Top 5 Insurtech Trends for 2020 (and beyond)03 December 2019 | Smart Saving Water
2019 is drawing to a close, and what a year it’s been for Insurtech. For the uninitiated, Insurtech is a catch-all term for a new breed of companies marrying traditional insurance to emerging technologies like artificial intelligence, the internet of things and blockchain, often with a radical new twist on the original product. Considering insurance has been around since 1347, a little update here and there certainly doesn’t go amiss – especially in a world that’s increasingly digital, on-demand and driven by rapid innovation.
And it really has been a great year! Lemonade, Root and Hippo all smashed the billion-dollar-valuation mark and joined the unicorn club; Insurtechs collectively raised a record-breaking 4.4 billion dollars and even Microsoft wanted in on the action. As we close out 2019, we thought we’d take a moment to reflect on the trends that defined the past year and look ahead to what we can expect from 2020 and beyond. Here are our top five predictions for how Insurtech will continue to switch up the game in the year to come. Let us know if you agree!
Let’s get to it with prediction number 1:
AI is Coming, But Maybe Not As You Know It.
There’s a lot of hype about artificial intelligence in insurance. One of the most common applications that gets bandied around is applying machine learning to the vast books of data that insurers accrue, in order to develop new insights into risk and improve on existing price modelling methodologies. That’s okay, and we’re sure it will happen – in fact, companies like Lemonade already report massive improvements in acquisition, loss ratio and customers-per-staff from doing more or less exactly this. But it doesn’t get me excited as a customer.
Arguably the best AI available to end-users right now is Amazon’s Alexa. She’s genuinely smart, helpful and context-aware. You can talk to her, she talks back, and most of the time says something sensible that more or less answers the question. Under the hood, she’s rummaging through colossal reams of data – transcribing your voice to text, analysing the resulting words, mapping the semantics, interpreting your intent, blending in heuristics like location and search history, issuing a canonical query based on the dominant probability tree… and somehow returning with the right answer from billions upon billions of contrasting possibilities. Almost as complicated as say, trying to understand your insurance policy wording.
That’s somewhere we see natural-language-processing AI like Alexa delivering real benefits. We’d love to be able to ask: “Alexa, does my travel insurance cover me for skiing off-piste?”. Not only would it help us to unpick our labyrinthine policy documents without suffering an instant migraine, but it could also help to protect insurers from the wrath of the FCA – for whom the idea of a customer not understanding what they’ve bought is a serious red flag. It even provides the insurer with a chance to say, “actually, you’re not – but would you like to buy winter sports cover now for an extra £6 a year?”. Everyone’s a winner. There is a valid reason why insurance policies are written in Legalese – they do need to have a precise meaning that’s not always easy to convey in Human. But conversational AI can help to bridge that gap and provide a more natural interface between policy and person. So, our first prediction for 2020 is that’ll we’ll see more natural-language bots in sales and support functions as incumbent insurers seek to automate regular processes and chase early trailblazers like Lemonade. The big question is – will they be better or worse than human representatives? Time will tell.
Prediction number two:
This Might Be The Year Blockchain Insurance Gets Interesting.
Blockchain can be a pretty meaty subject. For this article, let’s just say it’s a way of connecting people directly to one another (peer to peer), without needing a central entity to oversee things.
So say I sell something through eBay – I’m selling it to another person, “peer to peer” – but I still need eBay to sit in between us as neutral third party; hosting the marketplace, administrating the sales process, resolving disputes and connecting buyers to sellers. In “Blockchain Bay”, I would be able to connect directly to my buyer without needing eBay to act as an intermediary – peer to peer, with all the rules and processes we need to transact the sale represented by a software called “smart contracts” instead of company policies.
Why do we care? Well, the idea of an inherently peer-to-peer network gets pretty interesting when you apply it to insurance.
In traditional insurance, I pay my insurer for a policy; they put all the policy fees into a pot, and if something bad happens then they dip in and (hopefully) pay out. Whatever’s left at the end of the year – after all the pay-outs and expenses and operating costs – becomes their profit to keep.
The problem with this approach is that right from the get-go we have a conflict of interests: pay-outs and profits come from the same pot. In other words, insurers have an incentive not to pay claims, because they come directly off their bottom line – and that can lead to serious trust issues. Depending on who you ask, only about one in two people currently believes that insurers process claims fairly. Among younger consumers, it drops to one in three.
A peer-to-peer model changes the whole incentive structure to eliminate that conflict of interests. People still pay into a pot and receive pay-outs in turn if disaster befalls them. But instead of the balance converting to profit, it’s either returned to the pool members at the end of the year or even just donated to charity. The party who writes the smart contracts (the insurer, per se) might take a flat fee per person, but many of them operate on an entirely non-profit basis. Essentially, it’s a framework that allows people to club together and share risks between them, without needing to pay a third party to match them with other people – or needing to trust that third party to treat the “pot of gold” fairly and not just keep it for themselves.
Proponents also argue that it reduces fraud, because fraudulent claims mean stealing directly from another human being (and yourself – or a charity) – rather than nameless, faceless corporations that you don’t necessarily like or trust in the first place. There are even companies that allow you to create a pool from your family and friends to double-down on that honesty and transparency mechanic.
Fun fact, peer-to-peer insurance is kind of where we came from several hundred years ago. In very early examples, small farming communities would strike up simple contracts to share risks among themselves and hedge against disaster – if your crops fail, I’ll give you some of my eggs and milk; if a fox gets my chickens, you’ll give me some grain so I don’t starve. It was also common for tradespeople like stonecutters to each put a little money into a community pot so that if one of them died on the job the guild would see to it that their families could still live. Although some skimming would certainly have taken place, on paper it was peer-to-peer, not-for-profit and quite egalitarian.
Will blockchain-based insurance cut costs and bureaucracy by aligning everyone’s interests, or are we decades away from being able to administer complex financial products with software smart contracts? Nobody knows, but 2020 is looking like the year we really start to test it out.
Prediction number three:
Prevention is Better Than Pay-Out (And Not Just For Insurers)
Another interesting way to address the conflict of interests is to eliminate pay-outs altogether.
How do you do that? Well, by eliminating the causes of pay-outs.
Take Insurtech Neos for example – which as it happens, is one of Hero Labs CEO Krystian Zajac’s previous ventures. Neos does insure customers in the traditional sense, but it also equips them with a range of smart home devices aimed at actively preventing the most common causes of damage to property in the UK – burglaries, fires, and escape of water (AKA, leaky plumbing and appliances).
By actively protecting customers’ homes and belongings from harm, Neos hopes to relegate pay-outs from the central feature of an insurance policy to a “last resort”, which really is better for insurer and policyholder alike. You buy insurance because you like your stuff and you want to protect it – so actually protecting it definitely beats quibbling with a call centre over compensation while knee-deep in water or sweeping up broken glass. The better the insurer gets at protecting your stuff, the happier you are; and the better their bottom line, because of claims costs and administration fall accordingly (and customer retention probably rises as well). There’s also the things that a pay-out simply can’t replace – photographs, data on hard-drives, your time; the feeling that your home is a safe place. Some things can only be protected, not bought.
As someone once said: “traditional insurance markets itself as a raincoat, but really it’s just a towel”. Those things can only be protected, not bought.
So, one of our top predictions for 2020 and beyond is that we see more and more insurance companies adopting products that help to mitigate the most common causes of claims. Interestingly, the biggest drivers aren’t the obvious culprits like fires, explosions or storms – in the UK at least, water leaks absolutely dominate the statistics and account for more damage than all fires and burglaries combined.
It almost sounds like someone should create some sort of intelligent, whole-home leak protection system to solve the problem… Oh wait, we did!
Shameless self-promotion aside, this kind of “active insurance” reconciles everyone’s interests so nicely that we wouldn’t be at all surprised to see more of it in future – especially considering the dominant smart-home headwinds driven by tech giants like Amazon and Google. Watch this space!
Prediction number four:
Black Boxes for Buildings
Insurance is all about understanding risk. Historically, that meant compiling a big old book of all your historical claim data – your “actuarial database” – and using it as a lens to evaluate the “riskiness” of prospective new customers. You’d look at factors like historical crime stats for a given postcode, or plot the chance of a car accident against the number of years of driving experience. It’s not a terrible way to study risk, but it has some obvious drawbacks – these kinds of broad-brush “proxy indicators” will only tell you so much.
Take car insurance, for example. I’m 19 years old, newly qualified, I drive a Vauxhall Corsa with an aftermarket rear spoiler and my postcode is in a council estate in central London. An actuarial database might tell you that historically, I’m not a great risk. But in reality, I’m an excellent driver – my dad’s an instructor, he took me out every evening and I grew up knowing the critical importance of safety and etiquette behind the wheel. The real-world data bucks the actuarial trend in a big way. That’s bad for customers, because I effectively get penalised for the behaviour of others. It’s also bad for insurers, because if my understanding of risk is decoupled from the real world then I’m at risk of mispricing policies or limiting my competitive footprint.
There’s also the optics of it all. It’s not nice to be treated as a stereotype – least of all when that stereotype directly affects your rights, like how much you pay for products, or whether you can buy them at all. Imagine it like a shop – there’s a line of people queuing up to buy the same chocolate bar. The first guy gets it for 50p because he’s a middle-aged doctor from a nice neighbourhood. The next guy has to pay £2.50 for the same item because he’s 19 and unemployed. Another isn’t allowed to buy chocolate at all, because somebody from his demographic once stole something from the shop. Is that okay?
Fun fact number two: back in 2012, the UK passed a law prohibiting car insurers from using gender as an underwriting question – i.e., it became illegal to say, “men are better drivers than women, so we’ll offer them a better rate”. In reality, the data shows women are dramatically less accident-prone – about three times less likely to suffer a crash, at the time the law was passed. So here in the UK, we saw a sudden explosion in car insurance companies with pink, feminine branding and names like “Sheila’s Wheels” – aiming to attract mostly women drivers, while still complying with the new law and remaining blind to gender. Nothing stopped men from buying it – nothing but sparkly, butterfly livery and names like “Go Girl” – but needless to say, they ended up recruiting a pretty girl-heavy demographic. In a much less fun example, studies as recent as 2018 have uncovered car insurers using the applicant’s race as a proxy indicator – “you come from X ethnicity, so you’re a bad risk” – which is genuinely pretty shocking.
So, what can you do to make it fairer? Well, the biggest ever shake-up in car insurance came from the introduction of vehicular telematics, AKA “black box insurance”. For the uninitiated, a “black box” lives in your car and transmits real-world driving data to your insurer over the cellular network. That means you can pay a price based on how well you actually drive – a fair, individual price, not a broad-brush stereotype based on your age or your choice of car. It revolutionised actuarial methodologies (particularly for young drivers) and increased the data available to insurers by orders of magnitude, but perhaps most interestingly, it actually changed the way people drive as well. Drivers with telematics boxes drive more safely than drivers without. They’re incentivised in cold, hard cash to be better drivers.
We predict that in 2020 and years to come, we’ll see “telematics for property” start to really take hold – using data from smart home devices to help improve risk modelling in home insurance. We could see models where insurers reward customers for locking doors and windows, or offer better rates when customers are at home on the basis they’re less likely to be burgled, or they could quickly respond to a leak or fire. It’s not a drastically different concept to say, a health insurer giving you a Fitbit or a discounted gym membership – empowering you to improve your own risk profile because that’s what everyone wants. It’s another reason why our own Sonic is such a game-changer, although we can’t talk about details just yet due to a host of ongoing patents. But suffice it to say, 2020 should be an interesting year for Hero Labs too!
Moving on, our fifth and final prediction for the coming year is….
We Demand More On-Demand!
You’re driving at night, and you’re getting way too tired. Your eyes are drooping, your concentration is fading fast – it’s a very dangerous situation. You either need to pull over and rest, or ideally, swap with your passenger. But if they’re not on your policy, what do you do? That’s the idea behind on-demand insurance services like Cuvva – you launch the app, input a couple of details and in minutes your companion is covered and sitting behind the wheel while you doze off in the passenger seat.
Another take on insurance-on-demand is Trov, an app-first insurer that lets you turn single-item cover on and off with a swipe of your smartphone. In practice, that means you can do things like temporarily insure a camera that you borrowed from a friend, or only pay for bike insurance while you’re actually out and about. It’s a cool idea, and one that only grows in value when you combine it with heuristics from IoT devices, wearables and smartphones.
Our final predictions for 2020 and beyond is that we’ll start to see more and more of this dynamic handing-off of risk in the mainstream. We also predict it’ll become increasingly automatic and governed by data from IoT devices.
Travel cover that kicks in automatically when your phone is in another country? Yes, please. What about home insurance that links to AirBnb’s API and tags on landlord cover whenever you get a booking? We’ll take it. Why not let my drone broker its own insurance at the best market rate every time it takes off, and cancel it ten seconds after landing? Okay, that might take some work, but you can see the new and interesting possibilities that the on-demand revolution starts to open up. We live in a world where we can summon food, transport or the faces of our loved ones from the internet with a couple of taps and swipes. Why shouldn’t insurance be the same?
Well – that concludes our round-up of the top Insurtech predictions for 2020, but if you’d like to find out more about how IoT is changing the world we encourage you to fill out the form below and join our mailing list for regular insights and reports like this one. Alternatively, if you’re an insurance professional, head on over to our partnership pages to find out more about our world-class leak prevention system and apply for a trial.