Get a demo
Select the type of demo you'd like to experience with Indico Data.
Select the type of demo you'd like to experience with Indico Data.
Watch Christopher M. Wells, Ph. D., Indico VP of Research and Development, and Michelle Gouveia, VP at Sandbox Insurtech Ventures, in episode 38 of Unstructured Unlocked with Mandy Hunt, Former Chief Underwriting Officer – Commercial Lines, RSA.
Michelle Gouveia: Hi everybody. Welcome to a new episode of Unstructured Unlocked, Michelle Goveia.
Christopher Wells: And I’m co-host Chris Wells
MG: And we are very excited to be joined today by Mandy Hunt, the chairperson of the Underwriting Community Board for the Chartered Insurance Institute. Mandy, welcome to the podcast.
Mandy Hunt: Good morning. It’s brilliant to be here. Thanks for asking me to join.
MG: Absolutely. Looking forward to the conversation. Before we jump into some of the topics that we’ve discussed, can you just share a little bit about what the organization that you’re currently working with does, and then a little bit about your history prior to joining that group?
MH: Yes, of course. So the Charters Insurance Institute is the professional body for insurance in the UK and for some of the overseas territories around the world. So they have their main board that obviously runs the business, but to ensure they provide the right sort of technical content and guidance and support to the industry. They have set up three community boards, one underwriting, one insurance broking, and one for claims. And I’ve been part of that board now for four years and privileged to be the chair of it for the last two. And our job really is to bring a breadth of industry knowledge to the table so that we can pick topics that will help people in the underwriting part of the insurance industry learn, develop. So last year we spent a bit of time talking about careers and people’s journeys to getting to the jobs they had.
This year we want to focus a little bit more on some technical content around things like what is portfolio management and things like that. So that’s the volunteer role. I do a couple of other volunteer roles as well, but this is the one for insurance. Prior to that I was the CEO or Chief Underwriting Officer at RSA in the commercial lines business in the uk was with RSA for 25 years. Previous to that, I spent eight years working with national brokers and my last two jobs at RSA, one was the managing director of a business called Insurance Corporation to Channel Islands, which is one of the subsidiary companies of RSAs where well, MD I suppose speaks for itself. And then yeah, came back to the UK from my stint in Guernsey to the CO role, which was, yeah. So to be honest, the last 10 years of my career has been a real privilege, the kind of roles and activities I’ve been able to get involved with.
CW: Yeah, actually let’s drill in there. Do you want to talk about some of the highlights over the last 10 years? It’s been, I think kind of a wild ride in insurance.
MH: It’s been fun. It’s been fun. Yeah. So I mean, let’s talk about the Channel Islands because I think without a shadow of a doubt, that was a really a lovely period of time both personally and professionally. So Zi a small island south of the UK just off the coast of France, it is a dependency in the uk. So it is regulated by its own and has its own requirements to operate in certain ways. So being the MD was like running a mini insurance company despite the fact it was part of the RSA group of companies. And what was really special about that, apart from the day-to-day job, which was broader than underwriting, which is where I spent most of my career because it covered claims and marketing and hr, it was the impact of being in something so small and unique and getting to know the people and the culture.
So whilst I wasn’t lifting up and moving to a much further away place, it is different to the UK and it’s unique and quaint and it’s really important that you understood that because they’re the customers you’re serving. So yeah, really loved my time there. Got some really dear friends, had some great opportunities and had that opportunity to be the person on point with the main regulator for the island. So that’s very different to being in the UK business where obviously there would be a group of people responsible and the CEO, Ken Nor Grove would’ve been more responsible at a higher level. So yeah, really good from that point of view, ran was part of the board, so it’s an executive board and a non-exec number of directors. So that brings something different to your mix every day as well. So lots of opportunity there returned to the UK to become the CO.
So the role evolved over the five years I was back in the uk, but it covered in that time delegated authority arrangements or schemes, whatever kind of language you may use in the US and mid-market offering. So up to I guess a turnover of about quarter of a million, 250 million pounds. The online trading platforms, we call it E-Trade or SME in the uk. And then we at various points, I also had the London Market specialty lines business, but my role also included governance. I had a team of governance managers that worked for me for a period related to the FCA requirements for product governance. We had a data analytics team which was providing data not only from an underwriting point of view, but to the business for that day-to-day management. And I also looked after risk consulting, so the engineers that visit businesses and are the eyes of the underwriter and write reports back. So I had a real mix of responsibilities, which is fabulous.
MG: So many questions come to mind about all of those areas. There are things we talk on the podcast about a lot, but Minnie, one of the things that you said, I’m going to probably restructure maybe how we have this conversation. You did just talk about a lot of things that I’d love to dig into in terms of how data, data management all fit into that, but we typically wait a little longer in the episode to bring up the all powerful ai, but I’m going to do it now specifically because you mentioned when you were introducing what you were doing as the underwriting community chairperson for the Charter Institute, talking about last year, having a focus on professional development and how people got to the roles they were in. And something that we’ve talked a few times about on the podcast with different people from different roles is how do you think about how the role of an underwriter may change or shift as AI capabilities come in and is there a concern about the decision making capabilities that underwriters gain just by being in the day-to-day and having to vet through the data that they see and having their judgment?
How does bringing something automated in how does that change what professional development looks like? So I’d love to get your take on that.
MH: Yeah, so I’ll probably answer that in two ways. So I think first of all, from an underwriting day-to-day, clearly roles are going to change because consistency of data and being able to pull it through and make it easier and quicker for an underwriter to look at things will be great because from my days as an underwriter, you’d get a presentation in, you’d have to sift through every question you were trying to check everything, fill in systems, try and work out what you do with the rating engine and the amount of time you’re probably truly thinking about or what a risk looks like and the price and the terms that you want to put on that risk is probably a very small part of the whole time you are looking at it because you’re doing all of this processing stuff. So the joy of the future is a lot of that stuff.
Of course as you extract things out, presentations and you automate the way things get fed in, what an underwriter will spend their time doing is actually understanding risk. And that’s the part of their job that every underwriter loves. That’s where they bring their own expertise in very specialist areas. It’ll be quite unique, but even in the generic areas of motor understanding what is happening with vehicles, what we can do differently and they want to spend their time thinking about that, not figuring out how to get stuff in the systems. So it’s going to be exciting I think for underwriters in the first place and it’ll be something that we’ll have to embrace. I think personally the bit we’re going to have to do is educate underwriters how to use data and what to look for because we educated them to use systems and now we’re going to have to educate them to figure out what the systems are telling them because data is brilliant, but just occasionally we’ve all had example of it throws something really random out, doesn’t it?
And you need to be able to spot those things, not just assume they’re right and be able to challenge in the right ways back to, I dunno, pricing or systems people. So that will be great. I think from a development point of view, that’s where we have to start thinking differently. So clearly being able to understand insurance, understand technical terms, wordings the legal implications will still be a big part of what people need to learn. But we will have to recognize that there will be a whole range of people who will spend or understand data in different ways. So we’ll have to bring in layers of understanding around how to interpret data and how to think about it in the right ways to make a difference. Because I mean, I hear my kids talk about stuff they do at college and at uni and it’s very different to what I would’ve done, but they’re learning how to use it.
If you’re my age or even in between those two ranges, you are going to be in a space where there’s going to be a level of, some people are going to be amazing at it and some people are not going to be so brilliant at it and we’re going to have to help them all through it. So for me, that’s the big thing about ai. It’s just because it makes sense to people who create the tools and the automation and use it really well. We can’t assume the people who are receiving that data that whatever it happens to be, are able to use it in the right way. And the danger of course is that you don’t understand it and you come to the wrong decision. So you might turn risks away, you could quote or you might make a group of risks appear to be uninsurable because of the way the data is extracting it. And I dunno if there’s any 10 types of risk or 10 risks in a certain trade and they all become no quotes, then we start to create problems in our industry, which we don’t intend to do. So we’ve got to figure out how, make sure we kind of work through those gaps. Apps.
CW: Yeah, I like,
MH: Sorry. And I think the Chartered Insurance Institute has a role to play in that space because it has a really generic role over the industry. And actually we just did an article on AI on that basis, which we said we need to start introducing this, the thinking and thought around AI because it’s going to become more and more underwriters of course are going to be a bit fearful. They think something’s going to come and take their job. I just think their job will be much more fun actually in the future.
CW: Yeah, I agree with you. I really like the answer. I don’t want to drill in just a little bit. I’ve had a number of people ask me in various contexts what generative AI is going to do to their jobs or their industries. And I think where I’ve settled in my answer is that it’s going to take the distribution of capabilities and bifurcate it so that people who are already pretty good get better and faster and the people that aren’t doing so great, they start to do not so great things even faster and there are more ways for them to screw up. So you get these two peaks where you have really high performers and really low performers starting to separate. And in that context, I used to be a university professor and having a classroom of really high performers and really low performers simultaneously is an incredible challenge. As you think about educating people and bringing that whole cohort up the curve, what do you think are some of the salient aspects of the training and the content that you need to create?
MH: So I think first of all, one of the things I think that’s going to be really interesting is today as an underwriter or somebody overseeing underwriting, we need to do a lot of manual checks. We might call them self validations or audits to check that people are following guidelines. The more you automate, the more the system’s going to start drawing those things out earlier. So I think to your point, will there be a real separation? There may be, but I think we might get to some people that maybe are not interpreting things in the right way or have missed a guideline or something because the system will help us see those things quicker than waiting to do the audit every 3, 6, 7 weeks or years for some businesses. So I think in a way what we might start to see is some things flashed up that will make us go, oh, that wasn’t what we anticipated and that will then help us work in a really good way.
And I know in my previous employee, part of their validation process was absolutely to make sure there was a conversation with the underwriter if they found something because that’s the way you help people improve what they’re doing and we need to do that. And even just because I had a title that said to you in it doesn’t mean to say I know everything about underwriting. It really doesn’t. What I want is people that can come and talk to me about risks and help me understand why they think they should write it. And what we will have is data and AI that will help us do that much more quickly. And then therefore customers and brokers will feel much more confident in the timeframes that they’ll get to get information back. I sort of agree with you, but I also think there might be an opportunity for us to see things quicker, get on top of things quicker and help underwriters and traders to really develop their skills in areas that maybe we wouldn’t have seen before.
CW: By the way, you’re under no obligation to agree with anything that I say, push back as much as you want. Yeah, no, I think you’re right. The framing I set up is more you have a tool that everyone uses on their own and what you’re talking about is having a copilot that everyone sort of works with in the same way.
MH: Helpful. Yeah, I think it’s really hard when you are in my kind of role to be able to see everything that an underwriter is doing and you’re then relying on stuff being provided to you through audits and checks and all those kind of things. And I just hope that we start to see some of those things playing through in the automation that just help us figure that stuff out quicker because actually I haven’t come across an underwriter yet that when they’ve made mistake goes, oh yeah, I meant to do that. They always have. It is never been an intentional thing that people have missed a piece of guideline or Mr Zero. It’s always been just one of those things. So everyone wants to do a good job and on the basis that everyone wants to do a good job, we should all be open to receiving information that helps us be better and do it quicker and I guess more fully to help us get the results we’re looking for.
It’s going to be exciting for sure. Sorry.
MG: No, no, no. I’m going to jump to a different topic now. One that I’m personally excited to chat with you with just given your background and it’s an area that we’ve wanted to dig into in previous episodes and just never had the right background or combination of roles and experiences to talk about is how does the regulatory environment potentially shape decisions that an insurance carrier or a chief underwriting officer an underwriter would make as they’re thinking about a variety of things? There’s obviously the product development that a product owner or a product manager would do before they even get to developing the underwriting guidelines and decision-making. But then the types of data that you want to collect that you need to collect to do your pricing algorithms, right, let alone the data you need to collect from the submission to actually make the decision and then how that all ties together for claims. So when we think about here in the states specifically, you have, I think it’s 51 different jurisdictions, all different regulatory requirements, all different, I’ll call ’em, states that you need to file products with and obviously there’s differences in what data you can use, the pricing associated with that. In your experience, can you just walk us through how the regulatory environment maybe shapes strategy as it relates to product development and underwriting decision making?
MH: Yep, absolutely. So the FCA in the UK regulates from I guess a conduct point of view. The PRA does the financial elements, but what they have done over the last couple of years is introduce, I guess a two phased approach to how we manage products from an insurance point of view and from a broker’s point of view. And what we’re ultimately, or what they are ultimately trying to do is ensure that the products that are provided to consumers and customers have fair value. So IE and insurer, in a funny way, it’s a bit of an odd one, but there are claims that mean the product meets the requirements of what somebody wants. In a simple way, if you buy a product that covers your mobile phone but it doesn’t cover accidental damage or theft, it’s probably not meeting the requirements of somebody that owns a mobile phone.
So that’s what they’re trying to get to. Have. We got products that meet the intentions that we laid out. And so they’ve done that in two ways. So initially they launched the product governance rules, which was said you had to every year review each of your products around fair value and there’s some guidance around the types of questions and things you needed. And then more laterally they have introduced the consumer duty regulations, which widens out beyond the governance stuff into things like checking and the standards that you have for documentation and consumer communications so that people understand that there’s real clarity in the documentation people receive. So those things have been, they’re hard, there’s no doubt about it. And what that also has introduced is both parties, so both insurers and brokers needing to provide information on the products. Now that’s all great and it’s very well intentioned and is doing what I think broadly doing what they intended.
I think the challenge that people have is of course insurance has been going for many, many hundreds of years and we’ve got products that will be quite old and what we set when we set them up and the data that you might extract from systems will be very different to maybe what those requirements say today. So legacy systems, legacy products will be harder to extract that information from. That may be if you’re an insure tech or a new startup or you write a new product today where of course you’re going to start to think about the kind of questions the regulators are asking. And I think that’s the rub at the minute, which is how do you get the information out systems that may be your two or 300 years old, an insurer, you’ve had products that have been going for 50 or 60 years, how do you figure out how you get that data out from old blue screens?
And every insurer will be figuring their way through it. And certainly there will be lots of people looking about how you can do it. And I absolutely know we’re finding ways to get stuff out of it. It’s just not what was intended when people wrote those systems 50, 60 years ago. So the regulation is doing what it needs to do, but it has brought questions around how you get the data out systems and therefore when you get to new products for the future, you’ll have those kind of things in mind. And I think at a very simple level, how you might even articulate a product in a system might be different. So we might, I dunno, put everything under motor and yet there are different types of motor products, fleets, single vehicle, commercial vehicle, all of those things. So what would people have to start to do is think about the lowest level of splitting data down, which they may not have done in the past because there would’ve been a load of manual interventions to strip stuff out.
And I think we’ll start to think differently about wordings to make sure they’re clearer for people to understand. And certainly the covid bi claims in the UK will have driven a lot of that. And I was part of the working party for the CI around what the reading age should be and it’s like 12 and we’ll have had wordings that will be much more complex than that. And probably some in some areas will remain complex because they’re big complex risks. But for consumers they need to be more simple. We need to send documentation, we need to test it, we need to make sure that what we intend to say is what people receive and bring that consistency. And then, yeah, we’ve got to set new products up that have fair value in mind. So we are not giving in that very cheesy example, a mobile phone cover that doesn’t cover accidental damage or theft.
And again, people won’t have set them up with the intention of not giving fair value, but of course over time things evolve, life changes and we’ll see different things. So yeah, the regulatory environment is big in the UK and I think rightly so, we need to make sure that consumers can have confidence in what they’re buying and how it works. Those regulations don’t just apply to insurance, they do apply in all sorts of places. So if you buy a car and you get some insurances off the back of it, it’ll apply to all of those people. And for brokers, they have to provide data back to every insurer they deal with on the kind of commissions fees and things like that. But they’re providing because there’s been a massive amount of work for brokers too, and ultimately that should give customers the public much more confidence in the products that people are providing to them, which can only be a good thing for the industry.
CW: Absolutely. So you talked about the challenges. Are you seeing these challenges creating opportunities for, is it whether it’s new technologies, automation, smart automation with AI built in, how is the tech industry coming alongside these challenges and helping?
MH: So I think we’re definitely seeing a bit of an uptick in people with opportunity for data, ai, ingestion techniques, all of those things because they definitely can see the need for that. I think we’ve also got companies trying to do it for themselves. They’ll have tools that they’re using and they’ll be trying to figure out how they can do that maybe without the expense of a third party because that for some people will be a challenge, whatever happens and whoever gets the business. Ultimately being able to figure out how you can extract stuff out of any system, legacy system is going to be really important and how you can produce that into some standard formats and help. We’ve talked about underwriters before, but if you look at governance managers who need to be able to interpret data in a slightly different way, you talked about claims earlier, how many claims came in, how many claims were paid, how many claims were declined because the cover wasn’t appropriate, how many complaints did you get?
Were the complaints relevant to a particular product or schemes because we might have logged them under property again. And of course there could be lots of things. So I think the challenge for third parties is being able to pick which parts of that they can help with. And I think there are two types at the minute. I think there’s people that, again, we can help you get stuff out and I think there are other people saying there’s all these generic data sets in the world, how can we help you use these in your jobs and build it into your processes so that you can in the UK check company’s house really easily and make sure that there’s not a business that says, I don’t know, I’m a dentist, which would be insurable. And then there might be something that we will also make asbestos because they’ve declared it to come, companies have, but they haven’t stuck it on the proposal form for insurance. Now that’s a really extreme example, but you get the gist that you are relying on people disclosing what they do and those things like those checks will help us make sure that we’re assessing the risks we think we’re underwriting. So I think there’s the two types of data techie companies that we’re starting to see come in, how do we get stuff out and these are the data sets and how you could use them in the future
CW: In terms of the insurers or the brokers, whichever side of it you want to talk about. You talked about three areas, sort of like data and legacy systems, data and analytics, and then ingestion, which sort of feeds the whole pipeline. Where are you seeing companies focus right now in terms of priority?
MH: Probably in ingestion. I think if you were to measure the process from the minute you receive a broker’s submission in your inbox or into a central place through to when you quote possibly bind and win and bind, there’s a lot of time taken up and getting things onto systems, reading documentation, broker presentations. I’m not sure what they’re like in the us, but they can be 30 pages. You’re trying to find the nuggets of information in there and work out what you need to put in. So I think proportionately, if a case took a day, my guess is two or three hours of that day is just trying to figure that stuff out, log it in the way that every insurer wants it so it can be found allocated to an underwriter, pull out that information and get probably a shell of an underwriting pricing engine pulled together for somebody to start to look at.
So I think from what I’ve seen, there’s a lot of focus on ingestion in a lot of insurance companies. There’s a lot of people talking about it from a third party supplier point of view because if you can half that amount of time, could you take all of that and make it really easy? Then of course you can do two quotes a day, not one quote a day. That means you have more opportunity to win. That means you can be more successful. So definitely seen a lot more focus. And from people I’ve spoken to, ingestion is a really big part of what people are trying to manage at the moment because that will just help everybody. That’s the efficiency angle people are looking for.
CW: And it has knock on effects too, right? You’re more efficient, but you’ve also sort of memorialized your process by automating it and building technologies around it so you get cleaner data, which means better analytics. It’s
MH: Completely, I mean I think about when I started, and I’ve quoted this when I was at RA, I remember starting on a system and somebody said to me, you put this code in when you want to cancel a policy. So I just put the code in right now, 15 years later I discover that that code means it’s gone to a different competitor. I won’t quote one, but it’s gone to a named competitor. Whereas a different code might’ve been a different competitor or one that the business went out of. They stopped trading. Somebody just told me to use a number. So every risk I canceled had the same code because I didn’t know any different. And that’s the challenge with data and systems, isn’t it? You are as good as the person that trained you. So unless you probably are trained by the person who wrote the system, you are as good as the person who kind of found their own little ways through.
And automation from an ingestion point of view, we’ll take some of that out because we’ll bring that consistency in which like you say, will make better data. Because to my earlier point about education, I didn’t think about the consequence of my code. I was told I needed to cancel the policy, so I canceled the policy. I wasn’t thinking about what that meant from a claims point of view or a funding point of view where the premium went. I wasn’t thinking about what claims would do as a consequence of it. Those things are not going through my head because we weren’t telling people that whole journey and that’s not a problem that wasn’t happening. But I think this is the opportunity as you start to, when you educate, you start to get people to think, oh, when I do that, where does it end up when I do this? Where does it end up? It’s a bit like baking a cake. We were trying to get a cake at the end, but we didn’t know what ingredients we were doing. We didn’t know what to put in and then we were adding something extra. We really, and that’s no way to bake a cake and insurance is the same, right? We need to start to think about making sure people understand what they’re doing, why they’re putting it in, and the consequence of it so that actually everybody is better for it.
MG: I’m
CW: Imagining a cooking, sorry, go ahead. No,
MG: I love that example. As someone that worked at insurance carrier where also had to just type in codes into a green screen and hopefully I ended, my role ended up being to train people. So hopefully I was good enough for my job to train ’em to do it the right way. But there were a few stops and starts. Similar thing, you have to put in this producer code or this is for that line of business and this is how that form is indicative of that. And it is a lot of the manual audits done every six weeks or so to let you know of 2% of the work you did. Yeah, absolutely. Right or wrong, right? Absolutely.
MH: And that’s why you see people focusing on ES because that’s really important that we get to that. And then you’ll have people with rating engines that will guide you through some of they’ve existed for a long time about how you get to the price, but there’s real opportunity I think for people that want to look at ingestion and help brokers who want to run stuff for themselves as well, as well as insurers about making that easier. And there’ll be a group of underwriters across the UK going, hallelujah, thank goodness we’ve got something coming that will make our days easier because nobody likes admin. You don’t like your own admin at home when you have to find all your pieces of paper or catch up on your emails, it becomes, it’s just hard work. Everyone wants to do the exciting part, which is get a deal done, get something over the line, feel like you’ve helped a customer find a solution to a problem, but the admin side is the bit that you have to do. So if you can make that easier for people, that’d be brilliant.
CW: I’m also now imagining a cooking competition show where you swap chefs at every step. I think that would be a really good show. Yeah,
MH: That’d be a great show. Figure it out. What did I put in? I love an analogy about life. I’m afraid. It just makes me, makes my head work. I’m a bit odd, maybe abstract, somebody decided to tell me I’m abstract in my thinking, but it’s just my way of figuring out how the world works. I think,
MG: Well, it wouldn’t be an episode if Chris didn’t plan for a future job in a different industry somewhere, right? For this episode, Mandy, you touched on something in that example that you gave that has been, was sitting in the back of my mind as I’ve heard you answering some of our questions. And it’s a little bit different than the ingestion of data question, but specifically it’s on leveraging data and analytics for what is happening inside the carrier, right? So to your example of you unknowingly putting in a single cancellation code may have downstream led to decision-making in terms of pricing strategy or competitive analysis that without making it sound like catastrophic, may or may not have led to a strategy that was not actually solving the key problem as to why things were getting canceled or policies were moving over to different carriers. And I think a lot about the data loop of this was the underwriting decision that was made, this is what led to the claim.
How do we cycle that back to enhance the underwriting decision making? But all of that ties to product development, product pricing, strategy, market growth and strategy. How do you think about, especially in your experience of having essentially run a small carrier, how do you think about all of those things tying together and the value of having structured data or understanding of where the data can be found to then lift out some of those things that are not external data sets but that are truly internal data that’s been developed just by the actions in the day to day?
MH: Yeah, think one of the, it’s interesting. So structured data exists, I think in a funny way, structured data does exist. Legacy systems have load to structured data. The problem is not the data we necessarily thought we would need, therefore it’s the unstructured data we’ve got in all other places that we’re obviously trying to capture and figure out what we’re going to do with. So I think you’ve got a couple of things and it depends which part of the business you work in as to what you want to do with data. If you’re in a head office underwriting function, you are thinking about what you’ve just described. What kind of strategy do I want? How am I going to price these risks? Do I need to exit? Do I need to put rate through? Can I afford to let rate go actually because I’m making, it’s working well and I can do something different.
So you’ve got all of those things. If you are in the operational teams, you are probably thinking about how many quotes I’ve got waiting to be sent out the door, how many renewals I’ve got to do. So I thinking an insurance company, wherever you are, you’ve got a demand for data of different sorts. I think the challenge is to work out what’s the right data for everybody and make sure you build products that answer those questions as generically as you can. And certainly one of the challenges I think is for people to always make sure if I’m setting a product up that I ask claims to come and talk to me or I talk to the operational teams and say, what do you need to make this work? Because if I just create what I want from an underwriting point of view, there is a good chance other people aren’t going to get what they wanted out of it.
So when you want date leverage stuff, you’ve obviously got to collaborate really well to make sure those things come together. And I think people are getting really good at trying to do that today. I think that understanding of how everyone joins up is really important. So I think from when I was in the Channel Islands, I wanted all sorts of data from wanted claims data, I wanted underwriting data, I wanted service data, I wanted debt data, how many customers were not paying. So I think from an insurance company point of view, there is so much demand on data and if you are running a data and analytics team, you absolutely need to start to think about how you don’t just maybe go down a path of going, oh yeah, I’ll give you what you want and then I’ll give you what you want. And stepping back occasionally and going, that’s what everybody wants. How do I make sure I get the right, I focus my energy and my team’s time on getting the consistent stuff that will make the biggest amount of difference. And data’s a word, right? That’s all it is, is a word for collection of numbers or pieces of information. The important thing is what you do when you’ve got it. And I think the doing part of data is the stuff we have to spend our time on.
CW: Absolutely. And part of that doing is grooming your data, which people forget about all the time. Yeah. One question, one thread I wanted to pull there. You talk a lot about the benefits of having different teams, different parts of the process, collaborating and working together, thinking more holistically, what do you see as the risks that come along with that? Are there any downsides
MH: Are probably one of those people that tries to look on the bright side of life. And so yes, because you’re slower collaboration I think has a chance of meaning that you’re not as quick as you could be. So if Mandy wants some stuff, Mandy goes and does, it gets done. And if I’ve got to bring other people in, I’ve got to start to think about, oh, have you got some time? Can I explain this to you? And they’ll, oh, hang on, I need to talk to somebody else. So you can slow the process down with collaboration. I think what’s really important is having that kind of early conversation about this is what we want to do, who do we need and can we make sure we’ve got decision makers in the room because then we can be faster. And I think that’s the biggest risk or downside to collaboration and making sure everyone’s getting what they want.
I think the bit you have to then think about is in two years time when you’ve got to do a product governance review and you did your product, but you forgot to ask claims what data they wanted and then they can’t give you the data you need to do a review, you then stop and go, oh, maybe I should have done that. So I think we have to think what we need today and what will the future look like as best we can view it and make sure that you bring those people in that are going to help you. So yeah, I think speed is the issue, but I think there are some really challenging things in the future if you don’t do those things properly. I don’t think governance is going to go away. Regulation is going to be there to
CW: Stop.
MG: Dare I say, too many cooks in the kitchen trying to bake that cake
MH: That’s why there’s loads of different chefs in a big restaurant isn’t there because they’ve all got their arms to play. But if they don’t all come together and kind of collaborate, then somebody doesn’t get their dinner is how it all works.
CW: Cake is frosted with ketchup. It’s terrible. Absolutely. That answer Mandy sets up my absolute favorite question, which is what do you see, we talked a little bit about the last 10 years, what do you see coming in the next, I know at this 0.2 years seems like an infinite amount of time, but pick your timeframe. Where are we headed?
MH: I think that’s a really difficult question because I think we are moving at such pace. I wouldn’t want to go too far ahead because I don’t know that we even know what the future really looks like. If you look at ai, we kind of know it can be really useful, but we don’t really yet know from an insurance point of view exactly how useful and where it’s going to play. So I think what we can say safely is the next two years are going to be exciting as people figure it out. I think there’ll be a point where we might plateau for a little while. We don’t end up with risks that are uninsurable because we’ve done so much that there is this small group that can’t get insurance. Not because they’re uninsurable, but because the data tells ’em something or the systems do. I think what we have to do in that time is think really carefully about the type of people that we might bring into the industry.
Previously we might have thought about people who want a detailed technical career maybe, but actually we might need people that are really good at customer communications and being able to understand things. We’re going to need some data analysts. That’s a different skill. We’re going to need people that can help us with automation and systems. So we’re going to need some different people in all of our businesses. So I think over the next couple of years, what you’re going to start to see is people really expressing themselves in AI and figuring out what it looks like. I think we’ll start to see more and more use of data and probably a lot more ingestion. I think we’ll start to see providers and carriers working out how they can make that better. And then there’ll be this race to the who could get to brokers quickly and do things more efficiently for brokers.
And I think we’ll start to see people think about how they recruit, who they recruit and how you manage that in a hybrid working environment as well. Because you’re not learning by osmosis anymore. You’re not standing at the coffee machine or overhearing a conversation. It’s me in a room with no one else trying to figure out all of these things. And I mean that both ways actually. So if you are a junior underwriter, you’ll learn from a senior underwriter, but a senior underwriter will learn from a junior underwriter how data works, how systems work, how to use the internet, how to find things differently. And so it’s not a one way street in terms of just because you don’t maybe understand insurance at the beginning of your career. It’s because you can teach people other things and I think that’s the danger in a hybrid environment. So I think that’s the kind of stuff we’ll start to think about in the next couple of years. It’d be interesting for sure.
CW: Yeah, I want to follow that up with two things. One, going back to your answer about we’re sort of in this unstable period where we don’t know where we’re headed. I have seen repeatedly the last six months in talking to insurers where gen AI has appeared and they’ve forgotten that there are already great technology solutions to the problems they’re trying to solve. And so it just blows up budgets, it blows up projects, it blows up vendor relationships. So I too am hopeful that we get out of this phase of our relationship with gen AI quickly. And so that triggers a question in my mind, which is again, thinking about the future, where do you hope we are in the near term?
MH: So I really hope we figure out ingestion. I think we’ve done that quite well with claims. So I think if you’re in the claims world, your process from a, we call it fol first notification of loss I think in most areas is quite streamlined. But I don’t think an underwriter would say necessarily the same thing from a first receipt of inquiry through to Quo. I don’t even from an underwriting point of view, we’d feel the same. So I think to ensure we don’t lose people from our industry into others that are running in efficiency better, I think we need to really work on that efficiency ingestion bit. And that is my hope because the people who’ve become underwriters, I’m sure every insurance company is quite similar. They have a licensing scale, I dunno a level one is you arrive on day one and level five or something when you become the expert, you want as many people to be fives in my head as you possibly can have.
That will be quicker decision making, more consistency in your underwriting, better knowledge. You’ll write more risks, you’ll have them at the right price, you’ll get the terms and conditions without all the referrals and those kinds of things that go on. So if you get ingestion right, you kind of build that opportunity for people to develop their careers really quickly. And that can only be a good thing from an insurance point of view, from an insurance carrier’s point of view. And I think beyond that, I think we have so many things that happen in our industry at the moment. We are managing inflation regulations, regulation’s a big thing. Data’s a big thing. Efficiency is a big thing. Climate change is a big thing. There are so many things and sometimes we don’t know in our industry what the big thing will be, but we all know there’ll be a big thing. It was covid a couple of years ago and therefore we need to create space that we’re ready for the next thing even when we don’t know what it is. And so I think if we do some of that stuff on efficiency, we start to give ourselves space to think, not just do. And it’s the thinking that will give us the opportunity for the future as well.
MG: I think that’s really interesting. Maybe that point you just made because when I think about what was the next big thing a couple years ago, and I was at a carrier that was launching a brand new cyber product at this time was cyber insurance. And I remember as we were developing the product, but also the application, it was what are other carriers asking for? What types of information do they think is key to understanding this risk, which is brand new? What are the different coverages that they offer that we offer? What questions are tied to which type of coverage? And there are a lot of, I don’t want to say mistakes, but there are a lot of false assumptions of what was important to capture and what would actually drive performance and keep loss ratio down. And there’s also the challenge of extracting that information.
Where do you keep that? And so the design of systems being flexible enough to take in different data points because after that cyber product launched, I remember we had iterations two, three, and four that we were refiling with states because of the learnings we had really quickly. And so I think that point is just spot on of you need to be setting up your systems, your people to understand nuance and how things need to change and shift in order to keep that performance, that excellence to a certain level as the macro economy and things like that shift, those different landscapes shift.
MH: And I agree and I think if you get to an individual level of underwriting in that for an underwriter, sometimes you think you need so much information to be able to make a confident decision. And somebody isn’t going to say, why did you do that if you have a claim or something? Or even just a query that’s kind of got people confused or whatever. So I think what tends to happen as an underwriter is you go, well, I’ll just ask that extra bit of information or if I actually know the colors of the doors or you kind of keep going and going and going. And of course brokers get a bit frustrated because they’re trying to figure out why do you keep asking all these odd pieces of information? And it’s because we are just trying to give ourselves confidence that we can write a note, keep a file that demonstrates we knew what was going on and our job, I think from a central point of view is to make sure that people have the confidence to know the right information to your point that they should be asking for.
Not just keep asking and asking and asking so that actually they can have the confidence in the decisions they’re making. And then that gives people like me in a CO role have confidence that we’re making the consistent right decisions. That means we can have confidence in the pricing, we have confidence in the funding and therefore we can continue to invest. I just think it’s natural when you are unsure of things, you just keep asking and you hear it with your children, you hear it with people when they’re doing a job at home, should I do this? Do I put that mail in? Underwriting is exactly the same kind of characteristics come out. I just need to make sure I know exactly what I’m underwriting here. Is it really the same as the one I just did or is it different? And I think we have to help people try to do that as simply as we can.
CW: Yeah. Well my head is absolutely full right now, so I’m going to call this one. You’ve been listening to another episode of Unstructured Unlocked. Our guest today who is full of a wealth of information and excellent analogies for how underwriting and insurance work has been Mandy Hunt. She is the chairperson of the Underwriting community board at the Chartered Insurance Institute. Mandy, thank you so much.
MH: Been a pleasure. Great to talk to you.
MG: You too, Mandy. Thank you.
CW: Thank you for joining us for this episode of Unstructured Unlocked. You could find all of our episodes wherever you listen to podcasts today. Be sure to write a review if you like what you hear.
Check out the full Unstructured Unlocked podcast on your favorite platform, including:
