Insurance and the big data technology revolution

Automatic razors, jet packs and traffic jams on the skyways. The Jetsons made all sorts of predictions about life in the distant future. But the Hanna-Barbera cartoon, made in the 1960s and 1980s, skipped over the question of insurance.

That is hardly surprising. Risk management is rarely a popular topic in science fiction. And in any case, the insurance industry has not always been quick to pick up on new technology. “The entire industry is focused on how life was lived five or 10 years ago,” says Andrew Brem, chief digital officer at Aviva.

He is not the only one who feels that way. In a recent survey of insurers by broker Willis Towers Watson, 74 per cent of respondents felt the industry had failed to show leadership in digital innovation.

But change is coming as the industry wakes up to the idea that the cover we might buy 20 years from now will look very different from the cover we need today.

Big insurance companies are pouring millions into innovation and research. Mr Brem’s company has set up what it calls a “digital garage” in Hoxton Square in east London to work on new ideas. Axa is putting more than €3bn into IT and digital developments. And others, from Allianz and Munich Re in Europe to Manulife in Canada and XL Catlin in Bermuda, are also investing heavily.

Meanwhile, billions more are going into so-called “insurtech” start-ups which are aiming to challenge the big players. The likes of So-sure, Friendsurance, Lemonade, Guevara, Brolly and a host of others are aiming to disrupt insurance in the same way that Uber, Airbnb, Netflix and Spotify have caused upheaval in other industries. According to CB Insights, $1.7bn went into insurance start-ups last year, across 173 deals.

“Insurance will be enormously different in the future,” says Scott Walchek who runs Trov, an insurtech start-up that operates in Australia and the UK. “We’re at the beginning of technology being applied to risk management.”

How we buy

The first thing that could change over the next two decades is the way we buy insurance. Even with the development of price comparison sites and mobile apps, many people still find the experience time consuming and frustrating.

“Buying insurance is ridiculously retrograde, with endless questions resulting in a quote,” says Mr Brem. “We’ll be moving to a world where those questions are basically unnecessary. With the use of big data, we are discovering interesting and accurate predictors of risk that do not involve asking people questions.”

One example already in use is the way that Aviva prices its car insurance. Traditionally, this involved a lot of questions about the type of car, the location and the driving history. But the company has found a statistical link between the purchase of life insurance policies and safer driving. So life insurance policyholders get lower quotes.

This idea is likely to grow over the next couple of decades. “Not far into the future, the power of the data will mean that people who look the same [to an insurer] now will no longer look the same,” says Mr Brem.

Insurers are already experimenting with this. Last year, Admiral wanted to use the language used in Facebook posts to gauge how dangerously a person would drive. Facebook got cold feet over the experiment, but industry experts say that other insurers are likely to try to use data in similarly innovative ways to assess risk.

The result, say insurtech enthusiasts, will be a system that accurately prices risk without burdening the buyer with cumbersome, time-consuming questionnaires. In the Willis Towers Watson survey, 94 per cent of respondents said that distribution would be the area where technology would make the biggest difference over the next five years.

What we buy

What we choose to cover, and how we choose to cover it, is also likely to change radically. Today, we buy car insurance, home insurance and travel insurance that is much the same as it was a couple of decades ago, except that we can now buy online as well as via a broker.

But Mr Walchek believes the policies of the future will bear little resemblance to those of today.

“We still own things that we want to protect and we take risks with them,” he says. “What has changed is that life and its risk is being ever more precisely measured. In the past, for example, you might use a postcode to give you an indication of risk. Now it is being measured in feet and inches.”

Imagine that mobile phone signals or other sensors detect that a person is about to walk down a road where several people have recently fallen on ice, he says. The insurer will react by either sending a message warning the person to walk more carefully or else automatically increase the premium and cover while the policyholder is walking down that road.

More on Insurtech

Investors view insurance as staid and ripe for disruption

Insurance tech start-up investment surges

Cost of car insurance to plunge with rise of driverless vehicles

In the shorter term, more precision is likely to mean that insurance will be broken down into easier to digest chunks. The days of the annual policy may be numbered.

Trov is already offering coverage for specific items for limited periods of time. Customers can buy insurance covering a camera for a day, for example.

The same principle is being applied to car insurance, where a common complaint is that, thanks to annual policies, people are paying to insure cars that spend most of their lives sitting in car parks or driveways. Being able to switch policies on and off, the argument goes, will make them much more attractive to occasional users. Cuvva in the UK and Metromile in the US are among those offering pared-down policies.

But not everyone thinks this model will have universal appeal. “The consumer has to make the decision,” says Murray Raisbeck, an insurance technology specialist at KPMG. “But consumers don’t really want to think about their insurance.”

Andrew Rear, chief executive of Munich Re Digital Partners, agrees. “I can’t see people wanting to turn their insurance on and off. But one of the things that’s interesting about atomisation is that the customer doesn’t have to turn the insurance on and off. You can make a decision that you want insurance when you are doing X, and the machine decides when you are doing X [and so switches the cover on].”

Another possibility is that we will be insuring things that, currently, we do not insure at all.

“Over the long term what we’ll see is a shift in insurance premiums,” says Mr Rear. “You see risks going down in areas where sensor technology can make a difference and help to reduce risk and reduce premiums, but you also get new risks and new sectors that need insuring.”

Digital data are one example. “There are virtual assets such as images, or collections of items in games,” says Mr Brem. “The digital world is creating a lot of new types of assets.”

At the moment, cyber insurance tends to be a corporate policy, covering companies for the consequences of a data breach, for example. But, given the amount of personal data that is now held digitally, some insurers are starting to look at the question of whether there would be a market for personal cyber cover.

On the road

It is not all a question of buying more insurance. In some areas we may find that we need less. Car insurance is one of them. The question of how to insure driverless cars is the subject of much debate in the industry. If a car has an accident in driverless mode, who is responsible? The latest government proposals suggest that car owners will buy a single policy much as they do now, and the insurer will be able to claim from the manufacturer if it turns out that the car was at fault.

Over the long term, however, the nature of driving could change dramatically. At a recent conference on insuring autonomous vehicles, Lukas Neckermann, a consultant, predicted that: “Over the next five to 10 years, we will move to a point where we don’t need to own or drive a car and we’ll have mobility on demand. The classes and types of vehicles are going to change significantly to classes that the insurers are not familiar with yet.”

That, in turn, will change the nature of car insurance. “Personal motor insurance will decline but fleet insurance will increase,” says Mr Raisbeck. “The manufacturers of vehicles may be the providers of insurance.”

Prevention not cure

One of the big benefits of autonomous cars is expected to be a reduction in accidents. David Williams, technical director at Axa, says that 90 per cent of accidents are caused by driver error. Fewer accidents should mean fewer insurance claims and hence cheaper insurance. Companies using in-car telematics, such as Insure the Box and Carrot, already try to encourage their policyholders to drive more safely.

Many in the industry hope that a similar principle will start to apply elsewhere. They hope that by using technology, insurers can advise policyholders how to avoid incidents and hence cut the number of claims. The insurer, they hope, will change from being purely a source of claims payments to a trusted adviser.

Alerting the person about to walk down an icy road would be one example of this. But others are already available.

“This has been talked about for a while, but we’re at a tipping point, for example with connected homes,” says Mr Brem. “I have something called a leakbot at home which tells me if it thinks there may be a leak.”

Giving insurers access to vast swaths of personal data is not for everyone, and some companies may offer policies that require less information. But that could carry a financial cost — the less insurers know about the risks they have to cover, the more they will charge for providing that cover.

The progress of telematics policies in car insurance offers an example. Many younger drivers, who can save by allowing insurers to see their driving data, are happy to sign up to telematics policies. However, for older drivers the savings may be less significant because the insurance companies already know a lot about how likely they are to make a claim, and the market for telematics among older drivers has not really taken off.

Health is one area where the potential for monitoring technology is moving ahead fast. South African insurer Discovery, creator of the Vitality programme, uses devices such as Fitbits or other activity trackers to encourage its health insurance customers to exercise more. The more they exercise, the fitter they are and so the cheaper the insurance.

“There’s an imbalance — healthcare is over-consumed, and wellness is under-consumed,” says Adrian Gore, chief executive of Discovery. “We pioneered the idea of incentivised wellness a decade ago. Now it is part of health policy. There’s been a massive change. It’s about trying to understand risk factors better, trying to incentivise behavioural change and making the healthcare system more personalised.”

In the future, he says, insurers may be able to use genetic data to help tailor health plans for each person, although he warns that the same data should not be used as a way of pricing policies.

The far future

In the dim and distant future, Mr Brem believes we may not even have to think about insurance for individual items. “Our industry has created a false separation between different types of insurance. That is not how I think about my life. I’d like to have a single relationship that I can tailor. I’d like the insurance company to say ‘I’m going to evaluate you and here is our recommendation across the piece with a single premium you pay every month’.”

Mr Walchek agrees: “There will be a rich brew of data that identifies risk in real time. That changes the way insurance works. Maybe there’s a single insurance policy that goes with you at all times — call it a halo — and covers all kinds of risks. The distinction between risks diminishes.”

How very Jetsons.

Phone a friend

There is an alternative view of the future of insurance. Rather than using big data to price everybody’s risk individually, there is a school of thought that says insurance will return to its roots where small groups of people get together to pool their risks.

The modern name for this old concept is peer-to-peer insurance. One of the main benefits is expected to be a reduction of fraud — people are less likely to put in bogus claims if they think they are hurting their friends, or causes they care about.

US newcomer Lemonade is one of the highest profile promoters of the idea. It provides home insurance for the rental market, but customers are grouped together by selecting a charity. The premiums of each group go into a pot at the start of the year. Whatever is left at the end goes to the charity. Lemonade launched last September and says that it already has more than 2,200 policies.

Lots of others are working on similar models, including Guevara, So-sure, Friendsurance, Inspeer and Tongjubao. The models they are using vary. For example, the money left at the end does not always go to a charity. Some companies use it to reduce premiums for the following year.

One thing that unites most, however, is using the peer-to-peer concept to reduce the mistrust that has often existed between insurers and policyholders.

Mr Raisbeck says that there are limits to the model. “If someone can create a platform that allows lots of peer to peer niches to emerge then it can work. But individual peer to peer pools break down when they get over a couple of hundred people. You need a deep trust between the people in the pool.”