Subscribe
Share
Share
Embed
In this episode, William Faulks, Chief Actuary at Antares, explores how actuarial roles are evolving in a fast-moving, data-rich MGA landscape. The episode explores the growing importance of aligning actuarial insight with underwriting judgement, the challenges of managing vast and often unstructured data, and the balance between real-time analytics and structured reporting. Will also shares practical examples of data-driven decision making, reflects on the skills needed to stay relevant, and offers a grounded view on how AI and better data practices are reshaping risk and pricing decisions.
Creators and Guests
What is Insurance Technology: Fact or Fiction??
Host Georgie Simister speaks with key insurance industry figures to separate fact from fiction in a world of insurance, debunk industry myths and explore the game-changing innovations shaping its future.
Get in touch to suggest our next guest: https://artificial.io/contact/
What if the biggest risk in insurance isn't AI, but drowning in data and missing what really matters? My guest today is William Faulks, Chief Actuary at Antares Insurance Company. He spent much of his career at the intersection of insurance and technology working across tech led MGAs and insurers. At Antares, he helps provide underwriting capacity to MGAs and focuses on how data and analytics can improve underwriting, pricing, and risk decisions. Will, welcome to the podcast.
Will:Thank you for having me.
Georgie:I asked you for your intro, which I think was very humble. So I've got to add add some bits to it. So you studied at Oxford
Will:I did.
Georgie:Where you studied physics. Mhmm. And then I was looking at your LinkedIn. What I found most impressive is you started Zurich as a graduate and then your final role at Zurich was head of schemes in case pricing, where you managed a team of 10 analysts. Yeah.
Georgie:So in four years, that's quite a lot to achieve, wouldn't you say? Oh, thank you. Yes.
Will:Oh, that's the question. I would largely put that down to the way Zurich handled with graduates. Like the and this is not a sales pitch for Zurich. Zurich have a phenomenal graduate scheme and Actuar Academy. And one of the big things they focus on with Actuaries is in about four to five years, how can we turn you from a graduate?
Will:Who knows? Let's be real, very little.
Georgie:Yeah.
Will:Yeah. Into a fully functioning member of the team. And yeah, it was excellent.
Georgie:But not just fully functioning, ahead of a team managing 10 people is really impressive. So, yeah, I think you are really positioned well to discuss today's topic. So looking at today, you are now Chief Actuary at Antares. You're supporting MGAs with underwriting capacity. How would you describe your role?
Georgie:How would you describe the role of the actuary in today's MGA driven market?
Will:So the role of actuary nowadays is to sit between a lot of functions and to do a lot of the analytical work that helps make decisions happen. As what actuaries have historically been good at is taking data sets, looking at trends and then giving them to other people. Nowadays what we are doing is we have to be essentially what's known as front of house. Where what we do on a day to day basis and it can be anything, if we're working for marketing, working for underwriting, working for my distribution partners is looking for problems, looking for places where we think that something is suboptimal, working with the subject matter experts to figure out what data there is here, and then doing analysis and coming up with proposals. And sometimes the decision maker is the actuary.
Will:Great. Sometimes the decision maker isn't the actuary. So you will work with them and give proposals, trade offs, and ideally make a nice, quick, fast decision to make something better.
Georgie:What drew you to this space in particular?
Will:So the thing that I really love about the space is speed. There's a common metaphor that's used of MGAs are speedboats. They're built for quick, fast decision making. Carriers or oil tankers, they are about stability and long term views. So they are companies that are fundamentally built for different reasons.
Will:And as an actuary, what's really good is when you can do a piece of work, you can put it into practice quickly, you can see the result and you can learn from that feedback loop.
Georgie:And
Will:the great thing with the MGA space is that feedback loop can be very, very short. And especially in the retail space, it can be short and you can see the actual direct customer impact of it. So, yeah, that's honestly what I really like about it. It's you get to do you get to do the fun math. I love the fun math.
Will:I'm I'm never gonna shy away from that. But you actually get to see it in practice quickly, fast and have a massive impact.
Georgie:Mhmm. And so well, I guess there's two different lenses then for an actuary. So that's the carrier lens and the MGA one which you've taken. But through the lens of the MGA, how has the actuarial role evolved as the MGAs become more specialised and especially as they become more data led?
Will:So I'd say one of the main differences now is it is a default expectation from an Actuary in a business at an MGA especially, to really understand your products, understand your customers, understand every aspect of the business. You still are part of an actuarial function. Yeah. But nowadays, it is completely expected that you are fundamentally part of the business. Mhmm.
Will:And you are not a back office function, you're a front office function. And especially when you're talking about kind of niche product lines, you can't simply just get some data and do some analysis, give it back. Because the changes you make will have very large effects. And you need to be completely confident that with all of your stakeholders, you're all on the same page. And the role of actuary is the role of a conduit nowadays, where you need to make sure that the underwriting and marketing are on the same page.
Will:And a lot of the time, the analysis is done for both teams by you. And I think I said before, I thoroughly enjoy that.
Georgie:So then if we're looking at the business and you mentioned underwriters, where does actuarial insight and underwriting judgment complement each other kind of best today?
Will:What actuaries are very good at is looking at portfolio views, looking at trends, looking at what's happened in the last twelve months and see insight from that. What works really well is when you have underwriters who are willing to challenge that. And I'm aware that's dangerous to say as an actuary. But what I found works exceptionally well is where actuaries come up with a model, come up with predictions.
Georgie:And
Will:generally speaking, these will be good for, let's say, 95% of the time. But that 5% where there's something special about it, that's where underwriters really come in. Because they really fundamentally know the business. They know the insurance they've been operating in. They've been doing it for five, ten, twenty years.
Will:So these edge cases, they're not edge cases to them. They live and breathe them. So they can work with you and say, Okay, you're seeing something interesting over here. We don't have the numerical data for it. But this is why that's happening.
Will:And this is how you can adjust for it.
Georgie:And
Will:something that I think people don't give enough credit to is when someone says the phrase judgment. Historically, that phrase it can go either way. People can be like, Oh, we rely too much on judgment or We don't have enough judgment in here. What really judgment is, is it's data. It's a lot of experience that you are coalescing into a view.
Will:And so one of the best parts of being an actuary is to work with the underwrite to say, You have a lot of experience here. Let's quantify it and let's get it into our model somehow. And I found at least in the last ten years or so, when I started, there was a lot less of that. And now where I am, that's part of everyday life.
Georgie:You said how the role of the actuary has expanded, so it's no longer just a back office role. Do you think it will continue to expand? Or do you think the underwriters themselves will need to be upskilled and have a much more of an actuarial view and a skill set as well?
Will:I actually think both are true. So, yes, firstly, the underwriters, you can see very clearly from top notch underwriters that they I say expanding, a lot of them have expanded their skill sets already to be able to do the analytics at a first pass themselves. And the best underwriters I've worked with generally don't come to me with a question of what if this? They come to with a question of what if this? And this is what I looked at and this is why I think it already makes sense but let's tweak it together.
Will:The Actuary Scale Set also has to expand because nowadays what you're seeing in a of MGAs is there's also a data science team. It always used to be underwriter actuary. It's kind of now underwriter actuary data science a lot of the time. So it's not just underwriters who are expanding to be more technical. Actuaries also need to be able to work with people who, for instance, may have come out with three PhDs.
Will:And we need to be able to translate that unbelievably smart thinking into a business outcome. And throughout, I mean, a lot of the last ten, twenty, thirty years, the standard actuarial role has been the most mathematical in the room. It's often not nowadays. How intimidating. And you often have someone in the room who really doesn't have that much business understanding.
Will:But wow, they are so smart. And an actual draw was to help that person and the underwriter go in together and go, Okay, let's make this into an actual decision to make us money.
Georgie:Gosh, that's a scary world when the actuary is no longer the smartest person.
Will:I do hope that is not the one clip you use.
Georgie:Right. If we're looking at data and what is the actual reality of data, MGAs and what we're finding right now, so we're doing projects with MGAs who are just receiving a ton of unstructured data. We know that we well, we see that this is a challenge. But from your lens, what is the biggest challenge that this unstructured data creates for insurers and for actuaries?
Will:So the biggest challenge is really that when you have a small amount of data, you understand it really well. When you have a large amount of data and you know, nowadays by large, think I mean exceptionally large what you are seeing is inside that you almost certainly have nuggets of absolute gold. But you have a lot of data that either you don't understand exactly what it's doing and the time it takes to find that nugget of gold, you may end missing it. And the biggest problem is figuring out from a huge dataset what is useful. And yeah, that is one of the skill sets that we are really having to up as a profession
Georgie:of
Will:you can't just be sifting through one by one. There need to be much better, generally mathematical techniques that we are making to try and figure out quickly what is the best place to look at.
Georgie:Or
Will:go the other spectrum and ask a subject matter expert. You know this area well. I have a 100 different data points. Give me the top 10 to start with.
Georgie:And I guess as well, it's the tools that can help you take vast majorities of unstructured data to structure it, because that's day one where you should be starting rather than kind of, yeah, taking all of your time to sift through this unstructured data. So definitely, I think there's loads of tools that can help people get underway. But it's a very complex challenge anyway to try and cover. So on the point of having complex data and knowing what is valuable, how do you decide what generally does improve the underwriting decision and which data just kind of adds complexity?
Will:I actually normally start the other way around and start with what decision are we trying to make first and then work backwards. And if you start that way and look, if you cannot stop the decision, look at the data and figure out how this is going to change your decision, then it's probably not your top priority of what people can get bogged down in. And I've done this myself as well as we all. We we look to something that looks really interesting and we really want to work on it. But if you can't make that clear line from I will do this, I will get this analysis and don't get me wrong, I'm sure it will be really cool.
Will:But if you can't link it to here is a decision, here is a customer outcome, here is a financial outcome, then that's not going to be your first starting point. And that's, in my experience, gives you a much clearer set because you can almost always narrow it down to 10% of your visual dataset just with that. And from there, well, then you you can get more precise or, to be honest, spend an afternoon in the weeds with the data. We've all done that.
Georgie:Okay, Bordeaux, everyone's favorite topic. Bordeaux versus real time data. With data increasingly being traded in real time, does the traditional Bordeaux still have its place?
Will:It does. Okay. So with a lot of real time data and to be honest, any data that we're using day to day to make, what I'm going say, tactical decisions. This lets us do short, sharp changes that we believe are going to help the bottom line, help the top line, whatever it is. Clean, structured, regular border data is really helpful to assess on a regular basis without bias.
Will:Is it working? Because the great thing about a board row is it comes at the same time every month. It comes in the same format every month. You can run the same report on it every month and see, Okay, we did these seven changes based on our real time data. We think they were all very good, etc.
Will:But can we actually see in our nice standard clear reporting that things are getting better?
Georgie:Do you have a standard format for all of the MGAs to report to you? Or are you receiving different border sets in multiple different formats?
Will:Bit of both. So without going to the specifics, it's always nice when there's a standard format. However, insurers are getting much, much better about taking data that looks pretty similar and making it one format. Now, I'm not saying we should get one via carrier page and one via email. But yes, it's one of the great things when you are setting up a new relationship with somebody, you essentially have a blank page.
Georgie:Yeah, of course.
Will:So you can say, this is how we like to work. This is how you like to work. Let's find the most helpful way around.
Georgie:Mhmm. And then so if well, from an actuarial point of view and a capital point of view, is real time data always better, or does the structured periodic reporting still have advantages? I feel like you're going to say it still has advantages.
Will:I am. Yes. I'm gonna go slightly the other way of there are some disadvantages to real time data. Okay. As what you can find is if you get tons of stuff in every day, your instinct is to we have new data, let's work with it.
Will:Yeah. And you can spend a lot of time reacting to things very quickly. Mhmm. Which is generally great. But if spend all of your time reacting to short changes, you may lose focus of your wider strategy as we're So going with more and more real time data, you need a lot more discipline.
Will:As one of the most helpful things I was taught when I was an actuarial student is sometimes when you get information in, you need to understand if it's not the right time to react to it. And it may not be the priority. You cannot just react to every single thing you see Mhmm. Because you'll never actually get what you need done.
Georgie:Is there a middle ground, though? Because I feel like, you know, real time data is good to be able to pivot in real time to live changes, you know, adjust your appetite. And then, I guess, if you're always waiting for a bordeaux to come in, you're kind of being reactive. Is there a middle ground?
Will:Yes. So the middle ground is that your board row is there as essentially there's a check of if you are waiting every month for your board row to make any changes whatsoever, you are probably being too slow is the honest answer with this. If you are just using real time data and never going back and saying, Okay, for my border of data, is everything going as expected? You aren't essentially working with your governance that you need to be working with. So the ideal middle ground is you use as much as you reasonably can with your real time data.
Will:But never forget to anchor back to your board rows and make sure that everything you're doing short term tactical is excellent. But if you cannot line up to your boardroom showing that your strategy is working on a macro level, is going wrong somewhere. So it's essentially, it's your checks and balances.
Georgie:Yeah. Okay. That's a good that's a good insight. So if we're solving the problem and we're looking at what good looks like, so when MGAs, underwriters, and actuaries are truly working together and are aligned with data, what does good look like through your eyes?
Will:So we'll start with actuaries and underwriters.
Georgie:So
Will:I previously had a boss who was the chief underwriter, and he described this to me once in a way I really liked. He said, When you are working with the underwriter, generally speaking, you should be agreeing 80% of the time and politely disagreeing 20%. And you should be never scared to disagree. As the way he phrased it was actually, if you two go into the room, are there for an hour and come out saying, Yep, we completely agree, you've done it wrong. That's never going to be the case.
Will:Cultural. It's one of the best ways to explain that of you need an environment where you can disagree and you know that your reasons are going to be heard, understood and factored in.
Georgie:From
Will:the MGA side, the main thing that's important, I say, is alignment of incentives. When you are working together, as long as you have aligned incentives, you will be trying to get to the same answer. Therefore, you will have a very similar dynamic of what I just said with the unduisance actuaries. You may well disagree. You probably will sometimes.
Will:But you are both disagreeing trying to get to the same goal. Therefore, you're probably going to come up with similar answers and you're going to be using the data in similar ways.
Georgie:Do you have any examples of where better use of data and analytics has actually improved the underwriting or pricing outcome that you can share?
Will:Yes. I'm going to go actually slightly left field to one of my favorite examples. So previously at MGA, I was doing a piece of work with marketing. And this is one of my favorite examples because it's completely tangible. We were a small company and we were looking to go to customers who had lapsed their policy with us previously and say something along the lines of Our product has improved in the last couple of years.
Will:We've made all these enhancements. Perhaps you might want to renew your policy with us. We were going to also send out a nice little chocolate bar with us, as, you know, as you do. So our constraint was we only had so many physical chocolate bars, and we didn't have enough chocolate bars to send one to every customer that had ever lapsed a policy. Yeah.
Will:So what we were working through was who do we send it to to get the most customers in? And it was actually really fun. So we had an entire list of every customer, the reasons they had said they'd left. We had our product enhancers that we had with them. And what we went through is matched up and went, whose reason for leaving?
Will:Have we actually solved since then? And we basically ranked them and said, here are the people where they left for a very valid reason. We fixed that reason. Get them a chocolate bar. And it's one of the few times where you can physically hold in your hand And decision that you're a chocolate bar.
Will:Yeah. So, yeah, it was great. And it worked really well. Yeah. Like unsurprisingly, if you go to someone and say, hey, you had a problem.
Will:We have solved your problem. Do you want to come back? They often will.
Georgie:Mhmm. So I know you and I think you're very likable. I don't know many actuaries, but you're a great one to know. But what skills and mindsets do actuarial teams need to develop to stay relevant in the MGA led world? Be More Will?
Will:Oh, I would never say that on microphone. So I'd say the mindset first. The mindset is and I'm aware how this may end up sounding but you are not a mathematician first. That is a skill set that you You are a part of the business first who uses your mathematics as a skill. And as you've said from my CV earlier, I'm a physicist, I'm a mathematician.
Will:I absolutely love my maths. You've known me years, you know that. But that said, there's essentially a time and a place for it. And you need to approach every problem with we know what business outcomes we're trying to get to. How do we best do that?
Will:You don't need to approach the problem with how can we apply the greatest mathematical model to this? Because as much fun as that is, it really is fun. It won't necessarily get you to the right place.
Georgie:And
Will:that is one of the things that I found really interesting working like the InsureTech area of with all the data that you have, you really can apply some very incredible mathematical models to it. And finding that balance where you are using, you know, the most up to date techniques. But also you don't have all the time in the world. Yeah. You have to pick a lane here.
Will:So going with the best that you can to get the best answer as quickly as you can is often the way to do it.
Georgie:Mhmm. What would your advice be to the underwriter side, though, that's looking to strengthen their analytical analytical capabilities capabilities without without losing losing the the kind of the art of the underwriting?
Will:I'd say there are two things that I have seen in the best analytical underwriters that I've worked with. And they sound slightly counterintuitive to go together. The first one is be willing to be wrong. You may have a view that you've held for ten, twenty, thirty years, and then a large amount of analysis that you do shows actually maybe that isn't the case. But on the other side, be willing to be confident that actually this is a situation where you have enough experience and the data maybe isn't quite relevant enough, isn't up to date enough, that your view is the one you should go.
Will:And it's finding that balance which makes kind of a good underwriter great with it. And I've worked with some people where it's unbelievably impressive. You don't quite know how they switch that gear so quickly. And you see them go, Okay, well, the analysis does show this. That's interesting.
Will:I guess I was wrong. And then the next thing you talk about, they go, No, I I understand the analysis shows that, but here are all the reasons that actually, like my experience is correct. And it's phenomenal when you see it happen. It's a rare but growing skill set.
Georgie:Yeah. I guess it's where you're saying that 8020% comes in
Will:in the meetings.
Georgie:For any MGA listening, what's one thing that they could do to make their data more useful for insurers and actuaries? Help them help you.
Will:Interesting. The honest answer is when you are giving carriers data, make it clear what the data is. Okay. It's not the most exciting answer, I'm aware. But the biggest mistake that I can see happening in my time in the market is people receive data, think it's one thing, but it's actually another.
Will:And that's actually much worse than not getting anything. Because you will do a piece of work on an assumption that's wrong. The answer will almost certainly therefore be wrong. And you will make a decision that is detrimental. And on an incredibly practical level, if someone is sending me, say, a table of data, the first thing I do is go back to them and say, please just give me a definition of these data fields.
Will:What is this column? What is this? And the better MGAs I work with, they're so good at it. When they send you data, along with it, either in an email or somewhere else in the sheet, it will go, and here's the clear definition of everything we sent you. And if nothing else, the relationship is so much smoother because you don't need this back and forth.
Will:It's the trust it builds is amazing of I trust that you really understand what you're doing. And not only that, you're willing to spend the very short amount of extra time to make this relationship better. And I really appreciate it. And it makes us have much better decisions.
Georgie:I think as well you're saying about missing those like golden nuggets of information. It makes that a lot clearer to where to look for it as well.
Will:Oh, it really does. Yeah. I I am slightly sad thinking about the fact that almost certainly, at some point throughout my career, someone has sent me something that is the golden nugget. Yeah. And I just didn't know because I didn't understand what they'd labeled it as.
Will:Mhmm.
Georgie:Right. I'm excited to ask you this question because I'm excited for the answer. What is one thing that you wish more people outside actuarial teams understood about pricing and risk decisions and how they're actually made?
Will:Oh, that is fun.
Georgie:Teach me. So
Will:I'd say the main thing is that an actuary isn't trying to predict the future and say this is what's going to happen. That's often what people talk about. But really what we're trying to say is this is the most likely thing that's going to happen. Essentially, what we are doing is insurance. You are placing a lot of tiny bets and you are looking at a portfolio level for your bets to win.
Will:And the actual job is not to get every bet right. It is to make the underwriters have everything they need to have as good odds on everyone as possible. And almost certainly, you know, if you have a portfolio of risks, some will be priced wrong. And you'll do your best to minimize it. But honestly, that's life.
Will:Like you are doing your very best to make sure everything is good as it is. But you're not trying to be a fortune teller. And it's often what you hear in the conversations around of, Okay, well, actuaries are going predict the future. I wish I could. That'd be great.
Will:But no, we're just trying to say, hey, this is the most likely thing that'll happen. And here's how we can work with that.
Georgie:So funny because I'm about to ask you to predict the future. If you did have a crystal ball and if we were having this conversation in five years' time, what do you think will look fundamentally different about the way risk decisions are made?
Will:Into my crystal ball, go. I would say the main thing is when you're making a decision, there's a bunch of stages. There's you get information in, you work through it, you figure out what it is, you do a lot of processing at the start. And I'd say 10%, 15% of your time is actually doing the decision making.
Georgie:Maybe
Will:80%, 85% of the time, 90% is actually spent getting it in, understanding the data, doing the analysis. That ten fifteen is going to keep expanding. And I can already see it in my day to day life expanding, where more and more time can be spent thinking. And less and less time is spent cranking handles doing this. As much as I'm not one who thinks AI is going to take over the world, the help in that area is really good.
Will:Like one of the things that's like I have found, like the AI tools I use have been truly phenomenal at, is say, this is the sort of data I get regularly. Give me a start.
Georgie:And
Will:that's the phrase I use when I'm writing a report, if I'm doing analysis. It's help me start. And can you get me halfway through even the data prep stage? That's already hours of my life. And I can now spend that time thinking.
Georgie:Mhmm. I love that. I love a positive plug for AI and how it can be used in the market. So thank you. You're welcome.
Georgie:Also, we've reached the end. Oh. So thank you very much, Will. You have been absolutely brilliant. I will be having you back here again in five years' time.
Georgie:Thank you so much for listening to Fact or Fiction. If you like what you hear, then please subscribe to our channels. We'd also love to hear from you about future guests. So please get in touch via the link in the show notes or reach out via LinkedIn. Thanks so much again, and we'll see you next time.