Latest / Autonomy Insiders / How Apollo Is Building the Insurance Layer for Autonomous Vehicles

Transcript

How Apollo Is Building the Insurance Layer for Autonomous Vehicles · Autonomy Insiders

• Daniel Abreu Marques: Welcome back to Autonomy Insiders. I'm your host Daniel and today we're diving into a topic that sits at the intersection of two industries that are about to collide in a big way, autonomous vehicles and insurance. You can build the safest self-driving car in the world, but if you can't insure it, it doesn't go on the road. So insurance is not a back office function in the AV space. It's a prerequisite for deployment. And right now the insurance industry is trying to figure out how to price, underwrite and also cover a technology that Today with Chris Moore. Welcome to Autonomy Insiders, the show where global industry leaders in autonomous driving unpack their real-world insights. breaks almost every assumption traditional auto insurance was built on. And my guest today is someone who has spent over a decade building insurance products for the digital economy and in recent years has also become one of the leading voices on how to insure autonomous vehicles. I have Christopher Moore on he's the president of Apollo Ibot, the specialist insurance division operating as Lloyd's Syndicate 1971 focused entirely on the sharing economy, new mobility and emerging technology. And most recently, Apollo ibott also announced the Autonomous Vehicle Insurance Program. And we will get into details of the program, but first we want to understand the bigger picture. So Chris, welcome to the show and so happy to have you on today. Chris Moore: Yeah, thank you for having me Dan. Daniel Abreu Marques: So Chris, to start for of our audience, maybe and also Lloyds of London might not be familiar territory. Can you give us the 60 second version of what Apollo does and where ibott fits in? Chris Moore: Yeah, so Apollo Ibar is a syndicate under the Apollo Group. And there's multiple syndicates of Lloyds and we all trade off our own capital, but Lloyds is the overarching sort of governance structure that sits above all of us. Now Apollo, quite a cool origin story. We have the name Apollo because one of our original investors when we set up the syndicate in 2011 was actually Neil Armstrong, the first man on the moon. And one of the reasons he wanted to get involved in insurance was because he wanted insurance to be this sort of vehicle for innovation, this enabler rather than this blocker. So I feel quite proud to still be at Apollo and working on things like autonomous vehicles. I think if Neil was still with us today, think he'd be proud that Apollo are sponsoring, know, frontier technology for the mobility space. Daniel Abreu Marques: Yeah, that's super helpful context. IBOT stands for insuring businesses of tomorrow, today. And as you mentioned, I find the origin story fascinating. And it also predates the current AV hype cycle by nearly a decade. So take us back to ⁓ the sharing economy comes to you needing a product that doesn't exist. ⁓ What the moment you realized that traditional insurance couldn't serve these new business models? Chris Moore: Yeah. Yeah, it was a kind of an eye opening moment for me about how insurance is its own worst enemy sometimes. Just in the way that we've built the industry, the cultures that we fostered, we tend to put people and companies into silos and they're the silos that we've created. If you are property risk, you go to our property division, you your boats and your yachts, you go into our marine division, your cars, you go into our motor or auto division, and we separate them really nicely by our product. When you had this in technology and this transition away from ⁓ assets experiences and usage, ⁓ which what the sharing economy is all about. it created a huge challenge for insurance silos. so you had ⁓ better breakfasts and rideshare companies and micro mobility companies ⁓ into the insurance market and they just didn't fit into a silo. And so ⁓ take like an Airbnb example, which ⁓ a huge play now in insurance and insurance is very important for their business. But when, when that sort of company entered the insurance market, you had certain property insurers say, well, you don't own property. And that's how our product works. You can't buy property insurance and then you'd have liability insurers saying, well, this doesn't feel like you're owning the sort of liability of someone else. Where's the insurable interest because isn't the liability set with the host or the user? And it was, it was really difficult. And so what it opened our eyes to was we have to create a new solution for these new tech and sharing economy companies. And once we had started partnering with a few of the big players in that space, it turns out Silicon Valley is a lot smaller than I envisaged. word of mouth spread and said Apollo are open to creating new innovative for these types of risks. It wasn't till 2015, 2016 that we first saw our first autonomous vehicle. And as you can imagine, that was in the early stages of testing. But it was something that once we had built a new program for these different share in economy verticals, building a program that was fit for purpose and specific to autonomy was kind of no brainer for us. it was already our our aspirations were racing about what this could mean for the future and it's really nice that we're now seeing the sort of large scale commercialization. I wish it had come sooner because we were waiting and ready for it because we have the products and the insurance appetite for it. it's now a really exciting space because for me, autonomy is the biggest insurance opportunity that exists today. autonomy is the biggest insurance opportunity that exists today. Daniel Abreu Marques: let's start directly with the fundamental question. Do autonomous vehicles need a fundamentally different insurance product or can they insured under more or less the same framework that exists for human-driven vehicles today? Chris Moore: I think it absolutely needs a different product. And it needs it because of the way ⁓ the existing traditional products ⁓ operate. I often get asked the question, and it's the wrong question, but this is the question I get asked, is where will the insurance for an autonomous vehicle sit? And really it's normally an insurance based question and they'll say, well, it's got wheels and it looks like motor vehicle. So is it in an auto insurance or is it a motor insurance product? And then I struggle with that because I'll say, well, those policies are built for accidents. You know, if you are in an accident, this, you know, we will pay for that loss. Well, an autonomous vehicle, if you want to be really panickety, doesn't really have an accident. It's going to make a cognitive decision. It may have a collision, but you wouldn't necessarily call it an accident because you could argue it was intentional. It was designed to do this in that sort of scenario. So that product has its flaws. So then you start thinking, okay, is it a general liability or a products liability product? Well, it could be, but most people that do products liability, this is a very different risk and their products and their wordings and their models and the way they rate products liability risks, which is typically of revenue, not of mileage, doesn't really work for a moving vehicle. And then you finally could have the argument that, well, it's a tech and digital risk. So shouldn't it sit under the cyber liability policies? Again, cyber liability is a priced for the most part with third party body injury and property damage in mind. And that's the biggest risk, I suppose, for for autonomous vehicle so it doesn't fit in any of our traditional silos and I'm not going to ram it into a silo that wasn't built for autonomy. Autonomy is such a big enough opportunity and big enough exposure that it deserves its own solution. autonomous vehicle so it doesn't fit in any of our traditional silos and I'm not going to ram it into a silo that was built or wasn't built for autonomy. Autonomy is such a big enough opportunity and big enough exposure that it deserves its own solution. And so that's what we've been passionate about in all our interactions with autonomous vehicle companies and providing them with insurance solution is we want to provide insurance solution that works for your risk. we think it doesn't matter what the traditional policies of old are. We want to create an autonomous vehicle liability solution and give that to you and make sure that if there is something that happens and you're at fault and you're liable, you know, that promise that you've given to us to, ⁓ to, to help you and to pay in that situation, we want to honor that promise. Daniel Abreu Marques: And how do you price autonomous vehicles when there's so little developed claims data? auto insurance rests on decades of ⁓ driving history, so to speak, across billions of miles. ⁓ with AVs, that foundation doesn't exist right now. Chris Moore: Yeah, it's one of the biggest challenges. the sort sentence that I chuckle about and when I speak to lots of tech companies is, know, insurance companies, you're like the car, looking in the rear view mirror. And all you're looking at is you're like 10 years of loss data. And without that, you can't do anything. And that's kind of been the challenge us with autonomous vehicles. way we look at it is, if we think somebody can scale and they can... and grow and embrace a partnership, then I'm willing to go into something with a bit of uncertainty, maybe a bit more uncertainty than I'm used to, a bit more volatility. There's absolutely benchmarks and proxies and things that we can look at to ascertain what we think of rough estimate for prices. Will we be right? No, of course we won't. It may be we've charged too much or we've charged too little, but as long as I've got the right partner on the other side of the trade, then we can correct that over time. And that's where we have a really different client engagement model at Apollo than what a lot of other insurers are used to. Insurance still works in 12 month cycles. So if you do your home insurance, your car insurance, your company's corporate insurance, you're probably insuring it every 12 months. And it's not normally the nicest customer experience. And so you only really speak to insurers every 12 months. We're probably speaking to our clients every week, every month, certainly every quarter. Because we want that exchange of information, we want to keep looking at this data on a really dynamic basis to correcting our pricing because we don't have that 10 years of dwell up loss data. it's evolved over time. And whilst, if you break down ⁓ pricing of autonomous vehicle risk, You're looking at the frequency of accidents element and then you look at the severity of accidents. We can see pretty quickly the frequency of collisions. know, we can see per 100 million miles or per ⁓ miles per thousand miles how many times an autonomous vehicle collides with foreign object versus what we know from a human driver. So we can get that frequency view. It's the severity view that is somewhat uncertain and we keep refining that ⁓ assumption time in partnership with our clients. You don't have to have the perfect recipe at the start. We know that we can build it as we progress and autonomy does more and more miles and gives more and more credibility to the lost performance. Daniel Abreu Marques: Yeah, maybe talking about this frequency severity theme. I think frequency should drop dramatically with AVs, but maybe the ⁓ severity per may increase because you have all these sensors that ⁓ would cost to replace. also maybe these nuclear verdicts ⁓ against those companies could also be something dwarf anything we've seen in traditional auto. So. How do you model that frequency severity inversion, so to speak? Chris Moore: Yeah, I think the frequency is relatively easy for us to model. The severity, like I said, is the difficulty. You've absolutely referenced one element of that is the cost of these vehicles being drastically more than the cost of a human driven vehicle. Those costs are coming down every year and that's the challenge and it will depend on the layers of tech stack you have on the vehicles. There are some companies that have a much lower cost vehicle but don't have as much tech redundancy. So like camera technology only versus camera lidar radar. then if you look at the actual bodily injury piece, think that is the most uncertain. Unfortunately, in a US setting, what we're up against is the US legal system. If you about, ⁓ we know it's true, ⁓ whether you it social inflation, ⁓ whether you it legal system abuse, reptilian theory, nuclear thermonuclear verdicts, there is some really, really unfortunate proceedings that happen ⁓ among certain plaintiff's lawyers that try to get these humongous within ⁓ and just personal car accidents or incidents involving rideshare companies. And we've seen that. mean... You don't need to be immune to it. you're ever in the United States, whether it's on your television or you're driving down the motorway, you can't help but see billboards with plaintiffs' lawyers saying, have you been in an accident? Please call me. It's a hugely lucrative business because of these verdicts. And so if you read some of the big reports, Waymo have done amazing reports with Swiss Re that they make public available. And you can see some of the frequency reduction statistics that they are releasing. and they vary and then they're obviously all the time as they do much more miles but if we assume that the frequency of accidents ⁓ an autonomous versus a human driven vehicle is a 90 % reduction or something in that region, what the plaintiff's lawyers are feeling is I have 10 % of the chances that I used to have of getting, of even having an opportunity to get one of these nuclear verdicts. So the plaintiff's bar is so against autonomous vehicles. You know, they're going to ignore the amount of the millions of lives that the technology will save. What they're looking at is you're ruining my lucrative business ⁓ of getting these huge awards. And so what I think will happen is if ever there is incidents with an autonomous vehicle, the plaintiff's lawyer will go... and throw everything that they absolutely have at that case to drive and fear monger and scare and punish one of these AV companies purely because of the economic effect it will have on their business. And that's where you can see the severity of loss, not necessarily the severity of injury, but the severity of loss within the legal system being that much higher. And that's probably the uncertainty and volatility that we take on our balance sheet as an insurer and that's really the protection that we're giving our AV partners. Daniel Abreu Marques: Yeah, makes sense. And I think many insurers also simply don't participate in the AV space at all. And this is also one reason because of that uncertainty. So ⁓ also a question to you. So what makes Apollo so different? Why is ⁓ leading the AV insurance space when so many other insurers are sitting on the sidelines? Chris Moore: It's a great question and I'm not so to think it's all about us. ⁓ asked that question to ourselves, why aren't more insurers ⁓ more active in the space? And think could be one of two things. ⁓ Yes, it be ⁓ they're worried the unknown, the volatility and the potential for loss. But that's what we're here for. If you're scared of paying losses, then you really shouldn't be in the insurance industry. And for me, I don't think that is the reason. I think the reason ultimately is either market conditions or how you are built as your business. So Apollo, I told you, was built in 2011. Actually, it was built in 2009, but kind of launched our first thing in 2011. But as insurance goes, that's a really young company. Lloyds has been around for more than 325 years, of insurance companies have been out there for over 100 years. They have grown this empire, this humongous ⁓ of business and there's not probably the same incentive or the same hunger for them to grow into a new vertical because they can say, ⁓ well I can just continue doing the stuff that I've done for so long... And I know it so well, and there's such a standard deviation around my performance, I can consistently deliver great returns to my shareholders. What's the incentive for me to go down the unknown? And there is a little bit of that. You've also got some companies, a portion of their portfolio, their empire is human driven vehicles. So you also have the, the added sort of conflict of saying, why would I support this technology that's about to disrupt all the premium that I get on human driven vehicles of which I know so well, and have made so many profits for so many years. So they're probably sitting on the fence and saying, I'm not going to accelerate the cannibalism of another side of my business. And then there's others that I truly think are a little bit fearful of just the, the cultural impact of autonomy. And I've tried to ask myself, people are so scared of this because I know I'm a really bad driver and I've been in the car with lots of my friends who I would also say are pretty bad drivers. I've been in autonomous vehicles and felt incredibly safe, much safer than I would ⁓ in of my friends vehicles but there is that... emotive relationship with driving a car and I think people put it down to independence. You know when I was when I turned 17 in the UK the first thing I did was do the driving lessons because I immediately wanted to be able to drive because I associate with driving with freedom I could go wherever I like to have my independence. I think some people in my generation are thinking, you're going to take my independence away with autonomy. And so some people say, this is 30 years away, 50 years away. It's never going to happen because I'm always going to drive because it's my independence. I truly don't believe that. And so, you know, as a purely selfish point of view, I'm glad that there's not much, you know, insurance interest, but I think it's because people don't realize how disruptive it's going to be and how quick the timeline can shift. Daniel Abreu Marques: Yeah, it makes a lot of sense. And when an autonomous vehicle, insured by Apollo, ⁓ is in an accident, say a pedestrian collision, something like that, what actually happens on the claim side? So who gets the call first and how does fault attribution work when you're dealing with software and not a human driver? Chris Moore: It's a great question. spend a lot of time with our clients ⁓ before we insure them. And then when we insure them, we're very clear that we view it as a partnership. And in that partnership, we're giving you a promise to pay a loss if it happens. But what we want you to do is control and minimize and mitigate any form of loss that you can. So if ever one of my insurers got in an accident and my risk manager's phoning me to say, hey, we've had an accident, I'll be like, you need to hang up the phone. Don't worry about the insurance. We're here to pay and indemnify. What I want you to do now is focus on controlling the situation and doing the best possible thing in your time of need. We will give you any support that you need in that, but you don't need to worry about informing us. know what mean? In any claim, what I love about autonomous vehicles from a risk perspective is in any form of collision, they're going to have a huge amount of insight into what caused and what happened in that moment. They've got so many sensors and technology and cameras, so you can easily ascertain if they're at fault or not at fault. And that means we can make a speedy resolution on a claim. what I love about autonomous vehicles from a risk perspective is in any form of collision, they're going to have a huge amount of insight into what caused and what happened in that moment. They've got so many sensors and technology and cameras, so you can easily ascertain if they're at fault or not at fault. And that means we can make a speedy resolution on a claim. we can deal with it in a manner that we know what happened in and motor claims that have gone before. Quite often we don't know what's happened and then there's a lot of pointing fingers and claims can drag over years to ascertain who's at fault and what ⁓ the of those damages are. In autonomous vehicles, you have the ability to be really dynamic on the claim side. And that's where we lean quite heavily into saying to our clients in that claim side, we're to rely so much more on you because you have all this technology and all this information and what we're seeing. with the know big rideshare platforms and with big autonomous vehicle companies is they understand that and are actually investing quite heavily in their own risk and insurance teams so they've got employed risk professionals that are claims experts so when it happens they are handling a claim good if not better than what we could do from our perspective and we absolutely accept that and love that as part of the partnership. Daniel Abreu Marques: super interesting and maybe now let's talk about a specific program that you brought to life. on March 12th, you, Marsh and Uber announced autonomous vehicle insurance program. you've just explained why AVs need a different insurance approach. And how did that thinking then translate into an actual product? Chris Moore: Yes, sure Yeah, so Uber realized and I the press release ⁓ summarized really, really well. What they realized is there is still not today a uniform approach to how insurance solutions should look like for AV. Certainly Apollo have a view and other insurers, I'm sure, other brokers have a view. But Uber essentially said, there's a lot of confusion. And when some of these newer AV companies are developing, one of the big challenges is finding the right insurance and knowing what insurance is right. knowing what the right price is, what the different coverages are and how that works. And Uber in their history have always taken a very active involvement in insurance for human drivers on their platform. have a very sophisticated insurance team. They have a very sophisticated captive. And so as they enter into this building an autonomous ecosystem on the Uber platform, they wanted to build an insurance solution that complements and supports that ecosystem. And so they wanted a very, very clear coverage. they built a specific ⁓ coverage ⁓ embeds ⁓ of the different coverages that any AV partner on themselves would have on the deployment on the platform. And then they've also made it, you know, built a pricing model they think is viable for all these companies. And it is almost like an added service for those companies, if you like, because if all those companies were buying insurance independently, you don't get economies of scale. And some of them don't have the sophistication, the buying power of an Uber. And so effectively what I think they've built is a fit for purpose insurance solution that will help grow the amount of AVs on their platform. And if we believe the statistics, which I of course do from a frequency perspective, and I believe that the severity will be controlled and lowered over time, then ultimately what they're doing is making their platform safer. Safer for me, safer for you, and everyone who uses the platform. And so I applaud the innovation, the approach that they've done, and yeah, very proud to be a small part of that journey. Daniel Abreu Marques: Yeah, and maybe you could also walk us through the structure a bit more. So the AVIP is described as a single master policy covering all key participants. So who exactly is covered and what does all key participants mean then in practice. Chris Moore: if we go back to the claim scenario, what's really difficult, whilst you have all the cameras and all the technology, in a human scenario, if we got an accident, it's pretty easy in most cases to determine it was my fault or whether it was your fault. And then the insurance then worked pretty easily. When you look at an autonomous vehicle, there are lots of different players make up the ecosystem. You've got the rideshare platform, potentially. You've got then the AV software. Then you've got the OEM that provided the vehicle. The OEM might have another partnership that has some tech built into that OEM to enable AV. Then you've got all the third tier suppliers, whether it's the brakes or whether it's the engine itself or whatever it is. So you've got lots of different parties in that AV ecosystem. If there was a claim... the scenario becomes really difficult because you could have all those different parties have all different insurance programs and they're all pointing fingers at each other and well it's your insurance program that needs to respond. No, it's your insurance program that needs to respond. They're all buying insurance for essentially the same risk and it makes the process even worse and so concept is insure all of those parties ⁓ under to make it super simple. So we're not pointing fingers. If there is a collision this insurance product will pay. It makes it seamless, cost effective, And just, yeah, it makes those partnerships come together and it creates less friction because all those partners know exactly what happens when there's a claim. And it's so good to figure that out before there is ever a claim than to start trying to figure out a solution afterwards. So that's the concept of bringing everyone together in one simplistic policy. Daniel Abreu Marques: Yeah, that makes a lot of sense. announcement also says multiple markets around the world. But is AVIP live right now? Is it this ⁓ US only today? Chris Moore: Yeah, so I can't go into the specifics of all the things, but obviously there's been announcements for the US, there's been announcements for the UK, and then you've got future announcements across Asia and Europe as well. All of that is super exciting. ⁓ same as every Lloyd's syndicate, has global licenses. And we control those global licenses out of my team. My team is dedicated for mobility. And I think that's the real exciting part. of insuring new mobility in general, especially autonomous vehicles. Well, I, know, if I was a buyer, say I'm an autonomous vehicle company and I've gone to the U S and I buy my insurance from is it? Insurance one on one. And I've got a great relationship. spent time to get to know each other. They understand my technology. We've had a lot of engagement. Great. If I then go to another territory and say, hey, insurance 101, I'd like to buy it in this territory. ⁓ we only do one. We don't insure in that territory. It's really difficult. I don't think any tech company or any AV company, any mobility company wants a different partnership in every different country in the world. mean, some of these platforms are in 180 plus countries. If they can create one solution, then obviously that has a huge amount of benefit. It takes away a load of uncertainty. It's a lot simpler from a resource perspective. So I think part of the attraction to partnering with Apollo and partnering with Lloyds is our global licenses. So we're excited about the new announcements that will come, but yeah, certainly some big ones already sort of been announced in the UK and the US. Daniel Abreu Marques: And AV regulation also vary enormously from region to region. And you mentioned that also you have US, but also for sure other regions. And the European regulation make it easier or harder to underwrite here in Europe? Chris Moore: a great question. Yeah. as much as everyone talks about the US insurance market so difficult because of the US court system, ⁓ actually find it easier to insure autonomy in the United States because it's super clear and the limit requirements ⁓ super manageable. In the UK and Europe and a lot of territories, you have to provide unlimited liability for motor. It's real consumer protection type item. but it does impede a lot of insurers in providing those types of limits. And you also have really strict motor regulations and motor laws. And so the policy wordings don't permit any and ⁓ anything that nature. So you almost to... ⁓ insure an autonomous vehicle using a human insurance construct and some of that is really difficult and so what we have to do is try and create wraparound or policies that sit alongside the traditional regulatory policy to make sure there's no gaps in coverage. That's really difficult and really inefficient for us it means you've got sometimes one, two, three different insurance policies trying to insure the same thing and it's purely because of the regulation doesn't permit a an AV insurance solution. We were trying to build that and that's part of our job is speaking to the regulators about, look here the complexities of this, of the solution here. But you know, when you look at the UK, those regulations are kind of etched into stone. They're very old regulations and it's very hard to change them. Obviously in the US you're looking at regulations that done on a state level. So it's slightly easier to get change but... I think if we don't do the change in the UK and Europe, then potentially run the risk of falling in this AV I think if we don't do the change in the UK and Europe, then potentially run the risk of falling behind in this AV world. you know, for me, when I experienced Waymo, right, so my has never been in a Waymo, but whenever I get into a Waymo, it's a lovely experience. And ⁓ even the time is longer, or even if the cost of which is higher, you know, I would probably do that. And ⁓ I think... people are really looking forward to when that sort technology comes to the UK. Certainly my wife said something to me recently and I can remember when we were talking about the growth and I said, oh, I came back from one of my trips and she asked how many times do you use an autonomous vehicle? I actually not too many times. Every time I was trying to do it, was like twice the waiting time. And my wife said, but it was late at night. I said, yeah, it was late at night. And she said, well, I would have waited four times the wait time because I'd have felt so much safer in an autonomous vehicle than in a typical rideshare. And sometimes obviously my white male privilege, I don't think and think like that. And you see why Waymo are very open with their ridership and it's really heavily women. So. There are huge other additional benefits that come with the technology and I think the UK and EU will start from a public perspective really craving that technology once they've experienced it elsewhere in the world. And so I hope that pushes the regulators in the right direction to say, come on, we've got to do something and create the opportunities for things like insurance and everything goes with it to be economically viable that they can launch and be a big part of our mobility system. Daniel Abreu Marques: Yeah, absolutely agree. And this whole AVIP construct sounds ⁓ first conversation to launch, how long did it take to structure this AVIP and also what were maybe the hardest negotiation points? ⁓ Chris Moore: What? months, but think, you know, whilst insurance isn't the most dynamic industry the world, I think it's the best take time to build. And I think was a lot of people that had a lot of back and forth. And what I really loved about that construct was there were three key parties in ourselves, the platformer and the broker in Marsh, and everyone added a huge amount to it. And I think if any one of those parties tried to do it themselves, then I don't think the product would have been as robust and as good as what the outcome was. So it was, it was a lot of back and forth. may have to look at legal, you have to bring in legal and compliance. When you look at a new wording that's never been done before, you have to test the construct to make sure it works. lots of approval processes, but I think what, what has been built as a foundation it's in its early days, right? It's just, it's just launched, but I think. that could serve as the sort of solution that sort of empowers autonomy at scale. And if that's what ends up happening, then ⁓ those months of our work will be absolutely worth it. Daniel Abreu Marques: And you just brought up Marsh so maybe for our listeners who aren't familiar with ⁓ how Lloyd's works, so can you explain who does what in this program? What is Marsh's role What is Apollo ibott's role? And where does one end and the other begin? Chris Moore: Yeah, great question. So Apollo, think of Apollo as the risk taker. So in the event of a claim, it's Apollo that pays for those claims and in exchange for that promise to pay those claims, we generate for that. So build pricing models that talk about what that premium is and it's kind of our wording that dictates which claims we pay and which claims we're not liable to pay. What a broker like Marsh does is they are sort of working on behalf of the client to access the right insurance and negotiate on the client's behalf the pricing and the coverage. Marsh being the biggest global ⁓ broker in the world have a huge amount of IP and knowledge in every corner of the planet. And so if you're looking for a global platform, they're kind of an obvious people to be considered. And also been ⁓ equally in like ⁓ a Apollo has for over a decade. They've worked with so many different AV partners. they have a team that is absolutely dedicated and specialized in autonomy. so, know, was no surprise that they were selected as the broker partner for this program. you know, for a lot it, you know, a lot of what that program looks like would be ⁓ Marsh's input. they are the broker, the intermediary, if you like, between us and the buyer. and then we are the risk taker and the people that pay the claims. Daniel Abreu Marques: also mentioned already the Swiss Re and Waymo study that they published in December 2024. It covered like million autonomous miles and it found an 88 % reduction in property damage claims and also a 92 % reduction in bodily injury claims compared to the driven vehicles. So from an underwriting perspective, much weight do you put on that kind of data when ⁓ pricing the AV risk? Chris Moore: We actually put a lot of weight on the data. we have great relationships with both Waymo and Swiss Re both fantastic companies. And what I have a lot of admiration for Waymo in putting that information into the public domain. I think one of the major, major issues facing autonomy is trust. And sometimes you have to think of yourself, know, Waymo to launch their autonomous vehicle in robot taxi. You know, it's not too... unlike thinking, ⁓ we think we should go to Pluto before we go to the moon. Like there's easier AV deployments, you you could put them in mining settings, agriculture, and you see more of that, but they've chosen to put it in a really difficult operating domain. know, driving autonomous vehicles around San Francisco ⁓ isn't easy the topography and the road network in San Francisco, but I think the, some of the behind doing is, Autonomy is never going to scale unless people trust it and I think the way to get people to trust this technology is put them in the back of the vehicle and see firsthand how safe you feel, how the driver is really, really relatively assertive and ⁓ confident of doing ⁓ you're asking it to do. I think that is a big thing and releasing those statistics adds further help. helps with regulators. It certainly is bringing more insurance capacity to the table, which will be needed over time. So yeah, we have definitely relied on those statistics I think there's a lot of credibility in those statistics, but I do think it's a great thing that they're releasing them. Daniel Abreu Marques: And maybe let's also talk about a risk that is also often brought up. It's software versioning risks. when an AV company pushes an over the air update, this changes is the driving behavior across its entire fleet. driver has fundamentally changed overnight, more or less. And from an AV insurance perspective, does that trigger then a re-underwriting event? Or how do you insure the that the insurance risk is software? Chris Moore: Yeah, I think it's a great, great point. it's why philosophically I really enjoy ensuring autonomy. If I insure typical traditional human driven risk, it's ⁓ relationship between the risk and the miles driven or time is very linear. If the statistics are you have an accident every so miles, then the more miles you do, the more likely you are to have an accident, right? And like I say, it's linear. It's not a linear relationship between autonomy, because every mile driven, the next mile is slightly safer. It's slightly better, because it's learned from the previous mile. And so effectively what you're getting is every mile done, I'm getting a better risk. And so the argument is... my rate should always be coming down. Now there is elements that work against this right there are certain things that we can discover and maybe is a type of incident or a type of accident a type of maneuver that we haven't seen before ⁓ and AV doesn't perform as expected in which case you're not necessarily upping your rate because the performance got worse you're basically amending the base assumption. And maybe that happens, but with every 100 million miles happens or 50 million miles, and you don't get one of those scenarios, you help but lean into this is a hugely safe part of operation. I sometimes struggle when I'll be at a dinner party and someone says, I would never get one of these vehicles, I just wouldn't trust it. And I've obviously seen all this data and saying, the data is incredibly compelling, you really should trust it. And I'd go to thought there, I have three children, eight, six and four, I would happily have my kids in an autonomous vehicle. I'd be very happy with that. Daniel Abreu Marques: And maybe one last question about the AVIP before we zoom out again. So, Baidu, WeRide, Pony and also Momenta all AV partners. And just wanted to ask if AVIP can also cover Chinese developed AV technology operating in the US or European markets? ⁓ Or does the geopolitical dimension complicate underwriting Chris Moore: Yeah, for me, we have agreed is to support ⁓ this sort of unique insurance solution, any AV partner that Uber decide to partner with in US. So we're not necessarily risk selecting. ⁓ have agreed to insure an ecosystem rather than individually agree certain partners. So the selection is entirely on ⁓ the partner. we're not involved those conversations. So yeah, I'm sure that they are considered at different levels, but it's never entered into something that we're considering. We've signed up to support autonomy at scale, like I say, so it's not, want to insure this type of risk versus, and we don't want to insure this type of AV risk. And I do think that that's probably the biggest challenge in front of us is creating risk assessment frameworks and ways that we can see the difference in how all the different AVs will perform. it's... Waymo have done more miles than so many of anyone on the planet. And so you have to put Waymo at a level of that is something that has huge amounts of credibility when you're ensuring that account because they've got a huge amount of data that you can rely on. That doesn't exist for a lot of other partners. They will get there. what we want to do is partner on the AV ecosystem. And like you were saying, the absence of claim data, we want to support the ecosystem. And then I'm sure that we will refine our pricing, our rating, our risk assessment over time. But for now, we have signed up to insure an ecosystem. And that was a major, major point that we wanted to make. Daniel Abreu Marques: makes sense. I think Uber is not only AV ⁓ partner you ⁓ insure. you say a bit more about other projects that you also insure and give some context there as well? Chris Moore: Yeah, we, we, we insure them, you know, there's more, list of AV companies that we insure is a lot longer than the list of AV companies we don't insure. And that's partly because of, you the amount of insurance capacity available, the amount of insurance, companies that are interested in ensuring autonomous operations. you know, I the, the constructs are slightly different in terms of the distribution model. And I think you're kind of seeing ⁓ almost. ⁓ like three different insurance constructs or maybe AV partnerships. So you've got like, you've got the, we are AV software. Then you could be, we are the OEM. And then you've got, we are the like the transportation or the mobility app. So if you're Tesla, you've got all three. So Tesla, know, we've got, we build the software within our robot taxi and we're going to have our own mobility app. That is a completely different distribution insurance model partnership model that exists there. Then you've got the of Waymo. ⁓ We do the software, but we're going to partner with an OEM and we've got our own app. That's a slightly different construct because the OEM is slightly different. So their insurance solution is going to look slightly different. And then you've got we are going to be a platform that we have the app, but then the OEM potentially and the AV software come through to us. So that's more akin to that Uber sort of scenario. So you've also got three different insurance products for all those three different constructs. And then when they obviously in different regions of the world, they're going to look slightly different again. So there's lots of different approaches that's not to mention when we still have an autonomous drones and autonomous agriculture and autonomous mining equipment all of those look slightly different as well. So lots of different ways to approach it and I think ⁓ approach everyone on their own merits. We're not betting on which of those models is the best and which one of those is going to be successful. I think one of the other really interesting topics is that ⁓ personal versus commercial. ⁓ think that's going to, that ⁓ is happening a lot more frequently than been asked to have those discussions. So would I own a vehicle that has an autonomous capability, but I only switch it on for a certain period of time. So you've got the level two plus autonomy as in eyes on, hands off versus eyes off and hands off. And would you own a ⁓ four autonomous vehicle? Depending on the price, maybe it seems like a luxury ownership, you're looking at those ⁓ and those type of models and they look very different. And a lot of the ⁓ that are looking into that space are talking about subscription models. And if they're talking about subscription models, what you could see is like an embedded insurance model. So as part of your subscription, you get insurance. Daniel Abreu Marques: you Chris Moore: It'll be really interesting to see how it works when you're driving versus when the vehicle is driving itself because the two policies need to be completely different. Like you need an AV policy and a human policy. And then where's the price point? it make it, you know, it could be that autonomous vehicles become so safe and the pricing gets so low. you might be in your car so I could either drive myself but that's going to be super expensive because the insurance or I'll let it drive for me and then the question is why do you still need the capability to drive yourself so it's going to be super fascinating and insurance will play a big part in it just find it fascinating but I say the name of the company but I was speaking to somebody and their CEO was talking and they showed what they thought the split of autonomous miles would be between personal autonomy and commercial autonomy. And I think they had it at like 50%, 50-50. And I sort of scratched my head and said, you believe 50-50? I I would have said it would be so much more commercial than personal. Personal seems like a really luxury position when it's fully, you know, when we have level four autonomy, you know, all over the planet. what the person that was presenting said, like ⁓ of us in a room had different opinions and we were all sort of arguing with each other. And our CEO just said, ⁓ we're just going to talk about opinions, let's go with mine. And we'll move on. I was thinking, okay, so nobody knows the answer of what it's going to be. So we may as well look at supporting all of the verticals and then whichever vertical wins, we've a solution. We've got to play in that from an insurance perspective. Daniel Abreu Marques: Yeah. And you talked already about different verticals that you also have in autonomy, like agriculture, mining, and so on. And one vertical is also autonomous on-road trucking. And are you also already playing in this field? Because there's also a lot of activities going on in the US with autonomous trucking. also maybe is it completely different than autonomous ride-hail? Or what are the difference between autonomous trucking and autonomous right-hailing? Chris Moore: Yeah, we insure several of the major autonomous trucking companies. ⁓ think that's ⁓ super fascinating space. Insurance for trucking is, can put company out of business. It's such a hugely important customer balance sheet. And so lowering that cost for logistics and are absolutely... absolutely massive, it's a huge disruptor. I think it's been a slower rollout, because of the cost and the size and ⁓ have to get specific routes signed off at state level and so you almost have like a hub and spoke system where It can be autonomous for a certain period of the journey, but then it's got to be human driven for the second part. And I think those routes and the scope is expanding all the time. And if you think about the operating domain, it must be easier for an autonomous vehicle to drive straight down a major highway operating on the inner city streets of LA or San Francisco. So I absolutely believe in autonomous trucking. think that's another big, important part of the work that we do. ⁓ we partner with people there in a similar vein that we do in in AV transportation or AV ride share. I think that's a really, really fascinating part. I know a funny story of, I can't remember if it was an insurance broker or an insurance underwriter that maybe hadn't done as much knowledge as they should have on the company. And they asked a big AV trucking company if they could ride along on one of the journeys. And what they didn't realize was that the actual journey, the AV journey was six hours long. So they actually drove. Daniel Abreu Marques: you Chris Moore: Six hours in one direction to wait and I have to come six hours the way back I'm sure they learned a lot but if you've been in an autonomous vehicle you you pretty much get trust it in Seconds, maybe minutes probably doesn't need six hours. think that space is super interesting and absolutely will be the future Daniel Abreu Marques: Yeah. ⁓ So it was good road trip for him. ⁓ So maybe ⁓ point that already brought up ⁓ is also cyber ⁓ and the level 4 AV contains hundreds of millions of lines of code and a fleet wide software exploit or something like this could theoretically disable or misdirect thousands of vehicles simultaneously. does the insurance also cover Chris Moore: I hope they had fun. Daniel Abreu Marques: those cyber related incidents or would that be then a separate kind of policy? Chris Moore: Our policy encapsulates third party liability arising out of cyber. We want to create something that captures all the risks of autonomy. We want one simple policy that caters for all the different outcomes a collision. We want to create something that captures all the risks of autonomy. We want one simple policy that caters for all the different ⁓ of a collision. And so that was really key for us ⁓ making sure that third party cyber risks will come up. So yeah, that's been covered under our policy for many years. And be we have not seen losses arising out cyber attacks. I think a lot of people think about ⁓ cyber and Maybe watch too many movies, but they'll think about every single robot taxi, all of a sudden driving at a hundred miles per hour and, and, all turning left or driving off a cliff in reality. Once, these vehicles leave the sort and are operating, you can only really theoretically, if you even could, you can only actually aim a cyber attack at one particular vehicle. You can't hack all the vehicles at any one point. I'm not. I'm not a tech person, what I'm told is that's theoretically impossible. So some of the systemic loss scenarios, they're almost science fiction and belonging science fiction rather than the reality. But yes, if there was a cyber-related incident that caused third party bodily injury, that would be covered in our Apollo policy. Daniel Abreu Marques: So it's like a really all in one package for the AV companies. Chris Moore: Exactly. don't want our insureds sad or phoned. Like I said in that earlier example, I don't want one of my clients phoning me and asking me, we've had this, it covered? Because we've failed in the whole partnership process then. They should know we've created a policy that covers everything that you're worried about. And if there's something they don't want coverage for, we've had that conversation, maybe accepted that that risk sits on their balance sheet. But for the most part, we want to make sure comprehensively the risks that they face are covered by our policy. Daniel Abreu Marques: And ⁓ I read projection from Goldman Sachs about insurance costs that said the per mile ⁓ cost will drop to more than 50 % or something like this over the next years. And you also said that it gets cheaper and cheaper. So if projections hold, is the AV insurance market then bigger or smaller than the human driving market, to speak, that it replaces? And you also mentioned that it's somehow maybe a fear of other insurers. ⁓ Chris Moore: Yeah, it'll undoubtedly be smaller. And that's why I saying there's a lot of insurers that are probably not looking at this space because they know that the, let's call it wheels-based insurance, because that captures commercial, that captures personal. If the size of the personal and commercial wheels-based insurance combined is sort of this much, then I think you're talking about autonomy going down to this much. Like it will be a fraction of the size of the premium, but that's because the risk will be a fraction of the size. Now, obviously insurance companies invest premiums and larger premiums and make that money work and do other things with it. And there will be other liabilities that come from AI that I'm sure we need to create new solutions for. But ultimately, the autonomous vehicle insurance will be much smaller. Now is that a concern for me? No, because I don't have a huge presence in the personalised motor space. So I hope to have a large segment of that autonomous vehicle insurance market. But yes, it's going to be drastically smaller than the motor auto insurance market today. Daniel Abreu Marques: make sense. And maybe one last question ⁓ for today. So if you could change one thing about how the insurance industry is approaching autonomous vehicles right now, what would it be? Chris Moore: So. That's a great question. It's and leaning into ⁓ new and emerging risk. ⁓ think it is disruptive, so I understand it can be quite scary because of the potential impact it can have, not just on insurance, but in the world, how we live our lives, how we get around, but ultimately, this is to save so many people's lives and insurance cannot be a hindrance to that. ⁓ for me, it's more insurers need to trust this more insurers need to join the conversation at some point because of the legal system in the US we need lots more capacity, much more capacity than Apollo are going to be able to deploy on their own. And I welcome any insurers to say, hey, we're not in this space, but we'd love to discuss ⁓ we can join and ⁓ support some of your clients. ⁓ ⁓ it will be. so disruptive to every industry we know. So if you're writing marine insurance, there's going to be autonomous boats and that is going to affect your market. There's going to be autonomous construction, there's going to be mining agriculture we've already talked about. So every single ⁓ vertical we know today is about to be disrupted. So if you're not part of the conversation, ⁓ urge would be insurance needs to be. And I get a question a lot as well, and I'll probably end on this. I get asked, who's your biggest competitor? And a lot of people are expecting me to name another insurance company. My biggest competitor is my client. And what I mean by that is if I don't create a solution that is fit for purpose, both from a coverage and a price perspective, and I don't take into account how safe they are quick enough so I don't give them the price that the true risk deserves. Some of these companies with huge balance sheets in their own right can create their own insurance company. And we've already seen quite a big player in this space have their own insurance. It hasn't performed very well, but they have their own insurance company. I do not want my clients to build our insurance company. I want them to buy insurance from us. But if we don't provide the right product, then they could insource insurance. And that is for me. Daniel Abreu Marques: Yeah, super insightful. And Chris, thanks again for taking the time today. It was a very insightful conversation. And I have been wanting this conversation for a while because insurance is one of those topics that most people in the AV industry treat as a checkbox more or less. But as you've made it clear today, it's far more than that. So it's a structural enabler more or less, a structural barrier depending on whether the industry gets it right. Chris Moore: Yeah. Thank you for having me. I hope I've shed some light on how we're trying to tackle the problems and what certain major players are doing. I know there are load of questions about certain insurance programs, but hopefully that cleared a lot of it up today. Daniel Abreu Marques: Absolutely. So thanks again Chris for all the insights and until next time. Chris Moore: Thank you. Yeah, thank you very much.