Welcome to Scouting for Growth. 93 percent. That is the proportion of UK commercial properties insured for the wrong amount. Across the Atlantic, 90% of U.S. commercial buildings carry inadequate coverage, and underwriters rate their access to risk intelligence at just 3 to 5 out of 10 at the moment of decision. We are not talking about a data shortage. Data exists across the ecosystem. The problem is that it doesn't reach underwriters in a verified, structured form when they actually need it. It is an architecture problem, an integration problem, and fundamentally a trust problem. This is the risk intelligence gap, and it is costing the industry billions. Today, I'm joined by someone who is building the infrastructure to close that gap. My guest is Anthony Peake, CEO of Intelligent AI. Anthony leads a property risk intelligence company that delivers verified API-first property data directly into underwriting workflows, covering over one Android structured data points, pair building from construction attributes and fire protection to natural peril exposure and forward-looking climate protections. Developed with institutional credibility through Lloyd's lab, intelligent AI serves insurers, reinsurers, brokers, and MGAs across the UK and US markets. Why does this chat matter right now? The commercial property market is navigating a period of profound structural vulnerability. In 2024, global insured catastrophic losses reached $145 billion. The U.S. PNC industry posted consecutive net underwriting losses exceeding $20 billion in both 2022 and 2023. And the industry has invested billions in sophisticated AI decision engines and catastrophe models, only to fuel them with unreliable, unverified and often still input data. Meanwhile, underwriters spend 50 to 55 percent of their working day chasing data rather than making decisions. This is better engines running on worse fuel. Together, Antony and I have co-authored a new research paper, The Risk Intelligence Gap, drawing on 30 plus insurance annual reports, 20 executive interviews and multiple secondary research streams. And the findings are striking. In this episode, we will explore what the research reveals about the true scale of exposure data deficiency across the UK and the US markets. Why better technology alone won't solve the problem and what the real bottleneck is. How API-first risk intelligence can transform underwriting from reactive data chasing to decision-grade insight. And what the path from here to predictive cognitive underwriting infrastructure actually looks like. This isn't a conversation about incremental improvement. Remember that. It is about whether the industry can close a structural data gap before the next extreme weather season exposes it further. I'm Sabine van der Linden and this is Scouting for Growth. Let's step into the frontier and explore what happens when property underwriting finally gets the data it deserves. Anthony, thank you so much for joining me. Thank you, Sabine. I know you've spent over 30 years working for enterprises like Apple, GE, BT, Oracle, and you have delivered risk management system for six of the top 10 UK insurers. Why did you choose to focus on property data for commercial insurers? I've always been involved in data. So everything from the early days of the Macintosh and Apple and Oracle and GE, etc., has always been around data. And one of the markets that have the biggest problem in data and therefore the biggest opportunity has been insurance. We have been working on a piece of research together. for the property market and looking at data and risk for that specific target market. And we found something striking into the data. Carriers are investing millions in AI pricing engines and catastrophic models, actually. But the data feeding this model is rated 3 to 5 out of 10. I mean, don't you find this peculiar? I find it troubling. I find it sad. Having spent the last 12 years working with insurers, it's not unusual. It's fairly common. But at the same time, to me, it's a huge opportunity to work with them to improve the way they do underwriting. We actually call this automation paradox. And what do you think that actually looks like on the ground when you actually start interacting with those organizations? Interestingly, the first problem I find is that insurers have very poor addressing in the portfolios they underwrite. Less than 50% of the properties they're underwriting, they have a usable address. So postcodes can be wrong, street names can be wrong, company names can be wrong. And whilst an insurer can get away with saying, I'm insuring all the buildings at that site until there's a claim and then they need more detail. To help them do it digitally, we need better addresses. So the first thing we do is we have an AI engine that fixes the addresses. What we then find is that there is a model in property underwriting called COPE, C-O-P-E, Construction, Occupancy, Protection and Environment. There's about 100 pieces of data or so that each of the insurers need for their model. Paul. Invariably, the data that the client gives to the broker is poor, and the data that the broker then hands on to the insurer is poor. And there must have been a time 10 years ago when people looked at this and tried to fix it and said it was too hard, and so they've learned to deal with poor data. But they're not making profit in property underwriting. They're invariably losing money on property underwriting. And so now they see an urgency to improve that. So the underwriter and the risk engineering teams have to then manually look for this data. And I've seen people manually clicking around the footprint of a building. I've seen them literally on Google Earth measuring up the side of the screen with a ruler to try and work out the size of the building. And that's not fit for purpose today. So they're trying to underwrite tens of thousands of properties. they don't have time to be doing all the admin to pre-fill the data. They're sending risk engineers out then more to sites, and that's very expensive. And those risk engineers should be spending their time advising clients on risk, not just gathering bits of administrative data. So ultimately what they want is to have the right level of COBE information across every property in the portfolio in order to have that 360-degree view of risk. So conducting interviews with chief underwriting officers over the past three months, what we've realized is that some chief underwriting officers do understand what COPE means, but some do not. So can you explain what COPE stands for in a little bit more detail, please? Yeah, I mean, I started talking to... So having built systems for AXA QBE, RSAMs, Amelina Viva, et cetera, I started talking to them about COPE probably about seven years ago. And all of them saw it as the holy grail. If we have a sort of knowledge graph, if we have a proper schema across the industry that all of this data aligns with, then we can compare data more easily. So it's almost like in automotive, the insurers know whether somebody's had a crash or whatever, and they share that data. But in a lot of underwriting, particularly property underwriting, people use different schemas. So COPE is construction, occupancy, protection, environment. So from a construction point of view, what's the footprint? What's the height? How many store is in this building? What's the wall made of? What's the roof made of? What's the frame made of? Yeah, almost like the wolf blowing on a house. You know, if it's made of brick and stone, it's not going to fall down and it's probably not going to burn. if it was made of straw or wood or some other material it's likely to burn a lot so it helps for your combustibility to actually understand what it's made of equally if you if you're talking about is it made of brick or stone a brick terraced house in Leeds would probably be about 200,000 to buy for about 300,000 to rebuild however if it was made of stone it'd be 600,000 So knowing the difference between whether it's stone or brick. So on the construction, it's all of those variables. On the occupancy, occupancy depicts the types of risk. So if I am a perfectly safe office building rather than a petrochemical plant rather than maybe a synagogue for geopolitical risk understanding the occupancy and many buildings you know i i work in a in a property in extra science park sometimes um it invariably 40 companies of you know doing technology with officers but there are four wet labs doing chemistry and there's a company right next to reception who are doing research into fireworks actually treating everybody as a as a sort of inert office building and not knowing about someone doing research into fireworks makes a huge difference and then from the protections point of view do they have sprinklers do they have the right sort of sprinklers do they have alarm systems you know how far from a fire station and then ultimately the is usually environment or exposure, but the sort of earth, wind and fire, NatCat, you know, natural catastrophe things, flood, hail, etc. But also exposure from a, I'm a perfectly safe office building, but I'm next to a petrochemical plant, I'm next to a petrol station with an underground tank, and, you know, I'm next to a synagogue. So actually, that COPE model clearly defines, and as we move into very much into a computer age and we move into an agentic age, you need to have this schema. Yeah. Well, you already mentioned this, right? We are seeing an hidden data tax. We have documented, for example, that 50 to 55% of an underwriter's time goes to chasing, validating, recceing data, right? That's not all. Right. Instead of focusing on risk selection, right? We know that. So the industry estimates that there is over 150 billion, I mean, I think it's 160, 170 billion efficiency loss over five years. Where is all that time actually going, Anthony? So I've worked with some insurers recently who, the underwriting team, need them to go and visit 10,000 properties. And again, if you've got your risk engineers visiting 10,000 properties and all they're doing is gathering data, they're not actually helping the client reduce the risk. They're not actually helping the organization. And they're doing admin. We've actually, in that particular case, we managed to get them down to actually only having to do 3,000 visits. The other seven, 7,000, are actually done digitally now. That's freed the team up to spend twice as long with each client to actually educate them on risk. It's also freed the team up to do more capacity. So it's little things where we all feel the computer should be doing it. I say 80% of what humans do is admin, and less than 20% of what we do is adding value to our customers and our company. If I can half, even half the amount of admin and I can double the amount of value I'm delivering to the business, a business is hugely profitable. It's interesting because I'm having at the moment a lot of conversation around frontier transformation. And you are highlighting here part of the foundation to build a frontier insurer, right? The data foundation and making sure we have the data at the right place to actually take the right decision. So in the UK alone, according to some of our findings, 70% of commercial properties are martyly underinsured, right? A cumulative exposure gap of over £350 or £68 billion in the UK alone. In the United States, 90% of appraisal of buildings are underinsured as well. and after the LA wildfire destroyed 17,000 structures in January 2025. Many of those were carrying stale baselines, and it would be great for you to explain what that is. I know who is paying for it. You know, how do we deal with that? So to date, it is the end client who buys the insurance, who has been paying for it. So where does the problem come from? So, invariably, someone with a portfolio of properties, a corporate with a thousand properties, for instance, they would get a quarter of that portfolio valued each year by a professional chartered surveyor, valuer, et cetera. And that sounds good. But the other 75% of the portfolio, they'll add one or two or 3% to the increase to the insurers to think they're covered. Last year, construction, plant, machinery, labor went up 20%. So if you're only adding 1% or 2% or 3% a year and the market costs are going up 20%, you can see where you're falling to this sort of average globally of 80% of properties are underinsured by at least 50%. And it's getting worse. In the UK, as a good example, up until 2024, most of the insurers apply an averaging clause, which means that if I tell my insurer I want my property insured for half a million, but it burns down and I need a million, the insurer will go, well, you've only given me the premium for half a million. So here's your half a million and off you go, enjoy or not the rest of your life. And invariably people in that case will complain to the FCA and they've seen a 400%, the Financial Conduct Authority have seen a 400% increase in complaints. And so that's the way the market has been. In 2023, 2024, the FCA came out with the Consumer Duty Act. This has changed the landscape, actually for the benefit of both, but a little bit of pain, where it basically says, yes, Mr. Insurer, you can hide behind the averaging clause, but you need to demonstrate you've got a system internally that could calculate what the amount should be for and that you've advised all your clients. And none of the insurers have the capability. And so what the FCA is doing now is turning around to the insurer saying, you didn't advise the client correctly, you need to pay them a million. It's making a huge change in the market, but under insurance, good example in in commercial so i did a portfolio recently at lloyd's with with a big insurer where we looked at 355 commercial properties of between 1 and 20 million average 14 million thereby it was currently insured as a portfolio for 5 billion we put it through our platform and and it takes seconds to recalculate with all the right data and we were able to demonstrate it should have been insured for 6.17 billion. So there's only 1.17 billion underinsured and a relatively good portfolio for the insurer. More importantly, the insurer had only been getting 7.5 million of premiums when they should have been getting 10. Okay. And so it shows just on 355 properties when many of these insurers have 355,000 properties, you can easily find 2.5 million of additional premiums. It's interesting because it feels like, you know, property risk data are very important. Rebuild data as well are very critical. Sparning when we look at the inflation, social inflation we're experiencing today. So why do you think the market is not paying enough attention to that right now? So I think it's a sort of inflection point. The data hasn't necessarily been there with the models. The market has got used to under-insurers but hasn't had systems to be able to resolve it. And the legislation hasn't been there to drive people to do better. We've actually reached the point now where, as I say, with our platform, within 10 seconds, I can give you a more accurate valuation than you previously could. And today, insurers are spending 90 minutes manually doing something that the computer can do for 10 seconds. So we're not saying the computer will replace valuable risk engineers and underwrites, etc. We look at it very much as augmenting their role. But if we can take all that admin off them and free them up to actually save lives and do risk mitigation and better value properties for customers, then everyone wins. So let's go back to the definition of frontier transformation. intelligence on top, human and Asian collaborating with one another, and every human becoming an Asian manager or an Asian boss. Absolutely. So that's where we start freeing our times to take better decisions, to actually be more creative, and actually focus on system thinking and solving probably more complex problems. So you are building what you call digital twins of risk. Yes. Virtual replica of physical properties enriched with, I think, 300 plus data points. And it would be great for us to go into that from satellite imagery to real-time float scoring. For the people who are going to listen and view this interview with you, Anthony, what does a digital twin actually do that a traditional property record cannot do right now? So a digital twin, the term digital twin came out of aerospace and automotive, etc. If I wanted to pay to have a new car built and I wanted to test it in a wind tunnel, that would be very expensive. But I can build a computer model of what that car looks like and I can build a computer model of the wind. We've coined that term to be able to say, if I can create anywhere between 100 and say 300 pieces of data, depending on which insurers model, and I can simulate all of the data for the construction, for the occupancy, for the protection, and for the environment, then we have a virtual digital twin of all of that data. And then I can understand, you know, from the wall materials, the roof materials, and the frame, I can understand the combustibility of that building. If I can understand the occupancy, if I can understand the types of trades, I can understand the types of slips and trips that might go on in that building, or the amount of crime that might go in that building, etc. And it's really just a term we've phrased, you know, we've coined it, it shouldn't be something that people feel fearful for or big. It's just simply, we can create a model of all of the data that a risk engineer might manually go on site and collect. But we can do it very accurately, we can do it very quickly. And more importantly insurers today deal with data Risk engineers for instance right And a lot of people have this misconception that risk engineers go to 100 of properties The average across the industry is 3 to 5 There one very very technical and very prestige insurer who deals with a lot of technical and fire who at maximum do 20% visits. So over 80%, and in most cases over 90% of properties, they're underwritten or underwritten without much data at all. And on that side, the insurer doesn't know the level of risk they're taking on. And when they don't know the level of risk they're taking on, they charge more and they charge everyone more. Obviously, the more certainty we can give the insurers, the more competitive they can be. And that's my hope is that all of our insurance at least stabilizes if not comes down. But actually, we're using the data then to give the insurers the certainty. The other thing is most brokers will send a portfolio to multiple insurers. And some insurers I've seen can take two to three weeks to manually analyze and reading PDF brochures on Google and all this kind of stuff to try and work out is it a good portfolio or not. The insurer who can respond to the brokers the quickest has the most new business wins. So again, there's in client retention, in risk selection, and in new business, accelerating new business closers and revenue, data makes a huge difference. I'm going to go back to a statement I made yesterday as I was doing some research as we use more machine, we should be able to quote and bind faster. Absolutely. So discoverability, we need to make sure that the information we have access to is machine, readable. Quotability, we need to speed time to market, right? Yes. We need API ready data, but also platform. And lastly, bindable. We need to make sure we can actually close the deal as fast as possible so that we can access the good risk and alienate the bad one. But we need to do that with the right data to make sure we don't fall to adverse selection. Absolutely. Right? So we talked to a lot of underwriters over the past three months, and I'm so grateful for having been able to have access to a lot of those thinkers. But, you know, one thing they actually highlighted is that they will not act on data they cannot explain. They need explainability. And you have partners with organizations. You have won awards as well. You have mentioned the Lloyd's Lab. How do you build trust in AI generating risk data inside an organization that have been doing this on spreadsheet for years? Yeah, I think the first misnomer is the industry is dealing, making decisions on 10% of data. And in order for it to transform and really trust, it thinks it needs to be able to get 100% of data on every site. And that data needs to be 100% accurate. That will never happen. Data is never perfect, right? But what we do is we always look for three sources of every piece of data. We register the source of that. the date that the data was collected from, the period over which it was collected, and then we look for an algorithmic confidence. So from those three pieces of data, we can say this particular type of data is 98.2% accurate, this particular data is only 72% accurate. Then if that piece of data is really important for the insurer, they can make a decision whether what period they want that data from, what source they want that data from, And is 72% good enough for them or is 98%? And that's the huge difference we've found by not just telling people, oh, we've got great data and you can trust us honestly. Actually telling them, there's no such thing as perfect data, but this is the best you're going to get in the market. And here are all of the sources and here is all the level of accuracy. Furthermore, here's the work we're doing to make it even better. You have great case studies. I've seen some of them and I will let you actually discuss one of the greatest, actually, transformation you have enabled an insurance to go through. I understand one of your clients has gone through, you know, doing manually surveying 10% of their portfolio to now having pretty much 100% of that portfolio dealt with through technology. Can you tell us a bit more about this? Yeah, so I alluded to this a moment ago, you know, they have a, They've had an annual requirement to do about 10,000 surveys. The maximum capacity they've ever had, if no one goes on holiday, no one's sick, and they just work flat out, is maybe 5,000. So underwriting is expecting to get 10,000 surveys done a year. Maximum ever done is 5,000. And so you've got this huge risk anyway. And then as a business, they're looking to increase their business by 50%. So the problem's only going to get worse if you continue doing the same thing. So what we've been able to do is to build that level of trust within the risk engineering team and the underwriting team that the data quality is really good, but also the sources of it and the accuracy of it, et cetera. And what that means is they've been able to, with that initial 10,000, they've been able to triage 7,000 of them digitally. So now they only have to do 3,000 surveys and that's given them the capacity to grow their business 50% so that they can do sort of 5,000 of physical surveys and they can do 10,000 of digital surveys to a high enough quality that they can trust them. It also means that the risk engineers can actually spend more time with each client and actually teach them about risk rather than just gathering data. A leading insurer shared with me a case which actually I was really surprised about. That was a case where just 7% of a building caused $300 million of losses because 93% of the machinery was halted by supply chain disruption. Now, during the research, we actually talked about business interruption and the impact of that around repricing, underpricing. And we need to actually help insurers do that better. How do you see this improving when we look at property risk data improvement and the work you are doing with insurers globally? Great case I had with the U.S. insurer. And I think they came to us to really test. They said, we've got this site. We've just had a huge loss. here's the data we knew about the site what could you have told us great our board has asked us what could you tell us and basically they they had a description of a site it was about 9 000 square meters as far as they knew it was five buildings and it was a a storage facility as far as they knew it was well managed 100 sprinkled and it wasn't in it hadn't had any flood issues within In less than a day, we could tell them that actually it's not 10,000 square meters, it's 14,000 square meters. It wasn't five buildings, it was nine buildings with two large production plants. It wasn't a storage facility, it was actually a chicken rendering plant on two fronts. One, some of the oils were explosive, and two, it was part of the genome testing process with eggs. So actually it had huge contracts for business and a huge risk of business interruption if it couldn't service the biotech industry. It also in the main plant only had 52% sprinkler, not 100%. It had had four flooding incidents in the last 10 years, and it had been fined over 10 million dollars for pollution. That information exists. So the information that is often provided to an insurer might be the very best the customer knows or thinks, but actually the data is out there. So initially it had been insured for 85 million and we could tell them straight away it should have been insured for at least 144 million. So that gap. What we then subsequently found out is it had a business interruption exposure of another 100 million. So it was in the end, it was a quarter of a billion payout for something they'd been getting premiums on of only 85 million of exposure. You have proven your model in the United Kingdom. I know you for many years, Antoine. And we have to admit it's one of the most complex property markets in the world, right? With historical buildings, you know, rigid regulations. We love our regulations. Now you're expanding overseas. You're expanding in the United States where property tax data is two to three times richer. But we have wildfire. We have hurricanes. We have convicted storms, right? What differentiates what you are doing here with what you are going to do over there? So we've developed data in the UK on 40 million properties, every residential, every commercial, every type of commercial, etc. In the US, there's 150 million properties. It's slightly bigger. And so we've been spending last year and we'll spend the next year building out a data set on all of those properties. As you say, there are different types. In the UK, there's about five different residential classes. In the US, Neo-Georgian and Colonial and all these great names of different styles and North America, South America, etc. There's different styles of properties. So there's a slightly different range of occupancies and styles. There are a lot more wood-based buildings rather than stone and brick. So there's a few of those characteristics, but fundamentally, it's a bigger market. There is a lot more data. So, you know, in the US, we've been, as you mentioned, tax records. We've been able to acquire, scrape, buy property records, tax records on 150 million properties. But these are published for citizens as PDFs. So, it's a small task of being able to read 150 million completely unstructured PDFs in many different formats, take the data out of that using AI, validate the data, put it into the model, and then make that available through APIs and solutions for the insurers. But ultimately the data exists There will be some areas where there are gaps in the data I been having a lot of discussions today about large lots where the address may only take you to one of the buildings But equally, the data that we can get from the tax records gives us there's 18 buildings on that lot and we know what the footprints are. Maps and GIS systems have those polygons for those. So it's not beyond the wit of man and AI to match which building and which record. matches two of them. So there's a bit of intelligence that will go into that. But ultimately, we're really excited because, you know, it gives us a big growth market. We actually believe on the rebuild cost side, there's even more problems. Most of the current solutions insurers have access to lack the rigidity of things like how much for demolition, how much for professional fees, how much for external works and landscaping. And we've already built models for those. Well, congratulations. You talked about shifting from repair and replace to predict and prevent, right? And we are seeing that shift actually into the industry now, and it is a fundamental change of value proposition. So when you look at 2030, what does that look like for the industry? So I definitely see a lot more augmentation of people's roles. This whole taking people from doing admin and putting them into value adding. I see the industry being able to respond more quickly to client requirements and broker requirements. The insurance industry is a digital industry. Yeah, it may not like it, it may not think it, but it is a digital industry. And actually, the information the insurance industry has should be giving benefits to far more sectors. Yeah, they are the experts in the room when it comes to any type of problem a business might find. Everything from sort of directors and officers type of issues, everything from business interruption, everything from machinery breakdown, and then ultimately to the entire building. And so I just see the insurance industry becoming a much better partner. But it's going to go through some pain in getting up to speed with using data and then using agentic tools. Just to finish off, what role does real-time property data is going to play in our industry if we can enable those chief underwriting officers to access the right data at the right time to service the right customer? Real-time means different things to different people. In your average home insurance, having up-to-date information annually is fine. But equally, most people forget I've had an extension built, and that extension's happened between the policy years and how do I get information on the extension. Many farmers forget to mention a new outbuilding that they've built or an extension to the farm or new land they've bought. Again, a lot of this can be picked up through satellites and through other sources of data. Real time in claims can often mean, I saw a great example where the roof of a research laboratory had actually collapsed. And the research laboratory contacted the insurer and said, we've had an issue with hail, roof's collapsed, and we'd like to make a claim. And the insurance industry wants to be able to respond to their customers very quickly. They want to help them with the problem, resolve the problem, and get back up and running quickly. Because actually, it reduces costs for the insurer as well, especially on business disruption. In that particular case, we were able to help the insurer to be able to get satellite images of the day before the incident, the day after the incident, an IoT, Internet of Things sensor data from around the building. We could tell that there had been hail because the temperatures had dropped locally from the weather sensors. We could tell the day after immediately this is 43% of the roof has collapsed. What we actually found in that particular case is the day before there was still a hole. We went back two weeks there was still a hole, we went back three weeks and we planted machinery on the roof doing work they hadn't told the insurer about. So in that particular case it was fraud. But equally if a farmer calls and says I've lost a load of my crop through hail, the insurer these days can do it parametrically from a satellite image day before, satellite image day after. They don't even have to look at it. and now it's just completely go, okay, well, it's not 43% loss, it's 23%. If you're happy to accept that payout, I can pay you up today rather than taking months. Now, I remember that yesterday you actually made a measure announcement. Can you tell us more around some of the development happening at Engine Tayad? Yeah, so yesterday we announced the release of our risk API. We've been developing really good risk data for the last five years, and we've been delivering it through our platform. And what we've done with the risk API is to now make it available as a uniform API that anyone can plug into their existing underwriting systems, claim systems, risk management systems, etc. We believe we have the right level of data and the right level of tracking of that data to really give confidence to the insurers now. So to realize the risk API, we also announced our partnership with Guidewire. So we're fully integrating that into Guidewire and the Guidewire ecosystem. Guidewire have 540 of the biggest insurers globally. So we're excited about that. Interestingly, today, I've had loads of people from Guidewire coming to the site and asking to look at it. And actually, they've been saying, this is absolutely fantastic. We're going to use this in every demo because for the first time, we can actually show customers how easy it is to pre-fill all of that data. They don't have to put up with the gaps in the data. So actually, I think it's going to be a really good relationship. And so we're excited on the risk side. And then next month, we're launching in the US. If a chief underwriting officer listened to this on Monday morning, what would you like them to remember knowing that property data is a big problem for them? I'd like people to understand that when they looked at this problem 10 years ago and put in the too hard to handle box. Now is the time to take it back out of that box. I think it's the level of cloud computing and cost of computing, the access to AI and agentic technologies to support people, not replace them. You know, so we can transform the industry with greater insight. The quality of data, the auditability and tracking of that data is really there. ultimately personally, and I think there's a bit of a journey to go, I would like the insurer to share the data with the broker to share it with the client so they all own that same piece of data. Because, you know, as I say, data is never perfect. A data may say this property doesn't have a basement, but the property owner knows this got a basement. And if everyone's sharing and owning the same piece of data, they can all keep it live. So I think there's a bit more sharing that needs to go on. But ultimately, now is the time for them to really look at the data. You said to yourself, you know, people are making not million dollar decision, but billion dollar decisions on thousand dollar data. And it's really important that they understand the exposure, but also their business is able to transform and respond more quickly to the market. An analogy that comes to mind is Waze. You know, all of us in some ways in our car, we use Waze. And when there is something, we actually collaborate and update to enable other drivers to actually benefit from the information. So let's move to the Waze metaphor to finish the discussion. Thank you very much, Anthony, for joining me today. Thank you very much. Before we wrap up, let's pause for a moment. What this conversation with Anthony makes clear is that the risk intelligence gap isn't a future problem. It is a present one, hiding in plain sight inside every portfolio built on incomplete property data. We are operating in a market where 93% of UQ properties are insured for the wrong amount and underwriters spend more than half their day chasing information rather than making decisions. The industry has invested in better engines, but the fuel hasn't kept pace. The real question isn't whether data matters. Everyone agrees it does. The question is who will move first from asserted data to verified intelligence, from periodic snapshots to continuous monitoring, from reactive pricing to predictive underwriting. Because the carriers that close this gap won't just improve their loss ratios, right? They will own the best risk. If today's conversation challenged your assumption about what good data looks like in commercial property, I invite you to go deeper. Download our full research paper, The Risk Intelligence Gap, for the evidence base, the executive interviews and the implementation roadmap. You will find the link in the show notes. Share this episode with a colleague, continue the discussion, and ask yourself, is your portfolio built on data that would pass a first year audit? If you enjoy this discussion, make sure to subscribe to the podcast and DM me on LinkedIn to let me know what you would like me to dive into next. This year, the Frontier Firm project is in full motion, as I mentioned before. Today, we unpack what it means to build risk intelligence infrastructure that is fit for the 21st century, verified, structured, and delivered at the point of decision. So please leave a five-star rating. Your rating and comments are so valuable. I'm Sabine van der Linden, and this is Scouting for Growth, where we don't just talk about the future. We design for it with you. Until next time, stay tuned, stay bold, and keep scouting the frontier.
