The Internet of Things continues to offer insurers considerable opportunities in terms of maximising existing data to better understand customers and enter new markets previously considered unprofitable. With this in mind, Post, in association with Microsoft, held a roundtable to look at how data accrued through IoT might be used to inform underwriting and risk management decisions
The roundtable began with participants discussing where the insurance industry is currently with the Internet of Things.
Aleks Leonard Rawlings, product manager at UK General, said: “There’s an increased awareness of IoT and how it has evolved in the market. You’ve seen the push with some insurers already developing specific products that cater to that.”
One of the examples Rawlings used is cyber in both commercial and personal arenas.
He said: “As IoT proliferates everything, the cyber risk increases – whether that be from hacking or whether it be from breakdown.
“You’ve heard about the ‘home of the future’ and ‘car of the future’, but it’s now become ‘home of the near future’ because elements from that – your connected property or your semi-autonomous vehicle – are already here today.
“The push by some insurers to enter the market is a recognition that there is a risk.”
Meanwhile, Sandeep Varma, motor technology specialist at Direct Line, pointed out that not every area of insurance is the same.
“It probably helps to be specific about which part of insurance we’re talking about. IoT is all over the place as we know. For me, there’s probably only one area where it’s really developed from an insurer’s perspective and that’s in vehicles. Different rules apply in different areas obviously,” he said.
James Tucker, smart technologies manager at Allianz, agreed: “IoT does cover a number of things. I would say motor. I’d say healthcare to some extent as well, for example, the wearables.”
He also pointed out that there is some confusion as to what IoT is: “It’s a phrase that is bandied about by a lot of people, but when they ask what IoT is, they have no idea.”
Participants discussed various opportunities that can arise from IoT, including more sophisticated data.
Dave Preedy, insurance digital lead at Microsoft, pointed out that the initial use of IoT “has been on data that is known to not be of great quality”.
“We are seeing data science becoming really prominent in the insurance industry, fundamentally around underwriting, but for fraud detection as well and trying to eliminate the false positives. To make the customer experience smooth and only investigate the right people,” he said.
Meanwhile, Hélène Stanway, head of digital at Axa XL, believes IoT will help shift the intangible to the tangible.
She said: “The way that you manage and monitor physical assets will change because you don’t need somebody to go in on a specific day in February to maintain it. The longevity of assets will be greater because you will be alerted of a need to repair them, the instance of loss will hopefully decrease as a result of all this monitoring.”
She added that IoT data will help insurers have a better risk management insight and help them insure risks that are too expensive to insure or uninsurable today.
“Things that are uninsurable today – if you have sensors on them to collect data you can actually then craft an insurance product based on it. That’s starting to give a bit of a different opportunity to distressed risks out there,” Stanway said.
Rawlings responded: “There are massive implications for large commercial accounts due to risk monitoring.
“When the insurance industry starts playing in that area and includes policies that are tailored, then you’ll increase client retention. You increase loyalty.”
Meanwhile, Tucker explained IoT is very complex, and added: “We will start with a very simplified version of what we’re going to end up using.”
Citing pets insurance as an example, he highlighted how IoT could help the market evolve. “If you look at IoT with pets, it’s about activity and obesity – that’s one of the really simple things that you can judge. How much is a pet moving? We know that IoT can do that. We can understand that the data in its simpleness is a good starting point.”
“From there we can move into more sophisticated areas, for example predicting what will happen [to the animal in the future]. We’ve got to start somewhere and we all begin with the most simple version and then build,” he added.
Integrating into the ecosystem
Speaking of integrating IoT into the insurance ecosystem, Fredrik Motzfeldt, GB industry leader at Willis Towers Watson, said it is using a gradual sandbox approach.
He commented: “Certainly from Willis Towers Watson’s perspective, we are looking at core areas we could focus on. Some ring-fence sets of data.
“The word ring-fence is actually important because there are so many data ownership and regulatory issues you’ve got to get around.
“And once you have got permission from the clients to use the data you then need to get it into a format that IoT or artificial intelligence or analytics wants it but that can be overcome.”
“If you look at risk management, the use of IoT type devices can help with risk monitoring. If you look at national catastrophe modelling, they’re using drones to help with the claims adjustment in remote areas of Africa and Asia.”
He added that his firm is looking at the idea that clients can adapt sensors in, for example, an area affected by flood to get an early warning – although it’s in the early stages.
However, Pip Fox, strategic account manager at Microsoft, questioned whether insurance companies should be facilitating the use of IoT in the first place.
She said: “Are insurers IoT people? You don’t own the things that the IoTs been put into. Should it be that the people who are manufacturing the devices are regulated to provide data to anybody who wants it and put the security measures in place themselves?
“Then insurers could fight with everybody else to get that data and prove that they can do more with it and provide a better deal for the corporate or a person who provided the data in the first place.”
Graeme Howard, chief technology and information officer at Covéa, agreed that partnering with companies that can provide the data could be useful.
He said: “To me, it’s all about the data. I don’t want to buy the sensor or build the sensor. I just want the output from it. It’s working with people that really know their area of expertise, it’s that ecosystem.”
Tucker responded: “We’re not going to create any of this stuff ourselves and we’ll all have our individual understanding of what we think the right approach will be to create that advantage because we’re in a competitive market place. So we all think today we’ve got the magic sauce, pricing, underwriting and ours is better than someone else’s. We’re just trying to find a route to making more profit in what is actually a really difficult market to be in.
While insurers are seeing the possible opportunities IoT can bring, Motzfeldt believes people are too afraid of making mistakes, which is slowing down the adoption of IoT.
He said: “They are not given the chance to make mistakes, hence why we talk about sandboxes and ring-fencing and all those terms.
“There is a risk that certain parts of this industry might be shaken up because competitive insurers will come out of the blue and suddenly they’ve replaced the traditional carriers.”
Paul Middle, global partnering director of connected insurance at RSA, disagreed that IoT is driving a shift away from traditional insurers: “What IoT has done is give the opportunity for smaller organisations to create those propositions that we can then go and work with them on. We both need each other to succeed.”
He added: “To find ways so that we can react to new client needs, for example, the gig economy and propositions designed specifically for them as opposed to a traditional 12-month policy that just doesn’t fit anymore.
“We haven’t got the agility to create that proposition ourselves. So it’s about trying to identify where those partnerships are so that we can then get access to those new markets.”
Lack of communication
Another challenge is a lack of communication between insurers and their clients.
Stanway said: “There is a value add from an insurance perspective, but companies are implementing these devices and they are not telling their insurers.”
She added that often both insurer and the client will test the same devices, which is wasteful duplication. Stanway also said that the insurance industry is not ready to really understand the data.
“In today’s world we get a submission once a year. We just get data from a single point in time. I don’t think people are ready from an analysis point of view to really understand that constant flow of data – and what it actually means for risk and then for pricing. That’s the journey we’ve got to go on. And, for sure, we are not there yet,” she said.
Most participants agreed the problem with adapting IoT lays with companies not utilising the available talent.
Varma said: “Data science for the sake of data science is a bit empty as far as what it returns. What it really needs is to humanise [the data] and make those assessments, those impacts and those correlations bear some fruit whether it’s for a customer or for an insurer.
“There’s a dearth of people who can solve those kinds of problems today and perhaps investment needs to be done again sooner rather than later.”
Stanway agreed: “We’ve got thousands of really talented people already. So why are there not more programmes to enrich skills that we already have?”
She highlighted the availability of the apprentice levy: “Everybody thinks it’s for apprentices that are just starting their career. No, it’s for everybody, but people are not taking those opportunities up.”
Attracting the right people
Rawlings agreed it is not necessarily about making the “insurance industry attractive for new people”.
Howard responded: We’re pushing to have a couple of data scientists but the bulk will come from within this system, it’s people converting because if they understand the data, they understand underwriting. Employing staff to come and do data science from an underwriting data perspective.”
He added that it’s about creating a safe environment for people to look at things in a different way: “If we don’t look at it in a different way then we’ll fix the same problem that everyone’s trying to fix.
While integrating IoT has its opportunities and challenges, the participants also pointed out that to make IoT work, customers have to see a value in sharing their data.
However, “there’s a great difference” in values across commercial lines and personal lines customers, according to Middle.
He said: “On the personal lines side, I keep asking myself the question, what do customers really want from their insurance company? That’s irrelevant because actually I just want to buy the insurance and that’s all I’m interested in for that five minutes I’m thinking about insurance, versus on the commercial side, where it is different with larger risks.
“As a personal lines customer would I pay extra to have things like maintenance on my vehicle insurance? No, but when I buy my car from the garage and they offer a proposition where they’ll diagnose issues with the vehicle much better, I might be more interested. But from my insurance company – I’m not so sure.”
Howard responded: “How do you bring that in to give value back and how do you enhance that proposition to make it more than insurance? It needs to be more than insurance otherwise data is only about risk and it’ll only ever be about underwriting and risk.”
Middle added that the objective is to change the model from reactive to proactive.
“Ultimately, if you’re managing risk better you’d expect the insurance premium to go down so there is obviously less premium out there. We’d then have to do something else, charge for a service to make it up. And that service has to be something that they find valuable,” he concluded.