Three more “sofa sessions” in the second extract from our “MGA, the new frontier for Insurtech” evening event recorded live at The Steelyard London on 2nd April.
Robin Merttens talks to Matthew Reed founder of health insurance for business provider Equipsme and Andre Symes from their tech partners Genasys. (0.42)
Sebastien Piguet, co-founder of Descartes Underwriting, talks about the bespoke natural hazard cover they offer with their MGA (8.10)
Matthew Grant discovers from Richard Coleman how his recently launched Archipelago is getting on, and Justin Davies explains why Xceedance is partnering with them. (15.15)
Wondering what an MGA is? Matthew explains in his recent article published on Linkedin: MGAs: the fast track to insurance innovation?
Transcript for this podcast
00:00 Matthew Grant: Hello and welcome to the InsTech London Podcast. This is Matthew Grant, one of the partners from InsTech London. In this episode, we're delighted to bring you three more discussions from our recent event on the 2nd of April, ‘MGAs The Next Frontier for InsurTech’. You're going to hear from Equipsme, Genesys, Descartes, Archipelago, and Xceedance, and you can learn more about the event and see what we got up to from our website, www.instech.london and look out for the final section of the discussions we had, on our next podcast.
00:41 Robin Merttens: Now, we have Matthew, who has his own startup, Equipsme and Genesys in the shape of André who's their tech provider. Matthew, what does Equipsme do?
00:55 Matthew Reed: We provide health insurance for UK business, basically.
01:00 RM: That's simple enough. Where are you in the process?
01:01 Matthew Reed: Behind the Tate Modern. And we've just finished UAT on renewals. So we're into our second year of trading, which reduces our odds to survive by about 45% apparently.
01:13 RM: And what's your distribution model?
01:15 MR: Multi-channel.
01:17 RM: So you go direct and, sort of, split after one year?
01:21 MR: Well, our plan A was brokers and as I said to the team before we spun out, there'll be something that comes from left field, Plan Z, which probably eclipses Plan A. Brokers, I've been a broker so I can say this, are pretty slow. And so we've got some interesting things going on in Plan Z which have opened our eyes, that we never thought about.
01:41 RM: Are AXA the only capacity?
01:43 MR: They are the only insurance capacity. We then have services provided by remote GP, remote health check, and stress counselling.
01:53 RM: And there are a few people doing what you do. How would you make yourself special? What's the source?
02:01 MR: I suppose it's a very simple product that we actually built by asking customers what they wanted.
02:08 RM: So tell me the technology, what's involved? And I think I know who built it…
02:13 MR: Yes, he built it. It's quite complicated, I think we do ourselves a disservice just talking about it. It's simple but to make things simple in insurance is incredibly complicated. So we have a two-stage platform which has a company element and a payment engine, and an individual element and payment engine, with individual upgrades on both platforms. So, you're renewing an incredibly complicated thing. We met these guys, we did due diligence on six platforms. They won, but they won for many reasons, one of which was they said they'd never done it before, so we quite like the fact they were honest. And since then, it's been a journey.
02:56 André: It's been a journey.
02:56 MR: Yes, a journey as they say.
02:58 RM: So were you a supplier who, when asked to fill in an RFP said they couldn't do one of the things, because that would be a unique thing in my view.
03:07 André: Yes, but we were very honest about it from the beginning and we said to Matthew that we think we can do it, and he trusted us on that.
03:13 RM: And what challenges did he give you that you relished so much?
03:18 André: As you could hear from his opening, Plan Z is already in place, so it's about the plan changing the whole time and I think another very big important part for any startup is that every client matters, and because you own and manage, there is a lot of pressure they apply on to their vendors, and to their suppliers. So that was probably one of the challenges.
03:36 RM: And I see this a lot. Matthew, how did you pay for this?
03:41 MR: One of the decisive moments in the due diligence process was when we worked out a pricing scheme - I won't embarrass André because you guys were all wanted. We worked out a pricing scheme and it was a mixture between a fixed fee and an upside from our GWP, and these guys said, "Well okay, that's fine, that'll work but at some point, we will need to cap the GWP element because that will just hack you off." And we had other vendors that were like, "Well, actually that doesn't matter does it because at that point you'll be earning enough to pay us enough" and that's not the attitude actually.
04:13 MR: And so what we found with Genesys is that we've been able to talk almost transparently about where we are in our runway and how much money we've got. There's absolutely no point in them double charging us if we run out of money at Christmas, because then we're not here. So I think the flexibility of approach around pricing has been really beneficial for us, because when you go and raise money and you turn up with a fantastic pitch, the first thing that the guys in front of you do is halve it and then probably halve it again, and they're probably right actually. And so you've got to have that flexibility and that open relationship with a provider and say, "Look, you can charge this, but it will just run out quicker."
05:00 RM: I'm presuming it's a continual relationship. I mean, the build goes on, you must have other things you want to do.
05:03 MR: Yes, we just finished UAT on renewals, we are UAT on the API feed which allows us to put our product on other people's platforms. We've got two more products in development, so by October, we'll be able to insure every business in the UK. And it just keeps on going.
05:22 RM: Well done, you. Okay. We've got time for some Q&A.
05:29 Audience question: Hello. So you said that the time to delivery was three months. How many people were involved in that build?
05:37 André: So, on Matthew's team what we had was a stakeholder representing us in the organisation, being his Ops Director, which was pivotal. So it was effectively one Ops Director within the organisation with a couple of their staff doing the UAT and some of the specifications, and we've got a substantial team back in South Africa, but I think our delivery team for that project was about seven staff.
06:00 Audience question: André, I'm a big fan of the Genesys platform obviously. Why are people still buying Guidewire? How are you getting your brand across to the market?
06:10 André: I think, possibly, it is career management.
06:14 Matthew: You don't get sacked.
06:15 André: You don't get sacked for buying IBM, we all know that, and obviously, there is also pedigree. We are a young company, we're only 20 years old, we've got 100 staff, you can't compete on an RFP versus a Guidewire. When we come in at our price point, people don't trust it either, and also we promise we can go live in three months, people just go, "Oh you're talking shit, go away."
06:35 RM: Also, they didn't have £120 million to spend.
06:38 Matthew: Yes, Yes, Yes.
06:41 Audience question: André, hi, I also love Genesys. How long does it take for you to onboard new products, in comparison to normal policy admin systems?
06:50 André: We invested quite heavily about a year ago into building a zero code product, we call it 'Build' ironically. And with that product, we can onboard something, a simple single line of business, and take it live within about a week.
07:03 MR: Yes, we didn't start ‘speccing’ renewals until October, I suppose, and we've renewed our first policy in March. We haven't even started on a solo product. And that will be towards the summer, before going live in October. It works for us.
07:20 André: Yes. And that's about two week's ‘spec’ that we did for you.
07:22 RM: One more question.
07:24 Audience question: Hi Matthew, this is a question for you. So, we've heard about the how and the what, but where is the why? Why did you do this? What made you want to do this?
07:36 MR: Simple. Personally, I got fed up with making other people rich. Business-wise, 95% of businesses in the UK don't buy any health insurance for their staff, so we're not fishing in a small pond. So, we had to have a go.
07:51 RM: Good answer. Matthew, André, thank you. Good story. Thanks a lot.
08:00 RM: So, we have Sebastien Piguet from Descartes Underwriting, all the way from Paris. We thought we would have somebody from France here while we still can legally travel between the countries. Sebastian, your turn, what does Descartes Underwriting do? Tell me the story.
08:23 Sebastien Piguet: So, we develop a new insurance product on behalf of different risk areas. We cover all natural hazards. And to do so we leverage new data sources from satellite images, to all IoT devices.
08:43 RM: Nice. So, what classes are you doing now and what classes do you intend to do?
08:51 SP: We have no constraints. Among clients, we have companies operating in the energy sectors, in the agriculture sector, in many different sectors, basically all companies which are weather sensitive. So, the only constraint is that we have to cover natural hazards.
09:15 RM: So, that sounds complicated from a tech perspective. Who built the tech? How did you set about putting that together?
09:23 SP: So, data is very important for us. First, to improve the customer experience - we settle the claims automatically based on data. Data is also important for us to insure new risks. We do a lot of NDBI, to cover risks for it that traditional insurers are not able to cover. The last part is that we leverage data to have cutting edge models. Basically through parametric insurance, we avoid flood risks. We have access to a very large amount of data to improve our pricing. And we develop physical models, as opposed to statistical models. So, we model the underlying risks of the temperature, the precipitation, and we build a model to have a better understanding of the risks that we underwrite.
10:18 SP: In terms of technologies, our platform is coded in Python, and it is developed by a team of data scientists and meteorologists.
10:30 RM: So, would I call that parametric? Which of your products is parametric? Or are you parametric by depending on what the requirements of the customer are? Do you see what I mean? You said you have automatic claims payments. Is that always automatic or sometimes automatic?
10:51 SP: As soon as we receive the data, we are able to settle the claims.
10:58 RM: Cool. So, I was going to ask you what makes you special? But I've just worked it out for myself. [laughter] What are the datasets? I don't tell everyone where you get your special data but are they publicly available data sets? How did you set about getting them?
11:12 SP: Yes. We use a lot of data from NASA and from the European Space Agency... So, there are already plenty of data sets available for free. But we use also data from weather stations, from buoys, from gauges to measure water levels, for instance. So, as soon as we have a reliable data set, we are able to leverage it to cover new risks.
11:39 RM: If I don't ask this, then someone else is going to. What's the distribution model?
11:43 SP: So, we fully rely on workers. This is very important, thanks for asking this question, because for us it's very important. We want to build sustainable and strong relationships with workers. Basically, based on the needs that are collected by the workers, we are able to design a new product, within one month sometimes, to be very responsive, and to partner with them to better understand the needs of the client.
12:15 RM: And you haven't got a secret, distant mediation strategy that you're going to tell us about?
12:18 SP: No, I think for us it's really two different businesses. We are technology-driven, we develop new models, we propose new products. But the relationship with the clients is better managed by workers. We are covering risks over all continents. And basically, it would not be our goal to bypass the workers, as opposed to many different large risk carriers.
12:49 RM: And last one, is there an insurtech scene in Paris? I think I've seen three or four really cool MGAs emerging from market in the last 3-6 months. Is there a sort of vibrancy? Why suddenly all this cool stuff..?
13:07 SP: Yes, we saw a lot of new startups and advances in data analytics and data science, so I think that's one of the main advantages of the bio-fintech ecosystem, I would say. But I'm very often in London to meet people from the London market.
13:27 RM: Okay, we've run out of time. We've probably got one or two questions, anyone got a quick question for Sebastien?
13:32 Audience question: I was just wondering, when you talk about automatic claims payments, what happens in circumstances where the data is wrong or data has been, for example, hacked. How do you manage that risk?
13:45 SP: So we define fall-back methodologies in our contract, so we have a way to automatically replace the data, and of course, that takes more time. So usually that's how we solve this kind of problem. It could be an insurance against a drought. For farmers, it could be insuring a fish farm against a height wave that destroyed it. It could be insuring a hotel against a hurricane. It could be also an energy company against a warm winter. So everything that is related to weather.
14:24 Audience question: How do you help your customers decide what parameters to choose, because obviously there's a risk that they'll buy too much or too little cover and they may be unsophisticated.
14:34 SP: So we work closely with the workers to propose something which is relevant from the start for the clients, but we design ‘bespoke cover’. So really, the discussion is with the workers first and with the client afterwards, but it's only on a case-by-case basis. We use also historical losses, if the client is willing to share them, to calibrate our cover and to design it in a relevant way.
15:09 RM: Brilliant, Sebastien, thank you very much.
15:10 SP: Thanks to you.
15:11 RM: Thanks a lot.
15:14 MG: Next up, we have Richard Coleman, founder of Archipelago, a new insurance company providing coverage for home, motor, travel and pet, and he is joined by Justin Davis, who is the head of EMEA for Xceedance who built the technology for Archipelago. We join them a couple of minutes into the discussion.
15:37 MG: How do you connect what you're doing to the brokers and how do they in turn, link it into your end client?
15:43 Richard Coleman: So we set out to build a very intermediary friendly user experience, which is the platform that Xceedance have built, and then we've really used our connections in the industry to start building out our own network of small brokers.
16:00 MG: Okay, and given there's the competition out there, how do you market to the brokers or how do you incentivise the brokers to use you versus everybody else that's out there?
16:09 RC: Well, no one else is offering the breadth of product that we're offering, so it's actually incredibly hard outside of Chubb to go and buy that kind of package. We offer a very ‘easy to transact’ way, and people trust the expertise that we've got within our team.
16:26 MG: Great okay, so, easy to use, breadth of coverage and good tech. We'll talk about that in a minute from Justin, and your capacity provider is?
16:34 RC: Arch.
16:35 MG: Arch. Okay, fantastic, and so you've found your first risk two weeks ago, how's it been going since then?
16:40 RC: We're okay, I think we're going to be in double figures tomorrow.
16:42 MG: Fantastic. So 10 times...
16:44 MG: Your order of magnitude to be increased in two weeks, so if you keep going at that rate..!
16:47 RC: Yes, exactly that growth on a weekly basis is on a good curve..!
16:50 MG: Well done, it's a tough market out there. Okay, so Justin, Xceedance got involved in building the tech. It'd be interesting first of all, just to know about how did the two of you, or the two organisations get together in the first place?
17:04 Justin Davis: So, we've known Rich and Angus for a very long time. We do a lot of work with Aon, both ACT and GBC. So we met these guys there and actually, they had an idea for the MGA a long time ago. I've got the original NDA which goes back to something like 2015. Because we've known these guys a long time, and they showed us their business plan, we believe in them as people, we believe in the business plan, we were able to offer a three-partnership model. So similar to what one of the gentlemen was talking about earlier where it's as a percentage of earned revenue rather than GWP, so it's slightly different but that means you have to believe in the organisation. We are very pleased about the double figures because ultimately that means we're getting more revenue. Happy days.
18:02 MG: Just a quick word about, when you found your capacity how difficult was that, because I see it's been a big challenge to actually convince more people to put up capacity for their MGAs.
18:13 RC: Well, I think we were lucky in that the jobs we had at Aon gave us some very, very senior marketing contacts and therefore we were able to engage the discussions at CEO level rather than at an individual product line underwriting level.
18:31 MG: Okay.
18:32 RC: We also had some additional distribution that we can't talk about it, it’s not completely signed yet but that was very powerful.
18:38 MG: Good. We always love the stuff we can't talk about because it keeps the story moving with some anticipation.
18:43 RC: Yes, but I guess...
18:43 MG: So the more you can't talk about it, the better.
18:45 RC: Yes, but probably we've been very lucky in the whole number of organisations deciding that they backed us as individuals really. So in the same way, Xceedance have been extraordinarily supportive, in the way that they have worked with us. Arch made a decision that they really trusted our management team to execute for them. They were looking for a way of getting into the UK personal line space but didn't really just want to wade into the current melee. So they wanted a very differentiated offering.
19:16 MG: Great. So the key part of the MGA model is that you need the capacity and then you need the tech which is where Xceedance come in. Justin, could you talk a little bit about Xceedance? What are you looking for now you have this partnership with Archipelago? Is that an unusual arrangement or are you starting to see more companies you want to partner with? And when you say partner, what is that? Can you talk a little bit more about what that means?
19:38 JD: Yes, definitely. We are absolutely looking for more companies to partner with. We're looking for more companies who want MGA platforms, clearly. And actually, as a result of what we've been doing with Archipelago, we've seen two companies who've approached us and said, "We like what we see so far." And whilst we can't really show them the full Archipelago platform, because they might be seen as competitive, we can certainly show them something similar that we've done actually out in Australia. I've got another company, who I was talking to a year and a half ago, that's sort of resurrected the conversation. But we are also looking for companies that might be open to investment. So our founder is a guy called Arun Balakrishnan, he's a real proper entrepreneur. And in the same way, we've kind of invested in Archipelago because it's a huge risk for us and actually, it's a loss-making proposition to us for a long time. But slightly differently, we're also looking to invest in an MGA or MGAs.
20:41 MG: Great, so no pressure Richard, “loss-making for some time”. So Richard, one thing that you've learned from this that you would use to give advice to anybody else looking at developing a partnership with their technology provider?
20:54 RC: If you've got a good idea, don't quit your job until you are absolutely ready to go. [laughter] So my co-founder and I spent three years on someone else's pay, developing the idea until we were ready to go. So, always assume it's going to take longer and cost more money than you think. So, if you can, do it in the evenings and the weekends, then that actually gives you a huge amount of advantage over it; jumping at the right time. It's about trust. There are six of us in our business, none of us come from a technology background. Only one of us comes from a personal lines background. We entrusted the success of our business to Xceedance.
21:34 RC: If we'd have had someone in their office at all times, or they'd have been in our office, ie we'd been face-to-face at all points, we could have reduced the build time by a couple of months. So I think that's an important lesson.
21:50 MG: Okay, I think we've got time for one or two questions from the audience.
21:53 Audience question: Thank you, and probably directed on the tech side. It's a bit of ‘peeling the onion’, the solutions around payment, remittance, the user experience. Is that an all-in-one tech solution? And how long did it take you to get clarity on that solution and bring it to market?
22:23 JD: So I think the user experience was pretty much driven by these guys. They understood what question sets they wanted and that's part of their IP. And on the payment side, that's pretty simple nowadays, you find the right direct debit supplier and you just API it. That's relatively simple. But yes, it is part of the whole solution.
22:46 MG: Okay, and one final question? Thanks, Justin.
22:49 Audience question: Yes, for Justin, I don't know if you said this but how long was the build? And where you said “it would have been better if we were face-to-face”, what kind of cost implications would that have had for a start-up and was that really viable?
23:10 JD: Yes, you can... [laughter] You're better placed to answer that.
23:12 RC: So I think this is the thing about you need to raise enough money to do it properly. We spent quite lot of time in Delhi with the guys from Xceedance, we had them over a lot as well. So probably for 50% of the build-time maybe, we've had people from one camp in the other one's camp. To have done it for the whole build would have probably cost an extra 50,000 but for us that would have been significantly outweighed by getting to market with eight weeks quicker.
23:52 MG: Great. Okay, Richard, Justin, thank you very much.