Property Intelligence at Scale: How Geospatial Data Is Redefining Insurance Underwriting | Izik Lavy and Jacob Grob

Speaker A: Foreign.

Speaker B: Do a great job. If you drop a picture in doing some analysis, that looks good, looks decent, but it is still a probabilistic answer. And when you look at those probabilistic ANSWERS, there is 20% of the time or more that it is a hallucination and the answer looks great and people want to trust it because it sounds good, it's convincing, but it could be total nonsense.

Speaker C: Hello and welcome. My name is Jay Carding and you're listening to Making Risk Exploring the Ecosystem, a new companion series to the Making Risk Flow podcast where we sit down with data providers, technology leaders and Cytora partners to uncover actionable knowledge on how insurance can achieve frictionless risk flows. Today we're joined by Jacob Grub and Isaac Lavi of geox. Hi guys, thanks for joining us. How are you doing?

Speaker A: Hey Jake, pleasure to be here.

Speaker B: Hey, doing well.

Speaker C: Good to hear. So I guess to get started, would you guys mind giving me a quick introduction into yourselves, your backgrounds and then a bit of an overview of gox's mission and um, where it sits within the insurance data landscape?

Speaker A: Yeah, for sure. So my name is Isaac, I'm CEO of gox.

Speaker B: I'm Jacob Grob. I've been in the insurance space selling data to insurance companies for 20 years. Twelve of that was at CoreLogic. A lot of building characteristics there. When I joined CoreLogic, they were part of First American. They spun off and started just buying all sorts of data companies. So you look at Risk Meter, which is a lot of deterministic risk scores, Equicat, which is Cat Modeling, MarshallSwift, Beck, which was uh, replacement cost. And so really exposure to kind of all of these different pieces and parts of the insurance underwriting ecosystem. And as time went on there's a certain set of data attributes that there's just a blind spot in the market for. And that really is what led me to looking at computer vision and data extraction through that. And GEOX is just such a natural fit. They have been doing this since 2008. Are uh, absolute leaders in Australia where they built and scaled up in Japan where precision is absolutely key. And they're coming into the US now with a very mature product and one that can deal with problems at scale. When I look at us compared to kind of the other players out there, the other players are much more focused on high touch underwriting if you will, where you want to sit down and dissect a uh, high res image where you have somebody touching and feeling every single one of those policies. GEOX is taking a Different approach and is really focused on solving problems at scale. So that's high volumes, high, straight through processing, getting a submission to a point where the underwriter has all of the relevant detail at their fingertips and can bring that decisioning down from two hours to a matter of minutes. So that's really what we're focused on, is that high scale, high volume analysis versus the other. And really like our mission is to be that base layer, that base information that's being extracted from imagery and plugged into all of the AI underwriting tools that are coming to market right now. All of the underwriting workbenches that are out there, whether that's an underwriting workbench that is provided by a third party vendor or it's one that a uh, carrier is building on their own, at the end of the day they need to understand what's happening at that property, what's on that property. And we've built a database that includes all of that information. And so we want to be that first choice for that first piece of information.

Speaker C: Yeah, I mean a lot of that really resonates. A few things you mentioned there. I'm really excited to kind of COVID in the rest of our conversation in that how do you make it work at scale? What learnings can we take from other M markets as well? Isaac, I'd love to hear from you a little bit about I guess how you guys differ from competition and other providers within the market but within your kind of specific area of intelligence as well if you could.

Speaker A: Yeah, for sure. So I uh, will continue from Jacob point. So I think that when we start. So actually it was very interesting because we start in different markets. We start from Japanese market and Australian market as Jacob mentioned. And I think that our approach was a little bit different from what is existing today on the US market is create the full databases, all the properties across the countries. So it's mean that we cover all the properties that's existing in Japan, all the properties that existing in Australia and what make it very powerful that the, the instance response from our API to the workflow of the carrier, it's very important. Second one is a coverage like when eventually when you make insurance, you need to provide eventually the product to everyone. So to make one standout across the entire country this is very important as well. So we can support you. We are the only one the market can support carrier in United States to provide like 100% of the properties in the United States and also the accuracy and the accuracy statements that we can be behind that because we create very stable and very standalone product. You can measure that, you can provide accuracy statement across the state and it's very stable in terms of prediction of what is the results will be and where the mistakes can be. So this is make a wonderful trust point for the carrier and I think this is really what make us unique just to understand the scale. So in the last quarter uh, we ran update on 130 million building footprints to compare to what is right now in the market. Our uh, competitors can update around 90 million building footprint nationwide for the entire year once and we in one quarter running 130 million properties in average. So this gives you the scalability of what we are doing and how we can support the insurance and the massive of data that we're running every quarter. It's nothing comparable from what is existing today in the market. We saw the advantage in other markets. So Today we support 100% of Japanese market, we support almost 100% of Australian market. Also three out of the four biggest banks, for example infrastructure companies, governments and et cetera. And eventually it's only about the time when we'll be 100% of US market. So this is our mission to be the standard geospatial insights for every carrier to make the right decisions and with uh, accurate and trusted insights for their needs. So it can go from building data, roof data, it can go to the defensible space, environmental informations and more and more. So this is our specialty and this is eventually what we want to provide one API key or platform that can serve all the appetite of the insurance carrier from the geospatial area.

Speaker C: Brilliant, thank you. And I think there's a kind of key theme that's touched on there in that you need all the context, you need all the data and all the coverage, but you also need to balance that with I guess human capacity. Like how do you make it usable as you mentioned Jacob, to avoid that super high touch underwriting. And I guess building off a couple of the points in both of your responses there. So some markets uh, they're forced to live with natural catastrophe far more often and more severely than others. And you touched on Japan and Australia. So do markets under the most pressure end uh, up further ahead in how they understand and use property data? And is there anything that they've worked out that the rest of the industry perhaps hasn't had to confront yet?

Speaker B: Absolutely, yeah, there's no question about that. You know, and maybe this comes from a little bit of bias about myself because I've worked in the Florida Market for most of my career, I'm located in Tampa, Florida. And so to say that I'm biased because I'm in Florida. I agree that being in a high catastrophe state or with a lot of exposure, you just end up having to be better. You look at what's happened here in Florida, all of the big guys pulled out and all of the smaller players started cropping up. And these smaller players, in order to be competitive with each other, had to do things faster, better, more accurately. The competition is just so stiff here and they're exposed to greater hazards than anyone else in the country. We have wildfires in Florida. You don't hear about them a lot because they're very rural. But you have hail as well. You have hurricanes and hurricanes, of course, the big one in tornadoes as well, you're dealing with all of those. And the coastal storm surge occurs as well. So you have to be better and you have to move faster. So those two things together are what create a situation where you see higher and faster adoption of roof condition, scores of roof age products of defensible tree overhang. And a lot of the middle of the country, a lot of the players that don't have that same exposure are still trying to figure out how to implement these solutions within their workflows. So, I mean, short answer, yes, you've got to be better, you've got to be faster. Just having that exposure will always create more pressure to improve.

Speaker C: Yeah. And immediately you can kind of see the benefits there. Some of the things that you covered, I mean, turnaround times is one offering a better broker service, responding quotes, returning quotes more quickly. But I mean, from your description, taking Florida of an example, there's money left on the table for the tier 1s and tier 2s. Right. Obviously there's more risk there, but with better data, they can price more appropriately to cover their potential losses and access those kind of added revenue streams that perhaps they've discounted.

Speaker B: Oh yeah, without a doubt. I mean, you look at the roof in Florida, it is the number one greatest factor in losses in a hurricane, and that is the condition. And we know that through our work with fema, where we created and really understood the post event landscape, our roof condition is based off of analysis done there. And so understanding the quality of that roof and then the shape as well, you can absolutely underwrite very profitably in the state with those data elements. And so there's certainly plenty of business being left on the table by the big guys not coming in and not adopting those solutions.

Speaker A: Yeah. And what was Also very interesting about that. So many carrier requesting like what is the roof shape that you have? Like if you have gable roof, if you have like heap roof and et cetera. But according to our study like much more important not to know like if it's gable or heap and flat roof, the if there any GABA wall and what is the direction of the gable wall if it's directed to their coastline. So your exposure is much more higher even if you have like gabber roof and the hip side is directed to the coastline. Uh, so still the exposure will be much less if you look on where the roof position. So when you're looking on this is just one example but we can bring like so many examples like that. So when you deep dive inside of every key insights, very small insights can make tip of the ice of the differences from what you have on the book. And everyone can bring you like advantage. One of them can bring you 5% more like 5 or 2 or 1% better protection and better uh, knowledge. But eventually when you have many of them, we provide around 70 attributes and every quarter it's increasing. So you definitely can benefit from each attributes in parallel.

Speaker C: You've illustrated it really well with some of the examples you gave. There's so many variables and attributes that can inform a risk and an underwriting decision. And you need to strike a balance between gathering all of those attributes for context but also making that scalable. So I wanted to ask you about the scalable question. What's the real difference between proving a concept and running it dependably at a national scale? And um, perhaps where does the proof of concept start to fall apart in comparison to something that works well at scale?

Speaker B: I love this topic, right? I love this topic because LLMs are just everywhere. Everything is AI. LLMs do a great job. If you drop a picture in doing some analysis, that looks good, looks decent, but it is still a probabilistic answer. And when you look at those probabilistic ANSWERS, there is 20% of the time or more that it is a hallucination and the answer looks great and people want to trust it because it sounds good, it's convincing, but it could be total nonsense. And so understanding the difference between that small scale POC LLM um answer what is right and what is wrong is almost impossible unless you are doing high touch and looking at the images and making those assessments yourself. And if you're doing that, if you have to look at the image, you've just completely wasted all of your time putting it into an LLM like You're not getting any additional information. You're already doing the work. When you talk about, like what we're doing at scale, it comes with a confidence score. We are turning that probabilistic answer into a deterministic answer. We are providing confidence. Hey, you know what, the imagery that we have for this location isn't that great. The statistics that we see for the answer we're giving here aren't to our standards. And so we're going to say, hey, this probably isn't something that you should rely on or it's something that you should do a little bit of extra research and being able to provide that at scale and know that those high confidence answers, you can just slide those right through. And those low confidence ones, those are the ones that are flagged. And we tell you that is a fundamental difference between a probabilistic and deterministic model. But beyond that, when you are looking at things at scale at nationwide levels across the board, you get to some of those insights and those details that Isaac was talking about. So like he said, the aspect of the distance or, uh, the direction that the roof is facing makes a huge difference in an event. Well, you have to have that data first across an entire affected area before you can do the analysis to find out that that actually is an impact. And so there's so many advantages to building, you know, one, at uh, scale as far as a nationwide level goes, because you can find those additional details and insights that you wouldn't be able to do if you didn't have all the data. But then two, because we're using a deterministic model or because we're able to validate at that level, we're giving better answers back that are actionable, not look good.

Speaker C: Totally. And I guess to dig a little more into that then, so what changes for an underwriter themselves when they can see not just the answer, uh, but the confidence score and how sure they are of the accuracy of that answer?

Speaker B: Well, I mean, just life gets so much easier for them. They can trust that the answer that's coming out is one that they can move forward with and that they won't have issues down the line with their carriers or with their partners, with their brokers where, you know, they've given, you know, turned down coverage because an LLM, um, hallucinated or wrote that policy that they're going to have a massive loss on, like that uncertainty, not knowing for sure that the answer that's coming out is, you know, the level of stress that you have when you're writing a policy that's worth millions of dollars or writing hundreds of thousands of policies that add up to that, knowing that the data is correct, that you can trust the answers that you are using and being flagged for those you aren't are accurate.

Speaker A: Yeah. I also would like to add that eventually our uh, main goal is to provide a superpower for the underwriters. So this is our main goal as the company. So we give them the two different abilities. First one we give to the insurance carry ability when they don't have enough time to even put the handwriter on the policy. And this is happening very often, especially on the PI side like personal line and also on the SME side like small medium commercial enterprises. So what we see that the handwriters don't have even touch on these policies at all. So the carrier need to figure out like how I make it automated, how I can trust all the data. And this is exactly when we are popping up and the carrier say okay, now I have jerks jax data. So I can trust on that and I can like automated policies. And not only that, it's also connected to the last question, like how the carrier can enjoy it from the scalability that we have. And we see many carriers right now shifting to pre process and pre pricing every property in the United States. And this is give them like huge ability. Because when you're looking just on the policy coming through the quote system that you had and you just see the property that coming to you, it's different perspective. When you look on the entire market and say now I know like every property what is the price that I want. It's helping you also to manage the marketing side, it's helping you to manage the sales side. It's helping you to make better decisions and how you spread your exposure, how you manage your book and everything make much more sense from your side and you're not surprised at all. So this is quite incredible. And eventually give the superpower also for the carrier and for the underwriters, this uh, is my second point. The tradition is the underwriters have like 2 hours to underwrite the policy, especially on the commercial side, two hours to put a price tag on properties that worth between like $1 million to $100 million. It's almost impossible. So eventually you have to understand where immediately when you have the weaknesses of this property, when you have the strongest point of this property, you have to everything immediately so you can actually invest these two hours to make the right decision and to select much better policies that you want and don't want. And what is the right pricing? So we, our goal and our mission is to give the underwriter, like the superpower, to make right decision and also to invest these two hours to purchase our two hours that they have on the right decision and right move and where to investigate this policy.

Speaker C: Yeah, and a number of things you said there really kind of illustrate the cumulative benefit of accurate data that is, I guess, more expansive in the attributes it covers. Like we've discussed. If you can understand the area or the building's attributes to a greater degree with accuracy, you can quote more appropriately. You can quote more in less time as well. And you can also stop missing things that are perhaps worth quoting. That based on really generic data sets might otherwise look unwritable. And I want to dig a little bit into that as well, because that's something that's probably less discussed than the other points we've covered. There are whole areas and perils that carriers treat as effectively unwritable and they blanket them out of appetite. What are the specific factors that would tell you that a location inside one of those zones has perhaps move to being safer and in appetite? What are the things that genuinely separate exposed property from one that just looks exposed because of where it is?

Speaker B: I think the most clear example of this is in flood insurance and flood zones specifically. When you look at flood zones, they're really raiding territories more than risk zones. They're very political and their responsible party to create those flood zones is the county. They have to do, uh, surveys to bring that data up to speed or up to date. And a lot of times they'll delay these significantly. And so you end up with a lot of areas that are exposed that are outside of flood zones. And there's a lot of areas inside of the flood zone that actually have never flooded at all. And those locations inside of the flood zone that have never flooded, their first floor elevation is greater than the surrounding areas. And one of the things that we do is we provide first floor elevation and that gives combining our data with other data sources out there, it kind of sharpens that underwriter's pencil. It exposes and opens up new markets or uh, new properties, new policies that are going to be incredibly profitable but look terrible to everybody else on the market that don't have that insight. So specifically in a flood zone, if you're inside of a flood zone but never flood, you're paying really high premiums for that location, but you're not having any claims that detail that first floor elevation. Is what separates you from the others and what gives our customers the ability to cherry pick policies inside of areas that normally you wouldn't touch. Very similar case is wildfire. You can be in a very heavily wildfire prone area, but if you have really good defensible space, the likelihood of you burning is very low. So again it is having the right base map to start. If you have the good base map and understanding and all of those attributes at your fingertips, then all of the other underwriting that comes out of that or grows out of that is going to be more profitable, more predictable than anything else. And so you got to look at it one of two ways. One is it do you have the right base map to accurately underwrite or are you bringing in that last detail to put you over the top? And I would say you should start with the right base map. So that's kind of my take on the place position.

Speaker A: Yeah, I would like to add on that that for example what we saw that let's just look on the roof condition on uh, storm and hell position. So we can determine like all the roof condition evidence and what is the damages and what is the roof condition score that we provide. So for example, if we say like this roof is damaged so that for example we see top we see like enormous rust or we see like all the roof is with discoloration or something like that. So we give this insight for carriers and what we saw that the frequency of property that claimed the insurance carrier that Gox says uh, this property had damaged roof condition is three times or higher. The severity is around 1.8 to compare to the average policy. So now when the carrier going to insure this property and they know that joex actually notified him like look, this is damaged roof. So in total the risk the carrier take on this policy will be much more higher. And we can looking also on the other position like right now we are on growing market. So we see that the carrier going to compete with each other very hard on the next year. So we can provide the insights of excellent roof and excellent roof it means that the frequency is going down dramatically and severity going down dramatically. So the carrier can give actually a discount and we see some carrier that give discount around 20% and 40%. It depends on the market and also on the area. But we see that the carrier eventually with our insights ableing to provide discounts and compete on the market. Especially when we go to growing market and you need also to be more competitive on the environment and the infrastructure. So it means that you have to provide fast quote because there are many right now MGA's that provide automatic quotes and et cetera, so people don't have a patient anymore. So you need to be competitive, you need to be inside of the market and you need to grow a business. So we can support you on that as well.

Speaker C: I think that's a less touched on benefit of the kind of data we're talking about as well. Right. You can highlight the times when you're able to compete from a price perspective. If you know something a competitor doesn't as a carrier, you can understand the amount of play you have on, um, price and where you can provide a discount, as you say, that's more appropriately priced for the level of risk for each property there. And that's not even to mention, as we've previously discussed, things like time to quote, even whether a competing carrier covers that area at all. So I think it really highlights the opportunity there is to differentiate and kind of access more revenue and grow their business. A couple of other things that stood out to me. So one is, I think it really just highlighted that this level of data to gather and interpret manually is just not something carriers have the capacity to do from a headcount perspective. And so as you say, being able to integrate with these kind of headless systems that sit in the middle and can use LLMs to interpret that data is essential, but that requires accurate data and it requires real time data. But I think the last area that I guess really stood out, that some of the anecdotes you've given me have really illustrated is, and um, particularly for growth in the US for you guys, right, You've got significant experience in Japan, significant experience in Australia and they encountered different perils, different types of perils. So perhaps one is more prone to wildfire, one is more prone to floods. Same for earthquakes, for really intense winds. The US gets a little bit of all of that. So that really highlights how you guys can slot into that market, I think, and offer that insight to the US

Speaker B: I love you brought up the difference in hazards. They are different in, they have different names, but a lot of them are the same. There's still flooding there, there's still wildfire. And actually the flooding piece I wanted to talk about for Australia because I think it's a great story, a great anecdote for our data. So we started providing first floor elevation in Australia due to flooding and we gave the data back to our customers and they came back and said, this is wrong. These properties are completely different than the first floor Elevation you say they are? Well, the reason they're wrong is because in Australia, after a flooding event, they will actually lift up the building. They will take that claim and mitigate the flood risk by lifting the entire building and adding to that first floor elevation. And so we provided data from before the flood occurred. And so a lot of the houses were at slab, uh, on grade and after the flood they were lifted, which is just a wildly different way to address flooding and mitigate that you just don't see in the United States yet. But it was a really interesting kind of way to find out that what this data is not wrong. Like we have the images right here, you know, and then to come back and find out that they lifted the building, the whole building moved. So that was pretty fun.

Speaker C: It's a novel and unexpected way to mitigate a risk, I guess. But I mean, if policyholders are doing that, then as we say, it changes the profile of those areas and it puts something on the table for carriers that previously wasn't there. And so again, uh, it's a great anecdote, I think, to demonstrate how real time data is important, data that's up to date and data that isn't overly generalized when it comes to geolocation. One other area I want to get into with you guys, and that's the buy versus build conversation. So a lot of what makes what we've discussed today possible is infrastructure that many carriers maybe will consider building themselves, but we've probably seen equally many, if not more carriers try to build it themselves and perhaps achieve a level of success, but generally it will fall apart at scaling. So when you look at the work that goes into producing property truth across an entire market, what would you say is genuinely out of reach for a carrier to do in house? And what does that tell us about the build versus buy line and where it really sits today?

Speaker B: I think that one of the first places that I would look is just doing the compute. Processing the entire country for this data, or an entire geography for this data takes a lot of compute. And what we've done to mitigate that is gone out and bought our own compute. So we have the latest generation Nvidia Blackwell 300 AI servers that we're running our analysis on. That is really what opens up the gates for us to be able to pre process and reprocess the entire United States on a quarterly basis. And does a carrier want to get into the compute game and the server game and all of that? That's just the physical hardware to be able to Keep a data set like this up to date. If you're doing it in the cloud, your cloud costs are going to be astronomical. And if you're using kind of the standard LLMs that are out there, if you're using anthropic to try and process this level of data, your costs, you might as well hire an army of underwriters to do the work if you're going to be spending that type of money. So there's definitely a hurdle there for that scale. And then of course just the expertise. We have a team of PhD level computer vision and mathematics professors that are dedicating day in and day out to this problem, to applying computer vision. And I don't want to go and sell insurance. I want to do computer vision really well. And I don't think that an insurance company wants to learn how to implement computer vision. Their job is to go sell insurance. And if they're focused on building, building technology, they're going to take their eye off the ball of what they actually need to do. And that is apply the data and write the insurance. And so it's twofold, the infrastructure to do it and then the focus to do it.

Speaker A: You're absolutely right. And uh, I think the funny part of that that actually the carrier that try to develop these models, especially the roof condition models, they are the most appreciated one. So eventually I'll best clients that give us always feedback, appreciate our work and et cetera. This is the same carrier that tried to develop these models by themselves and eventually switch to us. So this is very funny part and I think that eventually when I tried to investigate it, like what was the reason behind it, like you're selling insurance, why you need to jump into the computer vision and to build the entire infrastructure of AI and LLA models and etc. So what you find out that many carrier afraiding that our models is not fit their appetite. Like this is the reason why I think, like when I investigate and I ask many carriers why they develop by themselves, this is the most common one. And this is exactly when we decide that we have to provide the evidence behind our score. We have to provide the material data. Like if you say this roof is very bad, so we need to give the reason. So until today no one is giving the reason to carriers, like okay, this damaged roof, this bad, et cetera. So they used to just receive this call from the provider and no one talk with them what is the reasoning behind it. So with our data we enable to them not only to see the reason but also to Customize that. So for example one carrier say if it's concrete roof, I don't care what is the damage that happened, just if it's for example top on the roof maybe or for example if there are any huge rafts on the roof and et cetera or if it's concrete, so huge ponding. But when you can see the damage on the roof and carrier can actually play with that and play with the skull and decide if they want to customize that. Because eventually every carrier know what is a right to him to himself and how to customize the roof condition, how to customize the attributes, how to digest the data, how to to the right fit. And we enable to carrier actually do that. And this is why our best client actually is a uh, carrier that try to develop that invest huge amount of years and money and eventually just consume from us today the data. And they appreciate that more than anyone else.

Speaker C: That makes total sense. The carrier that has tried to do it themselves and realized how difficult it is, but also realize the value of it. Kind of the most appreciative of what you guys provide and the cost of compute versus the expertise. Again, uh, when you put it that way, it's so intuitive, isn't it? There's no economies of scale for a carrier in producing that kind of thing. They're going to create it and they're going to use it for their own needs. You guys being in the position you're in as a vendor and a provider, you can create it with all carriers in mind, which means that it instantly kind of has more potential roi. And I think another historical maybe concern about using vendors that has come from carriers is what you mentioned, Isaac, in that it has to match their appetite and their view of risk. And perhaps that was a kind of understandable concern in the past, but in the today, vendors really are understanding that and their products are built to be customizable and to be able to be directed at exactly how carriers need to ingest data and how they want to run their businesses. So super intuitive there. I think I've got one more question for you both that we tend to end this podcast on. It's a little bit of fun, but if you could each change one thing about the insurance industry, what would that thing be and why?

Speaker B: Coming from a person who's been working and selling data to insurance companies for 20 years, I would say the speed of adoption, the speed of change. I can come to a carrier with the perfect solution, solve their pain, solve their problem, but it tends to take another year. And a half to get to the finish line. And everybody on the insurance side is just as anxious to make the thing happen as on the vendor side. But it's just like the industry as a whole just moves very, very cautiously. And I think that it's kind of time to start taking more risk and pushing things faster. I think that that's going to be the one thing that I think I'd like to see happen and I think it's for the better of the industry as a whole. Like move fast and break things. Like, let's go and do this, let's solve the problem. That's my take. But that's also coming from a sales guy. So liquor, uh, source, you know.

Speaker C: Yeah, I mean I think some of that's culturally embedded and some of it's due to inertia, but now is the time to change that because competition is, it's just exploding in terms of where it comes from and how fast the gap grows between competitors. So yeah, I think you're right. That's a huge priority. How about you Isaac?

Speaker A: I think it's maybe a big ask but today the carrier react like they are payers and not as um, predicted and prevented. So I think that we see more shift right now that some carrier, especially the big ones, try to be more predictive and prevent the damages and think it will be eventually end. Clients will appreciate that first like when they see the carrier actually make action to protect them. And I think the second one is eventually they will see that on the loss ratio that will reduce dramatically. So in some geographical areas we see more carrier like try to be like informative, actionable, try to protect and prevent the next damage from there, uh, policyholder. And I hope to see that more and more in United States and to all carrier take action and protect the uh, policyholder. But I think it's, it would take years to see that eventually. But this is where I think that the market need to go.

Speaker B: Seriously, like we are in a time of climate change. Like we are having more and more impacts in the world and there needs to be more hardening of our communities. There needs to be more resilience built into what we're doing. And I really do believe that Isaac is right, like insurance can lead the charge in protecting this country and the world civilization as a whole from the coming changes that we're experiencing. And that partnership is absolutely huge. And I love that answer.

Speaker C: It's a great answer. It's a win, win, win. If we can be more proactive around mitigating loss in that way, both from a financial perspective, but also from an, uh, impact in the lives of others and from an economical scale perspective. From my experience, I think the line of business that's doing that best at the minute is cyber. They've got the biggest opportunity to do it. It's so fast moving. There's so many eyes on in terms of where vulnerabilities come from and then implementing fixes to that. But with the technology we've been talking about today, there's no reason that that can't be applied to climate peril and risks in the same way. As you say, the climate's changing quickly, but so is our technology. So our, uh, understanding of it and ability to predict is growing equally rapidly. So a great answer to finish on and it's a positive answer as well. So a nice end to the conversation. But I just want to thank you both for joining me. It's been really interesting to learn about both of your own experiences and also the mission of gox in the insurance market.

Speaker A: Thank you very much, Jake.

Speaker C: Absolutely.

Speaker B: Uh, thank you. Making Risk Flow is brought to you by Cytora. If you enjoy this podcast, consider subscribing to Making Risk Flow in Apple Podcast, Spotify or wherever you get your podcast so you never miss an episode. To find out more about cytora, ah, visit cytora.com thanks for joining me. See you next time.