybody and have further conversation. Thank you very much.