This podcast features speakers from three very different companies all providing solutions for managing claims, but each at different stages in their growth. Recorded live from our Innovation in Claims Event in London on 5th March, with an introduction from Instech London Partners Matthew Grant and Robin Merttens.
Paul Stanley CEO of 360Globalnet (3.40)
Jeff Manricks and Thomas Verduzco-Weisel from Shift Technology joined us the day after picking up another $60 million in funding. (15:10)
Avi Baruch from Previsico, a spin out from Loughborough University, on forecasting flood and use of new techniques to help reduce claims. (25:10)
Listen here to InsTech London podcast 19. It is also available on iTunes, Spotify and Podbean.
……………………………………………………………………..
Transcript for this podcast
00:01 Matthew Grant: Hello and welcome to the InsTech London Podcast, this is Matthew Grant, one of the partners at InsTech London, and for those of you that don’t know us, InsTech London is the largest community in London bringing together people with an interest in innovation in insurance, and we have regular meetings of people from insurers, startups, technology companies, investors, professional services, and all those in between. This week, we ran a claims event and you’ll shortly be hearing from some of the speakers we had that evening but before that, a word from Robin and myself.
00:40 MG: So Robin, we’re about to hear the highlights of our claims event this week. Another great attendance. What were the highlights of the event for you?
00:49 Robin Merttens: Well, I think it’s a feature of our times that you can get more than 200 people to a claims event on a Tuesday night. The big things from my point of view, it was great to have Shift there, having raised 60 million Euros and announced it earlier that day, particularly as they have been friends of ours now since late 2017. Paul Stanley, I think the points he makes about how difficult it is to disrupt claims was a point well made on a day when we were listening to so many people who are trying so hard. How about you, what do you feel?
01:21 MG: Yes, I think the surprising thing about claims is, in one sense, this is an area where the market can actually save a lot of money, if it gets it right. A dollar saved in claim, or a pound saved in claims, goes straight to the bottom line. We hear a lot about claims in motor or in personal life, but we’re still not hearing very much about the commercial specialty space. So we’re starting to see a bit more of that, but I particularly like Previsico where they have the ability to monitor flood risk. A lot of people look at river flood or coastal storms flash flooding, but far few of them look at the kind of floods you get off-plain and what they’re able to do is to provide alerts to people about a flood that is coming and to help them take some actions to get their possessions out of the way of the flood. So I think they have got some real potential in there.
02:05 RM: I think if we had this thing in one year’s time, we would have a whole bunch of people doing parametric claims payments and when you hear Shift talking about how they’ve got completely automated claims processing these days, one does get a sense of how that part of the business is going to be fundamentally changed within a year or two. But then I’ve been saying that kind of stuff for years now, no one has taken notice, but I really do believe that.
02:28 MG: Yes, you and I tend to be ahead of the trend but at some point the trend will catch up, but on the parametrics, there is definitely a strong theme there for an event later in the year. A lot of applications are coming out from parametric. So you and I are putting together the event for the 2nd of April for MGAs, how are we getting on with speakers for that?
02:46 RM: Yes so we’re going back to the front-end and distribution. There are a lot of good new MGAs doing some really interesting stuff so we’ll be looking at new products. But most particularly this is about stories, how people brought their MGAs to market, who they used, who built the technology, how they got the capacity. It’s mostly about how MGAs are being taken from idea to trading and who is doing those things in a new and exciting way.
03:17 MG: Yes. What’s tricky about MGAs is it’s an opportunity for organisations that have traditionally been outside of the insurance space to actually use their clients and their distribution networks to get into the insurance space, so we are starting to see more in that space. I think we’re going to see a lot more happening in the next year. Good, well let’s transition now to the highlights of the event from last week and hear from our speakers.
03:44 Robin: Let me introduce you to Paul Stanley, he’s the CEO and founder of 360 Globalnet who are very much a leader in the claims technology space. He’s a serial entrepreneur though having been in insurance all his life. This is not a pitch. I’m going to ask Paul a series of questions around what it’s like to be in the InsurTech claims space, trying to digitise a world that is difficult to digitise, just to get a perspective from Paul on his journey, both as an individual and with 360. So Paul, starting off, how on earth did you end up as a claims innovator, it’s not something people choose to do, surely.
04:29 Paul Stanley: No, no. I’m a fairly ordinary person, I described myself as a simpleton in a post the other day, but I always say in insurance you only have to be slightly better, slightly smarter than the average bear, to succeed. Plenty of average bears out there, slow, average bears. So, how I got into doing what I’m doing now, was thanks to Royal Insurance who gave me every job under the sun in claims and some beyond, got me working with BCG, McKinsey, and the inevitable operational improvement programs they run when they were about to go bust. Then to Direct Line, fantastic experience at Direct Line with Peter Wood, and they let me do more or less what I liked, which was ultimately my downfall. Then when they got too bureaucratic, got 1.4 million, I started the first of five businesses, but I tell you this one is by far the most difficult. It’s got the biggest potential, but it has just felt like pushing a huge boulder up the hill, a $35 million boulder as it turns out.
05:48 Robin: Thank you. So tell the people what 360 does and why you think it’s been successful in this space that’s so difficult to break through in.
05:56 PS: Yes, it took us a couple of years to actually work out what we were going to do. We always had the ambition, well after two years – 24 months – to completely transform the claims experience. I worked in claims a long, long while and I still can’t believe how naffy it is for customers.
06:21 PS: And I’ve had the claim from hell actually as well, so I could tell you about that…What we do is a complete digital process for any claim of any complexity and any geography, we’re worldwide. The big thing that we do is hand the power back to business people from IT. So we give you a platform with all the digital tools in it, the ability to use video, imagery, self-service for everybody, capacity management of suppliers, manage all your outsources, just link everybody together around one single digital record.
07:00 PS: And it’s entirely self-configurable at the desktop by business users, which is massive. So we say to customers, “we hand the power back to business, so you don’t need IT anymore”. So the only thing we need IT departments for is to the plug the client into the internet and we do everything else. So it’s been a labour of love, we’ve got some fantastic people and it’s just been incredibly hard, lots of money, incredible amounts of patience required, but we’re just starting to see the market turn significantly over the last 12 months.
07:42 Robin: So why is innovation in claims more difficult? All the money seems to me to be going into product design, distribution channels, data improvements. Why, when it’s such important part of the process, why lessen claims and why have claims therefore become so difficult?
08:03 PS: I think my mantra is you’ve really got to understand claims in order to change it. There are so many things. When I first walked into a claims department when I was 18, and I saw people there dealing with a glass breakage to a multi-million pound fire, whatever it was. How on earth can these people do that? They must have brains the size of planets. And you realise that actually they’re just going by rote. That there is not a lot of original thought there. There are lots of people just content to do a job and go home. Fair enough, but I don’t know, I guess most of you’d be like me, I’m not really interested in doing anything unless I can be the best at it and the best I can be. What the hell are we doing here, if we’re not trying to do that? So I think it’s around the fact that you need to really understand it to change it.
09:00 PS: We’re not a technology company, we’re made up of insurance people and just by a couple of lucky breaks…well, sometimes you have to go with your gut instinct and make a decision. So I bought an Australian start-up company in 2012 pre-revenue, but I bought it from the guy that ran it, the chief architect, and he is something else. He’s not the easiest to manage, but he’s pretty fantastic. And one of the reasons that we do what we do is because we onshore all the development, everybody is employed by the company, it’s all onshore, very close relationship with the developers and we’re basically insurance people that have got some really clever technologists. But you need to understand that anybody that thinks they can make money in insurance perhaps will gravitate towards internet, distribution models, all that sort of thing. But if you’re going to do anything in claims, you’ve really got to know what you’re doing before you start, else you’ll fail.
10:09 Robin: I don’t want to impune any of our start-up friends. But is that why start-ups have not made the same impact in claims as they have perhaps in the front end of the business and around the data space?
10:21 PS: Well, I think that in claims, I see quite a lot of individual applications, but you’ve got legacy technology. Anything you can buy today is legacy. So either 10-15 years old, or even 25-40 years old. How are you going to put your innovative piece of decisioning software machine learning into Guidewire or a Duck Creek or whatever that’s going to cost the insurer an absolute fortune. It’s going to address a little part of the process, it’s not going to change the world. You’re going to go on the journey as a customer where it’s bad, it’s bad, it’s bad, it’s bad – oh – automated decision or something like that, it’s great. I’ve got an instant response – bad into BAU.
11:10 PS: So again from experience, I wanted the complete platform, I wanted everything included and I knew it takes a lot of doing, but actually it’s the only way to do it by having all of these digital tools embedded within a self-configurable platform. That’s really transformational. And that’s why you need lots of patience and lots of money, but we’ve now got global carriers with the technology, six of the top 12 motor insurers in the UK, half a dozen US carriers and stuff like that, so we can see it turning. But God it’s been stressful.
11:56 Robin: So I’ve been saying we’re at the tipping point for about 15 years without being right once, but how does the future look, are we really at the edge of a sort of transformation in the way claims are paid? How does the relationship between insurers and innovators play out from here over the next couple years?
12:15 PS: Yes that’s interesting, there’s a lot invested in legacy technology but it just doesn’t do what you want it to do, and if you attempt to make it do what you want it to do, it’s going to be costly, it’s going to be compromised. You really want to start with a clean sheet, and there’s not many people that can give you a complete clean sheet platform, which is the problem.
12:40 PS: I actually feel for those people that may have a good idea or may have a nice bit of technology, but they must find it terribly difficult trying to penetrate insurers with their legacy technology. I’ll tell you a little story that probably sums this all up. So, for a global insurer, we got the support of global procurement, the global claims manager, and in Summer 2017, they said, “We’re going to mandate your platform worldwide, so nobody can do anything unless they use your platform. If they want to do something different, they have got to make their case to the board.” Oh, so we thought, “Great yes, I’m going to retire now next week.” But what happened was the grant we were taking was immediately re-taken by a massive IT department, a spend of three billion pounds on IT. And we made them irrelevant. We are making them irrelevant for claims.
13:46 PS: We only needed to plug in the internet. And so they re-took it, “Oh we can do that, we can do that.” So two years on there’s not a sniff of what they’re doing. So you get involved in power-play politics at this level, and that’s not easy either, but you have just got to keep working at what you believe, in what you’re doing. But I would say it’s very difficult for anybody out there who’s got individual applications, because of the difficulty of making them work against a background of legacy technology and there’s such a lot of money invested in the status quo as well, and complexity. The big news is it’s not complex. You can simplify it all. Bit like a duck, we do all the paddling under the water. And the world has changed. It has absolutely changed in the last three years. What was only a dream now is day-to-day reality. And you’re faced with people who think it’s far more complex than it is and they don’t believe you, but they do when they get involved with it.
15:00 Robin: Paul thank you, we’ve run out of time. Paul Stanley, thank you very much.
15:03 PS: Thanks, everybody, thank you.
15:08 Robin: Now we just told you that it’s really difficult for startups to break into the claims world. So next up, we have Shift who today raised $60 million. They have broken into the claims world. No one else in the claims space is raising that kind of cash. I’ll let you guys introduce yourselves. But what we really want to know is, how do you make yourself worth $60 million and how are you going to spend it?
15:35 Jeff: Wow, [chuckle] interesting, and thanks for putting us on the spot, but actually I’m going to start with what are we going to spend it on? So last night, we celebrated in InsurTech fashion by going to an all-you-can-eat Chinese buffet so that was the start of where we went with our cash. But actually really for us, the story started like all good stories, it really was about being in the right place at the right time. So we have three founders, we’re all 28, 29 and 29, and they were working for AXA Insurance in Paris as interns, and their job really was to find a global fraud solution that met the needs of all the global customers for AXA. So, this in itself was a task, and I think about eight months in there was nothing actually out there that met the complex needs of an insurer of today to be able to industrialise the process. It is very manual.
16:39 Jeff: So maybe this is just part of their plan actually because the next step was, the AXA board in Paris were very trustworthy and confident that these three founders could actually go out and produce a working AI that could industrialise fraud detection within insurance. So it took us about eight months to create a solution that could work across several lines of business. We started with a head start really because as part of the findings we knew exactly what an insurance company needed to be able to progress in this area. And I suppose fast-forwarding on from then, from Converse and skinny jeans, today we wear suits and shirts and our insurer clients wear skinny jeans and converse, so it’s interesting how things have changed. And I think like everyone has said, it is a difficult market to break into and very frustrating actually at times because we see an end to what we want to do, but to get to those steps, we really have to jump through some hoops to get there.
17:56 Jeff: So today we have progressed, we raised our series C yesterday, we have a new VC, Bessemer, that has entrusted in us to invest 60 million. And I suppose answering the question about, where do we spend that money? We spend that money on people. So originally we started with four people, and today we have 215 people. From a customer perspective it’s really about saying that while other people have a car, Shift have a car and a driver. And we call ourselves a SaaS+ solution, so we supply a data scientist along with all of our clients.
18:42 Robin: Of your 200, how many are data scientists?
18:44 Jeff: So today we have 215 people and 110 of those people are data scientists. And with all start-ups really we’re judged basically not on the profit we make, but actually on the scale that we do. So we’re almost this kind of false economy of a business that’s not really judged on profit but judged on scaling. This year our objective is to be at 400 people by the end of the year. And so it’s my responsibility as a director and Thomas’s responsibility to make sure that we bring more people to Shift, and this is particularly difficult for us, as much as creating the technology.
19:26 Robin: And what’s in the products? Are you just sticking to the claims fraud, or you going to break out from there?
19:33 Jeff: Okay, so fraud really is the first step of automation for us. So it’s about industrialising a process that is very manual. So the epitome of that moment really was for Jeremy, our CEO. His first ever insurance claim was a flood in his flat in Paris, and it was a complete nightmare. I won’t give you the French terminology, but it’s probably not polite either. So someone at that age trying to log a claim is almost impossible. So it’s an area that we decided to work in. We believe getting your house in order to make sure you have a stringent forward process in place is really the first step of automation. So fraud isn’t really considered like that by people, but really it is a process for us.
20:21 Jeff: And considering actually that 10% of people, between 10 and 20% let’s say, are committing fraud, 80% of people are not committing fraud, so for us, it’s also about, how do we reward people? And certainly here in the UK market where the aggregator sites are making us wanting to change our insurance company on a yearly basis because of price competitiveness, how do we differentiate ourselves as an insurer to be able to keep our customers? So we want to look after the 80% of people. We want to say that, “If you’re doing the right thing, how do we straight-through process your claim to be able to pay you in real time?”
21:00 Jeff: So Shift provides a cleanliness score on a claim that allows people to process their claim in real-time to be able to pay customers. This works particularly well in claims like travel, and this is really our first step towards moving out of Force, which is our fraud detection solution and moving into Luke. And Luke is our claim automation solution. So you’re probably working out that we are huge Star Wars fans and Star Dust is the name of the program. We talked about it last night, I think we’re going to run out of names eventually.
21:51 Thomas: And since you mentioned Luke. Luke is all about our solution for claims automation. And it’s the logical next step following to the fraud detection, which is all about well, centreing, revolving around text recognition, and text recognition as we all know, dominates insurance still. And the claims process in particular. Of course, there’s image recognition as well, but for us, the expansion following fraud detection is really to look at the complete process of claims from the first notice of loss to text recognition. Then, having an actual AI in place that not only extracts the content, but also autonomously can decide when it doesn’t really recognise accurately enough what’s in that particular case and so it hands over to a human claim handler. And that is where, again, some of our experience comes in from the fraud detection part, which is all about enabling investigators to hand over efficiently and effectively to a human claim handler, so he can really take the next steps as effectively as possible.
23:37 Thomas: And the Luke claims automation is, of course, not 100% claims automation that will make us all useless, but in fact, it takes on the routine work that it can detect and identify and everything else where, on its own, it understands that it’s accuracy wouldn’t reach the 99% that we’re aiming for, it passes on to a human claim handler. And thereby allowing insurers to clearly separate which part of the process is handled automatically or is still handled by a human claim handler.
24:26 Jeff: And Luke was designed specifically this way, so as far as we know, it’s the only AI that has the ability to say, “I don’t know,” and this is not a fault, it’s actually designed in this way. So for Shift, we must be at least 99% accurate in our confidence level that we’re making the right decisions. So what we would say is that currently the numbers look like 60% of claims are 99% accurate. The other 40 are in learning and we’re continuing to do that, but actually we probably say that between 10 and 20% of the claims are actually reported digitally currently anyway. So it’s a learning process for us.
25:08 Robin: Guys, thank you. Jeff and Thomas, thank you very much indeed.
25:11 Jeff: Thank you, Robin.
25:12 Thomas: Thank you.
25:16 Robin: Next up we’ve got Avi Baruch from Previsico. Come and tell us what’s going on.
25:22 Avi Baruch: Hello, thank you very much. It’s very challenging to follow someone who’s just announced they’ve raised 60 million. So, I’ll try my best to not be upstaged. So I think the question that a lot more people, especially in the flood industry are asking themselves more and more now in 2019, is, “What would you do if you knew exactly what streets would flood, right before a storm event actually hits?” I know from my perspective, the first thing I would do is, I would move my car, I would go home and move my TV and get my family out of there.
26:01 AB: And I think this is something that we should all be thinking about because that’s where the technology is certainly heading. So as they mentioned, I’m Avi, I’m the Chief Operating Officer of Previsico. We’re a spin-out company from Loughborough University, who have taken our solution, which is called Flood Map Live, to market. We just launched this year in January. But the project has been ongoing for about 17 years, working on using a flood modelling system to predict exactly what streets will flood and in real-time. So globally flooding costs the global economy $90 billion a year.
26:43 AB: And that’s rising very significantly and it’s expected to be at least a trillion by 2050. And those are in my opinion, very conservative estimates, because they don’t include many of the other sorts of impacts that we see from flooding. And my imagery when I see flooding, especially in the TV is just cars bobbing up and down the streets, people looking devastated and showing us their homes, which are full of damages. And all of these, most of the vast majority of these are avoidable. I was in Carlisle after the Storm Desmond floods in 2015, and all you could see across all the streets were skips full of TVs, full of furniture, full of treasured belongings and sentimental items.
27:28 AB: In my opinion, none of them should have been in the skip, they should have all been protected and avoided. And I think that’s part of the problem that we’ve become used to, especially when it comes to flooding, it’s not having the right sort of insights to be able to take actionable decisions to reduce our risk. Currently, that’s why there’s so much work going into flood forecasting and some really good schemes in place such as the European Flood Warning System, it has generated returns of investment for $400 euros for every euro that was invested in it.
28:00 AB: The challenge with especially these sort of systems is that often for rivers and the sea, where you can detect how the levels are changing and then, based on a few pre-run scenarios, it is able to tell people roughly which areas are going to be affected. I have flood alerts on my phone, and they’re always buzzing. I can be on top of a hill on a sunny day and I’ll still get an alert saying you’re at-risk. And that’s part of the problem, we just get so many false alarms, and that at least from talking to people in Carlisle, that’s the sort of issue that they were having.
28:36 AB: The issue for surface water is even worse. So in surface water we don’t get any high-resolution impacts, we just get a national assessment because surface water can be very difficult to predict. Hourly changes in rainfall predictions can massively affect the types of streets that are going to be affected by flooding. So what we did at Loughborough University, we produced a system which takes combinations of different weather scenarios and continuously models them in real-time to produce a map showing exactly which streets are at risk and when to produce this sort of actionable insight; the worst case scenario and the best case scenario.
29:15 AB: And that was done together with a cabinet office to help emergency responders be able to tackle flooding. And that project was very successful so we got to innovate UK funding, to form the spin-out, and the insurance claims industry is a very good place to then take that to market. We’ve been talking to few insurers and I think there’s been a lot of interest, particularly in helping on the claim side. So I think one particular area, as well as knowing which areas are going to be affected so we know where to send loss adjusters, is warning customers. I think that’s where, prior to a flood and experiencing the claim actually happening, it’s really important to tell the customer, “Your street is at risk at 3:00 PM today, please do as much as you can to reduce that risk.” And I think that could radically change the sort of relationship insurers have with their customers, and it’s sort of the relationship that I really think should have happened a long time ago and it’s certainly possible now with the latest technological innovations.
30:18 AB: So that’s really where I hope that we’ll go. If we look at, for example, Storm Desmond, it cost the UK insurance industry $1.3 billion. So if in my case, I would’ve probably saved up at least 10% of the cost by acting in advance, that could have saved at least $130 million, so that’s quite conservative. So I certainly think that this solution can have a lot of impact for the claims industry. So thank you very much for listening.
30:44 Robin: Thank you.
30:48 MG: Well, have you found that useful? If you’d like to learn a little bit more about what we had going on for our claims event, you can find out more information on our website at www.InsTech.london and you’ll find the event specific page there and you’ll see some of the presenters for the night and also photographs, before we’re going on. If you’d like to learn more about what we’ve got coming up, you’ll also find that on the website, and indeed if you yourself would like to present what you are doing, then please register through the website or contact us at [email protected].