Most of us will be aware of the ongoing transportation and automotive revolution that within a decade will consign driving to the past and instead usher in a world where cars simply provide a service. The impact of such fundamental change will be widespread and among the sectors most affected will be insurance.
In fact, machine learning is already making itself felt in the insurance sector, leading to an informal classification between legacy or traditional insurers and the so-called insurtech players exploring the possibilities provided by data-intensive processes.
Car insurance is big business for many insurance companies: the law requires all vehicles to be covered, and what’s more, in most countries drivers must have a policy for each vehicle they own, regardless of whether it is being used or, as with most cars, they are parked up 95% of the time.
In response, Metromile suggests pricing policies on how often a vehicle is being used. The company has already attracted around $300 million from investors to design systems that not only charge on how many miles you drive, but also to automate accident reports based on the vehicle’s sensors. Other companies, such as Zubie, plug a device into a car’s OBD port to monitor use, the type of driving, or even to warn when a part needs replacing, basing the price of a policy on the driver’s score. Arity, a spinoff of the U.S. giant Allstate, has come up with a platform aimed at linking certain types of driving with accidents.
Kasko2go proposes keeping its prices low by using a smartphone app to create a portfolio of low-risk users: 90% of drivers have never had an accident and 95% obey traffic signals. Trying to discover fraudulent claims is the leverage the company intends to offer competitive prices: users keep the company’s app active while behind the wheel, building up a picture of how they drive, and can also be used to establish the circumstances of an accident and thus expedite payment in the case of a genuine claim.
Root, a company I’ve already talked about, offers a policy specifically for Tesla cars in which the owner pays less the longer the vehicle is used in self-driving mode, reflecting the data showing autonomous driving is inherently safer than having a human behind the wheel, and has come up with apps that use a smartphone’s sensors to determine how somebody drives, rather than basing the cost of a policy on where they live, how old they are or other socio-demographic factors used by algorithms and that have been shown to be discriminatory.
Are insurance companies taking a proactive enough approach to using technology now that so many of us carry a powerful computer around with us and vehicles can easily and cheaply be fitted with sensors? Old habits die hard and many companies will resist basing policy costs on car use or how we really drive; perhaps they’re right and such an approach is not sustainable.
Is the insurance industry living on borrowed time or will there be a race to generate spinoffs or acquisitions that will allow it to save face despite its technological backwardness? Aside from the changes the automobile insurance industry will have to make to deal with the transition to autonomous vehicles, many of them in fleets, the use of technology could help lower the cost of policies. The question is whether the traditional players, the so-called legacy insurers, are simply biding their time.
Source: Forbes
EmoticonEmoticon