View All

Introducing ISF: Insurability Sufficiency Framework for Autonomous Vehicles — Part 3

In ISF Part 2, we discussed the data sharing, risk scoring, and insurance rating mechanisms behind the Insurability Sufficiency Framework. Our mission is to underwrite autonomous vehicle fleets in a meaningful, reliable, and repeatable way, and ISF is the centerpiece of that. This article describes ISF’s impact on a new generation of insurance products that we are looking to launch for the world of automation.

Insurance Products for AVs

ISF is a technical methodology for estimating the cost of risk of autonomous vehicle fleets. We developed a set of risk scores based on fleet data, performance metrics, and safety features that can explain the risk profile of a given autonomous vehicle deployment, be it ride-hailing, trucking, delivery, or off-road use case. ISF is integrated into the Singularity Platform that is available for use to autonomous vehicle customers today. It is the only purpose-built and production-grade methodology for autonomous vehicle insurance on the market.

However, the methodology itself is not the end game for us. In the world of insurance, we work with products (in this case by way of policies and the coverages they provide) that fleets can purchase to protect themselves against different kinds of liability risks. Shipping new insurance products to market is not easy. We must align capital and paper, estimate loss performance, develop an underwriting approach, have a distribution plan, work with regulators, and so on. All of these things have to fall in the right place. Thankfully, our insurance software platform enables us to do all of that and live up to the expectations (and requirements) of our autonomous vehicle customers.

Source: Koop Technologies

To begin with, why all the attention around insuring autonomous vehicles? Since insurance is an effective method of transferring risk from a first party to a third party, the certainty that comes with risk transfer is invaluable to millions of commercial enterprises around the world, from startups to corporations. Autonomous vehicles are no exception.

A lot goes into making autonomous driving happen: base vehicle platform, sensor hardware (e.g. cameras, LiDARs, radars), HD maps, localization, perception, prediction, actuation, simulation, and the list goes on. Oftentimes, AV companies do not develop everything in-house but rather have to rely on Tier 1 suppliers and software vendors to put together what is known as the “AV stack”. A lot can go wrong within the AV stack, which is very different from the human driver risk, where behavioral factors such as distracted or impaired driving can lead to accidents. With AVs, the risk is much more technical since the root causes of accidents might have to do with failures at certain nodes of the AV stack.

From the perspective of an AV company, we are looking at the risk of the autonomous system. Understanding the risk allows us to work with a risk partner and utilize insurance to transfer some of it. An insurance product that enables such risk transfer at the autonomous system level is known as General & Product Liability.

Since AVs are expected to commercialize through the fleet model, where an AV developer licenses the technology to a fleet that owns and operates autonomous vehicles, it introduces a new party of interest in the value chain. The AV fleet is responsible for operating and maintaining AVs within the operating domain defined by the technical capabilities of each software driver. For example, companies such as Aurora and Embark Trucks publicly stated that they would pursue the fleet model with their fleet partners. If you are an AV fleet deploying vehicles on public roads, you are generally subject to financial responsibility requirements imposed by each state in the U.S. These requirements are most often met by the purchase of an appropriate Auto Liability insurance policy. Coverages provided by these policies are the financial “first responders” for auto-related accidents and allow fleets to transfer liability economically.

In addition to coverages offered by Auto Liability or Product Liability, there are other needs in the AV ecosystem that insurance can address. Fleets must consider costs associated with business interruption, property damage (at transfer hubs, for example), cargo liability (for shippers), intellectual property (for developers), and more. Today, these needs translate into an array of insurance products such as Technology Errors and Omissions, Cyber, Inland Marine, Property, and others.

As you might have noticed, there are many insurable interests within the world of AVs, with different insurance products available to best match their needs. However, the combination of today’s insurance products does not account for the fact that all of these interests are going to be affected by the change in the underlying nature of risk as we move to AV-based services, as opposed to human-based operations. That opens up a tremendous opportunity for companies like Koop Technologies to develop a suite of insurance products fit for insurable interests across the AV value chain. Our risk methodology, ISF, helps us build insurance products where the underlying risk has to do with software-defined deployments in real-world operating domains.

We are uniquely positioned to launch new insurance products as an MGA by taking advantage of our methodology and AV fleet integrations. The ability to have insights into each AV system and use those insights to drive insurance coverage terms and rates effectively is the future we are building for the autonomous vehicle world.

Source: Koop Technologies

Value Prop of Autonomy Insurance

A key value proposition for buying insurance is not only financial protection but also assistance from an insurance provider when a covered loss arises. In the case of autonomous vehicles, if there is an accident, the expectation for claim assistance would be the same. Understanding a claims process from an AV perspective is therefore critical.

For AVs, when a claim happens, there are unique factors that differentiate it from human-driven scenarios. For human drivers, if there is a crash, an insurance company would investigate to figure out who is at fault — it could be a first party (you were driving recklessly and hit someone), a third party (someone was driving recklessly and hit you), or no parties at fault (a tree falls on your car out of nowhere). This process is known as claim adjudication. For AVs, it is more complicated. If there were a problem with the AV system (meaning an AV was at fault), an insurance company would have to figure out if it was due to failure within the stack, vendor/supplier issues, or a problem with maintaining the fleet. Even if the AV system was not at fault, you can imagine scenarios where third parties would still blame an AV for performing or not performing a certain maneuver that could have avoided a collision. In simpler terms, assigning fault could get tricky, lengthy, and plainly complicated. Considering there are many interests at stake, even if an insurance company pays out a claim, it might decide to go upstream and subrogate other insured interests (the process known as subrogation). That in itself could end up being a major source of costs for insurance companies.

Additionally, it is important to consider the severity of claims. For traditional auto, the severity curves are pretty well developed for both liability and property. But for AVs, the wildcard is with the property since the hardware that goes into an autonomous vehicle is higher compared to its non-autonomous equivalent vehicle today. Understanding the complexity and cost of hardware repairs will be critical for building an AV severity curve eventually.

The key to handling AV claims is in data (just as it is for AV underwriting and rating). When we built our insurance platform, we knew data would be central to all insurance functions. That is why we developed a set of features that enables us to analyze AV data for claim adjudication purposes. As we launch our insurance products, we believe this in-house claim functionality will prove invaluable to our autonomous vehicle customers and will provide them with an unmatched end-to-end experience.

Source: Koop Technologies

In summary, autonomous vehicles are different from human drivers from data sharing to rating and underwriting to claims handling. AVs introduce a new risk that deserves its own insurance products and tools. Given the size and impact of this industry, it is a no-brainer.

Considering that there are more differences between AVs and humans than similarities, we believe it is appropriate to introduce a new category of insurance that will serve the autonomous model of transportation. We call that category Autonomy Insurance, which our company is pioneering with its technology and upcoming insurance products for a variety of interests in the AV ecosystem. At the fundamental level, Autonomy Insurance is grounded in understanding AVs and estimating the cost of risk for all kinds of deployments and use cases through data. With that, we are able to build new coverages and claims handling solutions that will define that new category of insurance.

At Koop, we have defined Autonomy Insurance and already brought it into existence. Going forward, we are excited to introduce new insurance products, platform features, and customers that will show you our bold vision not only for the future of autonomous vehicle insurance but also for the future of insurance itself.

In the meantime, please make sure to follow our progress on LinkedIn and sign up for company updates here. If you would like to learn more about Koop Technologies, feel free to check out our website or reach out to us directly at hello@koop.ai!

Previously in the ISF: “Introducing ISF: Insurability Sufficiency Framework for Autonomous Vehicles” Part 2.