Are you maximizing your telematics efforts?

– Nathan Golia, Celent

– Henry Kowal, Arity

– Sanjiv Ghate, Mobilisights

– Rick Lanter, CSAA

I’m Nathan Golia. I’m a senior analyst for Celent, the property and casualty group for North America. Celent is an advisory firm that focuses on helping insurance companies make the best tech decisions possible. I’m going to let my panel introduce themselves, starting right here with Henry. Let’s go down.

Okay. Hi, my name is Henry Kowal and I’m a product director at Arity. For those of you that aren’t familiar with Arity, we have the largest driving dataset tied to actual insurance claims in the world, and previous to Arity, I used to work for The Hartford. I was a new product development director there, and I was responsible for building out and launching and managing a smartphone-based telematics program in over 40 states.

Yeah, your [microphone] is on. Mine was not. That was the problem. Hi, everyone, I’m Nate. I’m from Celent, as if you didn’t hear me before. Sorry. No, you can go ahead, Sanjiv. Thank you.

Hello, everyone. My name is Sanjiv Ghate. I am senior vice president at Stellantis. Stellantis is owner of Jeep and Fiat and 14 auto OEM brands together. I’m also appointed the CEO of Mobilisights, which is a subsidiary company of Stellantis created specifically to enable a smarter world through licensing of telematics data. Insurance is really a big vertical for us and we really are looking forward to bringing all that information and data knowledge to make insurance a better place. Thank you.

Should I test as well?

Go? We got it. Okay.

Rick Lanter, senior vice president of product strategy and product development at CSAA. For those of you who don’t know, CSAA, we are part of the AAA network of companies and we write insurance on behalf of AAA and leverage the AAA brand within our business. We’re in 23 states, so we’re not in every state in the U.S. and look forward to talking to you guys today.

Great, thanks so much. We are going to start with you, Rick, at the end. To just start by setting the stage a little bit around the auto insurance industry. What are some of the challenges that auto insurers and policy holders are facing and how do you see standardizing telematics data as a way to help address some of those problems?

Yeah, if I say inflation, that sounds generic at this point I think, but it’s probably been…I’d love a show of hands for those of you who are carriers in the audience who think it has been the hardest market that we’ve ever worked through. I know it’s been the hardest one that I’ve definitely worked through in my career. And I know we’re in the auto insurance track or the auto track, but it’s been property well, right? So we’re getting squeezed in both of our major lines of business and personal lines, insurance. From an auto perspective, obviously we’ve seen driving behavior change post covid, we’re seeing spikes in frequency, we’re seeing sustained inflation, inflation being higher, which obviously puts pressure on severity and it’s been a big challenge for the industry. I think we have probably a bit of a base rate issue that I think we’ve all been working through along with a segmentation issue that we’ve all been working through, and from an industry perspective and then you layer rising interest rates on top of all that and then a lot of the OEMs changing their business models and us adapting to that as well.

And it’s just been a lot of volatility over the last three or four years that we’ve been working through.

Yeah, sure. Rick, I don’t know if you wanted to jump in on that question as well or anyone else. Okay, so…

We are coming at it more from the data side obviously, and I think the challenge is that Rick is outlining are clearly there for everybody to see. We are looking from the OEM side as to we have increasing fleet of connected vehicles, which have a fundamental…two things happen in these vehicles. There are a lot of ways, there are a lot of instrumentation devices to collect data, but there are also connectivities to transmit data. So how can we make insurance experience more efficient, both for the consumers as well as the insurance industry or the providers? So that’s really what we are focusing on. That’s where we are coming at it from.

Henry, would you tell us a little about how you see the universe of telematics data that’s available and what makes it a robust and mature data set that can help meet some of these challenges that Rick was talking about?

Sure. Well, historically though, telematics has had rather low penetration in utilization within the industry. The typical approach has been, “Sign up for my telematics program. I’ll give you a participation discount.” “Drive with my app that you’ve downloaded till your next renewal and you might get a better discount at renewal.” That has been the same kind of program construct for over a decade, right? Because that’s the same approach that was used back when dongles were being used for telematics and now even with smartphones. But I think part of the problem is because there’s a lack of telematics data at scale, or at least there was. When Arity was founded, our main mission was to be able to gather telematics data on every driver in the us. Now that’s a pretty ambitious and lofty goal, but we’re making great strides towards that. To date, we have about 40 million connections within our network, and that represents about 50% of the U.S. driving population. So now we actually have telematics data at scale. So I think that presents insurers with the opportunity to use telematics data not only within that renewal use case but across the entire insurance lifecycle. So whether that’s marketing, whether that’s point of sale pricing or underwriting, even at renewal, but also even claims.

Sanjiv, can you talk a little bit from the OEM side about just how that data is evolved and what you’re able to acquire now and maybe operationalize?

So the list of potential data attributes that can be collected is really long. I think that what we are looking constantly to do is Stellantis itself has 14 different brands and we have 13 million connected vehicles today. These vehicles come through different generations and different tiers make models. So if you go back to Hendrick’s point for an insurance company or an insurance company technology partner to look at this data, they need a little bit of standardization. You can’t deal with so many different attributes. So we are working with closely with the insurance providers, carriers, to build that ecosystem together because just putting out data doesn’t really help. There has to be a market, there has to be a product to use that data. So we are starting out small. We are looking at continuous monitoring as one way of saying, okay, “Who is really the good driver?”

We are also looking at the data to make the driver better. At the end of the day, we are a B2C company, we are an OEM company, so we want to induce a good driving behavior through the feedback loop. So while the insurance work as a monitoring incentive to drive better, there is also a way to feed back into the user what is a good behavior, what is a dangerous behavior. So we are also doing that. And lastly, we believe that collectively our goal is not only to reduce the insurance, the goal is to reduce the accidents as much as to promote a better driving behavior, that the end goal is to reduce accidents. So the data that goes into that is data such as distracted driving, hard braking, fast turn. So you can start very basic. We can advance it to, “Is this driver a typical speeding driver? Is this driver doing a lot of tailgating?” so the connected car data can begin to shed lights on a lot of these things, which could be useful.

For sure. So yeah, we’ve sort of established that we have this hard market in P&C insurance, which is something that we at sell have been tracking quite a bit in terms of its downstream effects on tech strategy. And then we have now a broader universe of availability of data. So let’s jump into some of these use cases starting with Rick. So let’s start with that marketing section. How do you see, from your perspective as the carrier, the opportunities to leverage this and the point of sale quote pricing realm?

Yeah, I think Henry hit on a really good point that we talk about a lot at CSAA, which is, we’re reaching this inflection point where we were looking for this data at levels of scale where we could use it across a plethora of use cases, the use cases that you just mentioned. So maybe I’ll hit on point of sale. Data is going to be coming to us from a lot of different places, whether it’s our own app, whether it’s an OEM type of a partnership, whether it’s a company like Arity that you would want to partner with. And so the challenge that I think we now have as an industry is, how do we use all this data? What are the best use cases? What are the most predictive elements, scores, sources, et cetera? And then how do we apply that at point of sale once we do figure that out, the ability to better segment customers will obviously be something that we’ll be able to do straight through process.

Customers potentially who sort of meet our underwriting standards and we know have good driving behaviors, being able to target customers more effectively based on driving behavior. And I think for a long time as an industry, we’ve been looking for ways to be a little bit more transparent around how a final rate gets generated, right? Helping agents understand it, helping customers understand it much more effectively. There’s no better parallel than driving behavior, right? Customers intuitively know if I drive more safely and drive better I’m going to get better rates in preferred underwriting, right? That’s just common sense. And so, from that perspective, it also solves something that the industry has been searching for a while as well, but I think we’re just scratching the surface of where this is going to actually go. And at CSAA, we’re just trying to be on the playing field partnering with a lot of different firms so that we can get to that end destination of being able to most adequately rate underwrite and market to our customers.

I dunno if Henry assigned you if wanted to dive in on the marketing point of sale, any applications that you’re seeing or testing or hearing from other clients?

Yeah, there’s also the opportunity for marketing and using telematics data for that use case, and that’s probably something that hasn’t been explored too much within the industry. But again, I think the barrier to that was the lack of having that data at scale. But there’s definitely an opportunity, especially within the insurance industry. When Rick was talking about some of the challenges that are occurring, he talked about the raising of rates for insurers to try to achieve profitability again and kind of sustain that profitability. But what you’re also seeing in the marketplace is that consumers have reached an inflection point. The rates have gone up so much that it’s reached a point where it’s like, for them it’s unsustainable, right? They’re having to contend with other inflation-driven issues such as grocery bills and utility bills – and now their auto insurance bills are much higher than they used to be.

So kind of an unintended consequence, I would say, are raising the rates is that consumers are starting to shop. And we’re seeing some of the highest rates of shopping that we’ve seen in a long while. And I would say that that’s just going to continue, and that’s going to increase.

So that’s where marketing and better being able to leverage your marketing budget and use it more efficiently by using telematics data is a really great opportunity. What I mean by that is that you’re able to identify, using telematics data, the risk profiles of customers that are out there, or leads that you would want. And if I am a carrier where I may prefer the better drivers – that seems intuitive, but we have other carriers that are focused on kind of niche and they’re focused on substandard – but I can say, okay, I can use these marketing dollars and target these good drivers as leads because these are the type that I want to bring to my book of business. So I think that’s really a useful and kind of important way of how marketing and telematics can be used in today’s market.

It’s funny you mentioned that because I, before I was selling, I was a reporter, and I was getting all these industry news, your rates are going to go up, your rates are going to go up, your rates are going to go up. And I got my renewal in December, and it went up, and I looked and it was all concentrated in one of my two cars. I have one car that is three years old and is nice, and I have another car that is 10 years old. That one did not increase my rate there. It was a new car with all the bells and whistles and stuff that really doesn’t cost you going up. What’s easier for me to do, change my car or change my insurance? So I didn’t, but…

Never thought about it. And it’s interesting because even just from where we’re all customers, and so from my own customer perspective, my rates went up as well, but guess what? I work from home, so I rarely drive my car. It just kind of sits there in the driveway. But it was just this, “Let’s increase rates kind of wholeheartedly across the book of business” versus “Well, do we know Henry and we do we know what his driving behavior is and does it really warrant for him to have his rates increased or is he a really good driver and is this the type of driver we want to keep on our book of business?” Because guess what, Henry did? Henry shopped for new insurance and went to another provider.

Sanja, did you want to jump in on marketing or anything that you’re seeing from the OEM side for possibilities there?

Yeah, I want to precede that with a very, very important thing. I think marketing is where the user consent is, drives the marketing decision for us. So again, as a consumer-oriented company, we always rely, so we do it on the consumer side, we really rely on the consent user needs to tell us what she wants to do. If the user wants us to really, if the user is shopping for insurance – for example, let’s say she’s up for a renewal and user is asking us to look for a better insurance quotes. We can work with all our partners and say, “Hey, this is a good driver. Here is the behavior the user has consented to share. Do you have some better quotes for her?” That’s one way to do it. The other way for us to do it is really on the business side or B2B side, is enable the market. This user could be going to multiple insurance providers. She has taken the step to go to the market, and then she is again, consenting, via the carriers or via the people who are offering the insurance, to use the telematics data. So marketing has a lot of significance, but we would rather it start with the user saying, “Hey, I’m looking for something or I’m consenting to something.” That’s really the starting point for us.

Well, circling back to consent for a second, and I’m going to attempt to start up to unlock my phone, get the next question from memory. This will be easy, Henry. So we talked about the marketing sort of front end applications of telematics data. What about during claims or customer service? Where can we see that come in?

Yeah, definitely we could utilize telematics data within that claims use case, and I would bifurcate it between, I’ll say safety, so in terms of crash detection. And then the other aspects would be more the claims processing or the claims adjudication. So we have the sensors in the phone, we have the technology, but we also have the machine learning and the data science and the algorithms to detect when a crash occurs and when that crash occurs, we can notify emergency contacts, whether that might be a spouse or a parent and so forth. And then at the same time, there’s the opportunity to dispatch emergency services if it’s actually warranted. Now, within that use case, if you think about it, when a policy holder has a crash, that’s a key moment of truth for that individual with their experience with their insurance provider. And I think I saw something like a stat of 30% of policy holders that had a poor claim experience switch providers.

So the opportunity for an insurance carrier to be able to provide this type of service, of safety and security to the policy holder, is really important. The other aspect is we’re able to capture within the crash the data right before the crash actually occurs during the crash and then after the crash. And so an insurer can ingest this data and use it as part of crash forensics analysis to help adjudicate the claim. And what’s the benefit there? Well, the likelihood that they could close the claim as fast as possible. Because the longer it takes to close out the claim, the more expensive it is. So again, it’s a win-win in that scenario too. It’s a win for the insurance company to keep their claims costs down, but it’s also a win for the customer. They have that claim settled much faster.

Rick, from the carrier perspective, can you talk about the appetite to apply some of those things? Is that something you’re looking into from CSAA’s perspective?

Yeah, I think the claims use case is really exciting. It’s one thing to know what happened at the time of crash, how fast someone was going, did they swerve, braking, all of those things. As cars also get smarter and we start sharing data back and forth with OEMs, the ability to really dissect what’s wrong with the vehicle as well becomes a really exciting use case. So imagine a world, and I think we’re going to talk about generative AI in a few minutes, so I’ll leave that part of it out, and how that would help. I don’t think you can have a panel this year without talking about generative AI, but we will get there. But I think the ability to not only know what happened, but also diagnose the vehicle, be able to adjudicate the claim really, really rapidly. You mentioned, I think you said 40% of customers leave due to a poor claims experience. What a great claims experience. If we know exactly what happened with your vehicle, exactly what happened at the time of crash, and we can get you back on the road again sooner than what we’re able to today. Maybe we also dispatched some AAA services as well in that process, to give a shameless plug, but I think that’s the exciting use case, at least from our perspective. Are we there yet? No, we’re not there yet, but I think that’s the destination that we’re all going to be driving to.

Can I add something quickly? Yeah, yeah. I think this is interesting. I was having this conversation with Rick literally like 10 minutes back outside. I think some of these lateral use cases, whether that’s roadside assistance or preventive maintenance, we are actually at a point that those services can in fact be supported at a scale. So if you’re out there looking for those kinds of use cases, happy to have that conversation, too. Also want to quickly go back to Henry. I think the broader point of forensics or crash detection or prevention, we obviously feel that the car is the source of truth and we know the degree of details as to what actually led to the crash, and we also know what got damaged in the crash because we have a very deep understanding instrumentation in the car. So we are all plugging information in the overall picture here, but we feel that as an OEM with native access to data and the one who built the car, we can really provide the source of truth in a lot of these cases.

So we’ve talked about these potential use cases. Let’s jump to some of the questions that people in the audience might have. Starting with you, Rick, what are some of the challenges inside an insurance carrier to operationalize in this data? What does it take to be able to do all these things that we’re talking about and how should people start in implementing a program?

It’s a challenge that I think we’re all trying to solve. I think it’s one thing when you just had your own app or you had a dongle plugged in the vehicle and you’re recapturing data. We think about the future use case of data coming from a lot of different places. It’s a different type of a challenge. You mentioned customer consent, right? So that’s one area that presents a challenge when you have data coming from a lot of different places, making sure that you have the right consent. I think the regulatory environment also is going to be something that we’ll have to keep working through from a carrier perspective as well. Thinking about these different models, what these different data sets are saying, departments of insurance, I think we’re seeing some trends that say they don’t want us to just understand our models, but they really want us to really understand and underwrite the models that we’re using as well.

So that becomes another thing that we have to think about. And then, from our perspective, it’s how do we apply the right data source at the right time so that we’re able to understand the risk most effectively? Whether that’s – I can mention OEM data, data from Arity, data from your own app, which of those is most predictive? Which of those should you use at certain points in your quote flow? How do you think about that? So those are all challenges that we are working through from…how to operationalize it. And I think that’s just something that we’re going to have to keep working through as an industry. But when you get data coming from a lot of different places, it can be very predictive and very fun. It can also be very expensive. And so making sure that you’re optimizing your orders and how you’re thinking about that data is going to be something that we’ll have to solve as an industry as well.

Hey Henry, did you want to jump in on that question? Just what are you hearing? How do you try and make that transition easier for carrier customers?

Yeah, so earlier I talked about the barriers of being able to have greater penetration of telematics, but also greater utilization and not kind of just focused on that one renewal pricing use case. Now with telematics data at scale, we have the opportunity to do that, and part of why Arity undertook that mission was to break down that barrier for insurance, for insurance carriers. Because they were pretty much constrained to the policy holders that were coming down their funnel, buying their insurance policy, and then enrolling within a telematics program. But that’s pretty finite. I mean, we have Allstate [as a customer]. Allstate has about two million, I think, policy holders, give or take, within their telematics program. But if you think about the number of policies sold within the U.S., auto insurance policies sold in the U.S., that’s still a really small fraction. So our view is that we’re helping to enable the power of telematics and its usage for insurers.

Again, not only for renewal, but across these different use cases, including point of sale, which is a really critical moment in time. Why have somebody download an app and tell them six months later like, “Oh, you’re actually a pretty good driver. We’ll give you a 30% discount.” If you’re able to access the Arity network of 40 million connections and determine what their driving behavior is like right there and then, you can give that customer that 30% discount right there. So I think those are ways that we’re viewing how we’re helping to operationalize the use of telematics data within the insurance industry and across these different use cases.

We’ve mentioned consent a few times. I know Sanjiv, this is something that you’re, I just wanted to tell a quick story, which is about my dad. So I went to visit my dad. He lives in Rochester, New York. I walked in and I looked on the dining room table and he had a little box from his insurance company with a telematics device in it. This was less than a year ago. And I was so excited because I’m like, “Oh dad, this is what I’ve been writing about for a living for the past 15 years. I’m so excited that I’m actually getting to see one in action from someone that’s not me.” I had tested a couple of them myself, and I was like, “Great.” He’s like, “Nathan, Nate, Nate, that’s the return box. I got to send it back. I never installed it.” I went to put it in and I felt queasy about it and then I just didn’t put it in.

And I said, “Okay, well you work at home, you could probably save some money.” He’s like, fine, fine, fine. So he sent it back, re-requested it and he did install it. I don’t know, I haven’t checked in with him, it says, but I will say that when the next time I was there, he had gotten it back. He said, let’s install the device together, plugged it into the car, and then we couldn’t close the cover on the OBD-II port, so we to duct tape it back. So that’s why I said that’s why you got to switch to a mobile version.

Anyway, but given your example, so switch to mobile version, that’s an instance where I don’t know if your dad is a super user with his mobile phone and whether he uses different consumer apps like, say, a Life360 or a GasBuddy or a Get Miles type of loyalty rewards app. These are some of the partnerships that we’ve established to collect the telematics data. So it has nothing to do with insurance. It’s the apps that consumers are using every day and they’ve consented to share their data in exchange for some value. And so an example of that is, within Life 360, we power their crash detection feature. So for that feature, consumers say, “Yeah, I’m willing to exchange my driving behavior data for the value of that feature.” But then I guess my other question is, if your dad’s not maybe a super user of consumer apps, does he maybe have a connected car? Right? Because if he does…

By next year most cars will be connected. So I think this will, yeah…

That’s another opportunity to access.

We are at that inflection point where…

We’re getting to Sanjiv, the question I was going to pose to you, Sanjiv, from that point was like, how do we get people to get past that? What are you seeing there? And I think part of it is you were talking about building into the car.

I think, like most features, the technology-driven features, they come at a higher extreme first and at some point they become fairly mainstream. So I think we are at a point, I cannot give brand by brand, by brand within Stellantis, but almost everything will become connected. Users still have to activate the service because activation of the service is still, that’s where again, we come back to consent, but I think there are a lot of the friction points that we just talked about, whether it’s dongle or something else, duct taping it, et cetera, et cetera. We feel that in a modern vehicle, connectivity should be native.

It’s like, do you ever have a phone without data today? You don’t use phone just to make a telco call. So for a car, and it is just customization, user benefits, user safety, there are so many reasons you want to have the connectivity. That’s why we are going back to your original question very briefly. The industry is emerging. We are all sort of evolving together. And even if you look at point of sales, should we be going towards standardization? Should Stellantis and Ford and GM and any big – Toyota – should they come together and create standard of data exchange? I think that will be a good idea for the industry. Some standardization needs to happen. And then, finally, how our insurance providers or people are processing the data. There’s all a latency issue when you have a live customers trying to get a quote in front of you. So those are some of the early challenges, but every industry goes through those type of challenges and those are solvable problems. We just need to work on them together.

Rick, as someone who’s on the front lines of this, can you talk a little bit about how you put that question in front of people? What’s the regulatory environment around that? Fill us in a little bit on your view of the consent and regulation framework around telematics data.

Yeah, I see some of my team in the back, and this is something we talk about I think at least weekly if not daily, as we’re thinking about different types of partnerships. I think I mentioned data coming from a lot of different places, and it would be great for that to be streamlined from a customer perspective or from a consent perspective, but that’s not always the case. There are certain partnerships that we have where we feel good about the consent that’s happening within the partnership model. There’s others where we don’t feel as good about that. So we’re actually integrating within our point of sale infrastructure a way to ensure that we are gaining consent. I mean, we’ve been doing this as an industry for a long time with FCRA, so it’s not much different than that. And so we’re just making sure that our quote flows and our user experiences are gaining consent and doing it in an elegant way that doesn’t introduce friction for the customer or the agent. I think that’s a big point for us, and the way we handle it within our agent experience and customer experience is a little bit different, but that’s something that we spend a lot of time working through. We’ve been testing around how to do this effectively, and I think it’s just something that as we strike more partnerships and we see more data coming from more place is just something that we’ll have to keep an eye on from an industry perspective.

Henry, do you have a view into the regulatory side here and how that’s changing?

Yeah, I think just to kind of build on what Rick said, I think it’s having some guiding principles around the consent. So as a company, Arity is privacy-centric. And we want to be transparent about consent because this is an individual’s data, it’s personal data about themselves. And so we want to make sure that we’re transparent in how this data will be used, when it will be collected, and then how it’s safeguarded using best practices to make sure that it’s safe, secure, and private. But I think it’s important to have these guiding principles of: we will be transparent, we will convey to the consumer of how this data is being collected, what data is being collected, and how it’ll be used. But at the same time, I think it’s really important to convey what that value exchange is. What are you receiving in return for sharing this data? And in our experience, if individuals see value, then they’ll share their data.

Are you at all tracking the regulations? We talked, we’re going to talk about Gen AI more specifically in a second as it applies to this, but the fact that there’s going to be a lot of attention paid to this, to algorithms and algorithmic models, do you see that interacting with these sort of existing models of data collection and optimization? Is that something that you guys are looking at at all or tracking?

Rick Lanter, CSAA:

Yeah, I’ll go. Yeah, I mean, I think generative AI means everything and means nothing. I think, at this point, I think we’re all trying to solve from an industry perspective, we’re all playing in that sandbox, if you will. We’re all trying to figure out what it’s going to mean. We’re all trying to figure out how we build the right data infrastructure. I think where we’ve spent a lot of time is, I can build all the models that I want, but if I can’t operationalize those models effectively, then it loses value. So we’ve been spending a lot of time thinking about how do you build a foundational infrastructure and business model that can take advantage of generative AI. So not only are we building the models, we’re also building the business model that can move rapidly. As we gain insights, we can implore those insights into the marketplace much more rapidly. Now a rating use case is different where you have to file, obviously, but I also think that we were talking to someone yesterday about continuous, and depending on who you talk to, different people have different, 10 different people have 10 different opinions on continuous and the value of continuous monitoring from a telematics perspective. But I believe that we’re going to be able to identify new rating variables through that data. I believe we’re going to be able to identify new underwriting standards through that data, and generative AI is going to help us really do that.

But I think you also double down on this. You also have to have the business model set up to take advantage of what the AI model is telling.

I’d like to get your opinion on that.

I think I’m a hundred percent with Rick. I think generative AI is sometimes over indexed. It becomes the shiny penny that everybody wants to look at. But if you look at AI broadly, there are a lot of real applications even today, and I will just give you two. AIs in general do really good job at pattern recognition and also understanding what is odd in the pattern. So if we are, one of the important application of AI is really detecting objects on the road.

Yes, there are road signs which could be changing Jira, for example, speed signs change dynamically. AI can catch it on onboard of the car using the camera system, detect it to be, let’s say 45 kilometers per hour speed limit and not just transmit the image because that will be too data intensive, data intensive, but interpret it and then communicate the metadata that this is 45 [kilometers] per hour. But imagine also the applications where the same technology can detect hazards on the road. Hazards go on the road, there may be a mattress lying down on the road or some other way you can detect that with AI. If you go back to our original or other thing we talked about, we have EV vehicles now, batteries are becoming sort of a everybody’s top of mind issue. AI can detect potential issues with the battery proactively, and we can turn it into predictive maintenance keys. So yes, there is a generative AI and that has significant applications. There are fundamental applications that AI that are equally or very, very powerful and interesting.

Do you want to jump in on that before we go to the audience questions, of which there are several?

Yeah. I think one aspect, too, is with regards to AI, and plus Rick could speak to this as well, but the insurance industry has been using this for decades, particularly in terms of their modeling and so forth. Generative AI is just kind of taking it to the next level, but in either use case, it’s what’s required is data and a lot of data to be able to train that model, test that model, and then iterate and improve on it. And that’s exciting about…with telematics data, is that we could actually use this data to train the models. The other thing I’ll say, too, though is from a consumer perspective, I think it’s much more intuitive from a rating perspective to say that we’ve determined your insurance, your auto insurance price, based on the way you drive and whether you’re a risky driver or not, that just seems intuitive.

It seems right, versus, well, we used some aspect of your credit information. We also took into account your education. We may have asked about your occupation and so forth. And suddenly when consumers become aware of this, and we’ve done some research around this, they’re astounded. They’re like, what is this? What does my credit information have anything to do with my auto insurance rate? So I think just moving towards rating variables and factors that rely on driving behavior information I think is just something, at least from a consumer perspective that makes much more sense. But then from a business perspective, we see the predictive value. I don’t think there’s anybody here in the audience that we need to convince that telematics driving behavior is predictive of driving risk. So we see the value of that.

We do have some audience questions. Thank you. I almost was about to toss it to the audience and I remembered the iPad was back here. So glad I caught myself. So I think the first question I want to start with is maybe something maybe more for you, Rick. The question is, how can carriers learn to connect data across customer life cycle? And it really gets to the second part of it, which is, well, how do you test it? And so I guess maybe a little bit of, how do you go about testing and putting in these things in, what’s your strategy for onboarding this kind of, these kinds of new ideas?

About three years ago, we went on a journey to build an infrastructure where we could rapidly test data. Part of it was because we felt like we were behind the industry in our rating algorithms, in our underwriting, and so we needed to catch up. And so we built this – and we actually run data through a traditional technology sprint cycle where we’ll bring in data, we do a two-week sprint, we can see if that data is predictive. We then do readouts on that. We then go back and we look at it again if we need to. And we run it like a technology project almost where we’re kind of rapidly bringing in data. And so having that infrastructure that we’ve set up and that process that we’ve set up enables us to take data from a lot of different places and test it effectively.

I think at one point, we tested in one year – this is on the property side, we brought in 28 different new data sets, hundreds of thousands, millions even data elements and tested them all in a single year. But while we were also doing auto research as well. So I think it’s about, one, identifying the data that you want to test, having the right process and being diligent about that process. And then obviously the results are going to be what they’re going to be. But being able to do that rapidly for us at least, has been a differentiator.

And I wanted to jump in on that. There’s one that I think Sanjiv, this might’ve come in right before you talked about it a little bit, but it was about not all OEMs have the same kind of standards or telematics data and how carriers deal with this. So maybe you could just talk about or reiterate what the sort of plans are in the OEM industry to start to standardize data. How serious are those conversations?

So before I came into the auto industry, I spent a lot of years in the advertising industry, and that industry coalesced around standards very quickly. I think part of the reason was there are so many players that you cannot really function without a standard. I think one thing that is leading to delays in the auto industry, because there are just five or six big players.

So I do not have a plan today, but I’m looking forward to mobility sites was literally created earlier this year. So we are still in that very early sprint phase, but I’m looking out to have conversations with my counterparts in other big OEMs to really create – because at this point we’re not really competing. We are trying to facilitate new applications and new services and we want to speed up that innovation and having different formats, different way of packaging data, different way of looking at data that just creates friction, which we can all live without. So no good answer right now, but this is something I certainly think we should work together with other OEMs and I’m looking forward to it. So if there are people in the audience, let’s talk.

And I think from Arity’s perspective, we’re looking to help facilitate that. So I’ve talked about mobile data, but at the same time we’re a data company. So whether telematics data comes from OBD to dongle devices, mobile apps or connected vehicles, we want that data. And so we’ve been establishing relationships with OEMs so that we’re able to collect this data coming from the vehicle. And I definitely think there’s a role for Arity to play here because if you think of mobile data, so I’m going to exaggerate it a little bit, but there’s like thousands of different phone models out there. You have your Android, you have your Apple, and then you have, I don’t know, again exaggeration, but gazillion versions of Android phones, whether it’s Google Pixel, Samsung, blah, blah, blah, blah, blah. And then all these phones are running different operating systems versions of operating systems. At Arity, we’re able to take the data off these different mobile phones and kind of bring that together and help normalize it and use that data for various insurance applications. So I think we could definitely kind of take that expertise, that analytics capability, and apply it to OEM data as well. And Sanjiv, I’m glad you mentioned about all the OEMs here, you’re not in competition. It should be a collaboration because actually I’ve heard that from some of our other OEM partners as well. So I think it’s potentially collaboration with OEMs but with Arity as well.

The last question, there’s two questions here that I kind of want to meld into one ’cause I think they sort of get to the same point. One was, do drivers feel like use of their telematics data directly leads to a better claims experience? The other one was how does a best-in-class carrier use telematics data different than their peers? I think the question is here is we talked about we need to get consent, we’re going to operationalize across the enterprise. How do we pay it off? What’s the vision for the customer experience that they’ve given us all this data and we’re giving them what back? Is that something we can close on? Rick? Maybe just talk about how do you want people to feel once they’ve given you all this information?

Yeah, it’s all about the customer, at the end of the day. I mean, that’s what we’re here to do is serve our customers. And the insurance product in and of itself is a claim. That’s the moment of truth. And so we think about point of sale use cases, the ability to give someone a seamless experience if I know you’re a great driver and we get to the point where we can truly correlate that to an effective price and that data starts to overtake some of our other predictive variables, well now I’ve given you a really solid quoting experience. Maybe I’ve straight through processed you. Maybe the data tells me you’re a perfect risk, and I don’t need to ask you any more questions. So there’s that point of sale use case. And then from a claims perspective, I think we hit on it earlier, but the ability to eventually take an FNOL that tells you literally everything that you need to know about what happened within the crash is a differentiator. And that’s going to lead to a much more effective claims experience. And so I think we’ve hit on various points of that question throughout the day, but it’s all about delivering value to the customer, whether it be at point of sale, how we think about servicing them, how we think about engaging them through the renewal period, and then obviously at that moment of truth, the time of claim and being able to act swiftly fast and get them back on the road within their vehicle.

Sure.

Any last words?

Yeah, I was just going to add, too – and again, there’s something about the safety and security aspect at least. So I was on the insurance side, but I’ll speak as a customer. So it’s not just the, “I’ve paid money to insurance company and I’ve got this piece of paper.” What else are they providing for me? And I think that safety and security aspect – again, in this kind of moment of truth where, God forbid, a crash or accident happens – is something that really resonates. And so we power crash detection for some of our customers, and I think the stats are something like we detect 23,000 crashes a month. But in addition to that, we’ve helped to initiate the dispatch of first responders to the tune of I think like 3,000 a month. And we’ve seen customer testimonials from our clients where the customer has said, this has saved my life. It notified emergency responders. They came out and provided emergency services. So I think that’s almost like invaluable. How do you prescribe an ROI on that? But I think it just lends itself to kind of building that relationship with an insurer and building that loyalty with the insurer as well.

30 seconds. Sanjiv, 30 seconds.

Great point. I think this is a great panel. I would just also add to that we also touched on that the best thing would be to avoid the incident. So whether it’s you can use the data to better coach the drivers to become better drivers, you can also give them alerts to avoid accidents. You can give them alerts, predictive alerts to say your car is about to break down. Go get the battery fixed, or go get the engine fixed. So while there is ton of value to simplify a customer’s life, there is a lot of value of data, telematic data to actually avoid the incident to start.

A great point.

Thanks to all our panelists. Thanks to all of you for sticking around. Appreciate it. Thank you.