Introducing ISF: Insurability Sufficiency Framework for Autonomous Vehicles — Part 1

In March, we announced the release of the Singularity Platform, an insurance platform designed for autonomous vehicles, robotics, and other automation risks. The platform allows AV developers to securely share data to quantify autonomy risk for insurance applications. Insurance companies looking for an edge in underwriting autonomy now have a commercially available tool for that.

Today, we would like to shed some light on the methodology behind the platform, which we call the Insurability Sufficiency Framework™ (ISF). ISF is built on the first principles of insurance, data science, and systems engineering. It encompasses autonomous vehicle data sharing and risk scoring while accounting for the complexity of AV development and its operating environment. ISF contains many features to bolster underwriting more effectively and efficiently and in particular, serves two important goals:

  • It provides a practical way to understand and underwrite the risk associated with autonomy software, hardware, and operations; and
  • It provides a scoring model that can take advantage of alternative metrics and datasets to compensate for the lack of claims data to drive AV-specific insurance rates and programs.

In this series, we are looking to provide a high-level “what, why, and how” of ISF, which we hope will inspire insurance professionals to engage on this topic and explore what we do at Koop Technologies.

To begin with, we are an insurance technology company established to tackle the challenge of insuring autonomous vehicle fleets. Currently, there is no underwriting model capable of producing meaningful, reliable, and repeatable results to rate AV exposure (e.g. whether underwriting for primary Commercial Auto or excess General Liability/Products). Insurance companies have to rely on client anecdotes, public documents, and subjective judgment to rate tens of millions of dollars worth of coverage. We believe this is not enough to service an emerging industry estimated to generate trillions of dollars of value in the coming decades. Already today, there are AV developers running fully autonomous operations in multiple operating domains and responsibly pushing for the commercialization of AV taxi, freight, and delivery services. In the next few years, this is only going to accelerate. Nobody wants to have insurance become a hurdle for AV adoption with patchwork underwriting and a lack of fit-for-purpose coverages. This is why we created ISF.

Our thesis is that autonomy requires a practical methodology to estimate risk in order to underwrite it. A capable commercial insurance product should embody such a methodology that will let AV developer-operators continue to grow by making insurance available to them with as little disruption and as cost-effective as possible.

So what exactly is ISF? ISF is a risk assessment and insurance rating methodology for autonomous vehicles. On one end, our platform collects autonomous vehicle data through Koop API as well as other means. On the other end, we have a set of ISF Scores that explains the insurability of AVs. Eventually, a given ISF Score will incorporate a rating basis for insurance coverage. The scoring is standardized across certain factors so that insurance carriers can reliably use it from one insured to another. Meanwhile, AV developers and fleets can be assured of the security of the data sharing process and end-use of the selected data fields. ISF also solves the chicken-and-egg problem that has been around for a while in the industry: AV companies don’t want to share data unless there’s a practical use for it, and insurers are uncertain about what to ask for because of their unfamiliarity with this new technology.

How do we know that ISF is technically relevant and capable of providing high-quality results that could be used in an insurance transaction? When building out the parameters for ISF, we deliberately took an unbiased and safety standard agnostic approach, which resulted in a set of the first insurance and technical principles that laid the foundation for the methodology. Some of these principles are:

  1. Crashes will not go away completely with AVs. Autonomy is expected to deliver a magnitude of safety improvements across the board, but there will be no perfectly fault-proof system. A combination of unfavorable internal or external factors could push AVs into different kinds of accidents. Also, until AVs reach an extreme scale to account for the vast majority of vehicles on the road, human drivers might run into issues when sharing the roads with AVs. To put it simply, things that we thought were good to go could fail, and risks that we haven’t even thought of could come up as autonomous fleets start to scale.
  2. It is critical to sufficiently understand AVs for underwriting purposes. Autonomous vehicles are incredibly complex, so if we are looking to underwrite an AV risk, it is no surprise that underwriters must either have sufficient expertise in the system or have tools to help them do that. One topic that keeps getting debated in the insurance industry is how sufficient is good enough? Do underwriters need to look at just a handful of operating metrics (and if so, what are those) or do we need to thoroughly inspect and evaluate the whole AV stack, from base vehicle platform to perception, prediction, motion planning, and actuation? ISF provides an answer to this question, equipping the underwriting process with sufficient operating metrics and qualitative surveys that can explain the insurability of a given AV system and fleet deployment.
  3. Collecting AV data comes at a cost. To sufficiently understand AVs, underwriters need the insured’s experience. An underwriter can look into an individual insured’s experience or at the population of similar, homogeneous exposures. To do that, underwriters have to ask for data. The challenge stands that data sharing is complicated and usually comes at a cost for AV insureds. For example, even collecting simple data such as miles driven and the distribution of autonomous and manual modes could be burdensome. There are situations where some parties might require to perform a rigorous evaluation of the whole stack, which from the data perspective, is a prohibitively costly exercise for an AV company when procuring insurance. A key value proposition of ISF is the ability to formulate a sufficient data sharing request that comes with the tools to securely provide that data.
  4. AVs have to satisfy that they are an acceptable risk for an insurer. This means that insurers must sort out the following:
  • Calculate and measure the probability and cost of loss associated with AVs;
  • Exposure has to be understood across various AV fleet deployments; and
  • Insurance premium for any given coverage has to be affordable to AV insureds.

We believe there is a strong positive correlation between determining the risk characteristics of a given AV operation and the data and analysis that underwriters can use to understand that operation. This not only gives existing insurers a better ability to underwrite AVs but also provides a foundation for new, dedicated autonomy insurance programs.

In summary, ISF is an underwriting tool for autonomous vehicles, built on the first principles, and is available through the Singularity Platform. In the upcoming articles, we will touch upon a few technical topics pertaining to the methodology. If you are an AV developer, risk manager, insurance carrier, or broker looking to learn more about ISF or Koop Technologies, we would love you to reach out to us! Also, please make sure to follow our progress on LinkedIn and subscribe to our product marketing list.

In the news: Koop Technologies Announces Singularity Platform — Insurance Platform for Autonomous Vehicle Data Sharing, Underwriting, and Claims Handling.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store