The case for change: Transforming underwriting for tomorrow
Successful pricing and underwriting transformation requires a clear vision and the momentum to see it to fruition. TEC and EY's James Anderson dig into why its needed and how to achieve it.
Pricing transformation means different things to different businesses, but there is no shortage of efforts to bring it about. It goes hand-in-hand with underwriting transformation – both pricing actuaries and underwriters are fundamentally contributing to the same outcome: underwriting risk effectively.
How can companies build better underwriting workflows and make the most of their data assets to make the best risk decisions and achieve the strongest combined ratio?
When the market softens, the insurers that have invested in implementing a strong pricing and underwriting framework will be the ones to come out on top. From "augmented underwriting" to systematic data capture, and enhanced portfolio strategy – there are powerful levers insurers can pull, with the right leadership and vision.
In this episode of TEC Talks, TEC chats to James Anderson, Partner at EY, about the case for transformation, the opportunities to play for, and common challenges along the way.
Never miss an episode
Subscribe on your favourite podcast platform:
"A lot of companies lost a lot of money in the specialty market over the last few years, and I think have recognized that a lack of pricing framework, a lack of underwriting steering, has hampered them."
Full episode transcript
[00:00:00] TEC: Welcome to TEC Talks. I'm Tom Chamberlain, otherwise known as TEC. That's pretty much all you need to know about me. In this episode, we'll be discussing pricing transformation, and I'm delighted to be joined by a veteran of the insurance industry, James Anderson. Hi James.
[00:00:26] James: Hi Tom. I'm not sure about veteran, but…
[00:00:29] TEC: Well, I was, I was gonna do a little bit more of an intro for you. So, James has worked at some of the largest insurance reinsurance companies in the world, Britts, Allianz, Swiss Re, and currently a partner at EY. So I think that pretty much counts as a veteran. And, just as a side note, also a pretty decent poker for my understanding.
[00:00:46] James: Very kind. Very kind. Tom's won a lot of money from me.
[00:00:51] TEC: Only, only a little bit. So James, yes. Thanks for joining. Looking forward to the chat. Yeah. Pricing transformation. So before we kind of get into the sort of the whats and the whys, perhaps we can just, take a little time just to define what it means. A few different people will see things slightly differently.
So how would you describe, how, how do you see pricing transformation?
[00:01:11] James: Yeah, that’s a good description already, right? That a lot of people see it in different ways. I don't think there's a definition of pricing transformation, but I guess for me, right, it's kind of, it's something about, its. Something you can't just do as your business as usual, incremental improvement. It's about having a vision of where you want to go to and then kind of a step change towards that vision. So it's kind of, it can be on a lot of different things, right? It could be your modeling capability, your deployment capability, how you engage with your underwriters, you know, how you steer your portfolio.
It could be lots of different things all under the banner of pricing, and you probably don't wanna do all of them all at once. But it is this concept of let's take a big step in the direction of your future vision.
[00:01:54] TEC: And I mean, you kind of mentioned a few words in there. Which sort of resonated with me. It's, it's, it does seem to be wider than just pricing, right? So, so this is not just an actuarial exercise. I, I think this very much encompasses, particularly in the specialty market, which I think we're kind of focusing on here, very much the underwriting side as well.
And I think pricing and underwriting transformation sort of go hand in hand. I, I guess you'd agree with that.
[00:02:18] James: Yeah, completely. I think it's one of those things, if you do a pricing transformation purely focused on what the pricing actuaries need, you know that's probably gonna fail to some degree in that, you know, huge amount of the benefit you're gonna get from any pricing transformation should be. How do you underwrite the risks better?
How do you make best decisions? How do you improve your loss ratio?
[00:02:39] TEC: Absolutely, and I think, we'll, we'll kind of get onto that, which is the sort of the why of the change and, the, the big benefits that you can get out of it. I think that there's also a part, and you might have just briefly mentioned it as well, it's, it's also just not sort of pushing in technology and, and sort of hoping for the best.
There's that, there's more nuance than that. So we can get onto the sort of the more cultural shift and, and that side of things. But I think, you also say it's, it's about setting the vision. Right. It's, it's about understanding where you're going.
[00:03:06] James: Yeah, and that's, I mean, you, you mentioned, I, I work, work at EY and I joined there nine months ago and that's actually been, One of the really interesting things for me is seeing how all the different pricing teams and all the different companies work and how varied it is and you know, that vision of pricing differs a lot and I think sometimes is a sort of legacy of who set that team up originally or you know, who, who, what the personal preferences of one or two people are. So really taking that step back and saying, How will pricing add the most value to the company? How does pricing fit into the kind of philosophy of the company? You know, is it a very traditional underwriting face-to-face relationship sort of company? Is it a data-driven, technology-driven company? 'cause that's gonna have quite different needs from a pricing function. And you really need to be clear on your vision before you start doing things like technology and actually starting to execute. Because I think otherwise you. Run the real risk of being disappointed by the end result.
[00:04:07] TEC: Absolutely. Projects, projects like that are doomed to fail. I, I think that's sort, sort of any, any transformation type of project, being in pricing or across insurance, that is absolutely critical to know where you're going and why you're doing that as opposed to just, diving headlong into new technology that's, that's go and implement that.
It's quite interesting so that you talk about the sort of legacy side and, and what's been happening. I mean, we, we analyzed 13 of the top insurance and reinsurance companies and over the last five years and. Sort of pulled out from reports and stuff around what, what were their main priorities as a business and very much we've seen a shift over the last five years.
Previously, claims transformation and other areas operations as well have been very much at the top priority list. But actually more recently, over the last five years, pricing has sort of come more to the forefront. So what's been really happening over the market in those, in the last sort of 10 years and, and what's changing now that that pricing has really coming to the forefront.
[00:05:04] James: Yeah, and firstly just say that totally resonates to me as well in terms of both my previous experience of being a pricey act in industry and the change over the last 10 years and now in consultancy seeing that. But to answer your question, a lot of the kind of pain points and reasons why you might want to change have been true for all of those 10 years.
You know that your models, a lot of the pricing models are sitting in Excel, which has a lot of great features, but also has a lot, brings a lot of challenges. If you're trying to be a, sort of, have a systematic pricing process, you know, there's not enough data capture. You are asking your underwriters and your actuaries to do a lot of manual, not particularly value added work. it often leads to a lack of. Underwriter buy-in. You're not doing that steering side of things and you're not capturing the data to then be able to build better models in the future. And it you, you can end up in this sort of sad state of you spend. You know, a company invests a lot of money in pricing actuaries and pricing tools and the underwriters time on it, but actually the underwriters don't buy into it. The pricing tools are filled in after the risks written with not much care. Whatever data's captured isn't of high quality 'cause no one's putting much effort into it. And you kind of see where's the value being added. So there's definitely been a need to change. Sorry, I'm not saying that's what every company at all.
Uh, I think that's kind of a worst case. Catastrophizing, but you know, there's been these drivers of change for a long time. So to come to your question of, I guess why now and why over the last few years, and there are a few. This is personal opinion, but there are a few things that I see driving it. One is kind of the market cycle in terms of a lot of companies lost a lot of money in the specialty market over the last few years, and I think have recognized that a lack of pricing framework, like a lack of underwriting steering has hampered them. And it's one of the things that they, they're now trying to invest in while the market's hard, while they've got kind of a good combined ratio so that they're in a much stronger position. Once this market does start to soften, EY does quite a lot of sort of analysis on specialty results and from that tries to kinda tease out drivers of performance.
And the two that always seem to come out is one is just simply good underwriting at a line of business level. And the second is strong management of the cycle in terms of, you know, not, not growing aggressively when the market's soft and then growing aggressively when it's hard, rather than remediating, you know, all of those things.
Just talk to pricing. You know, you need to, to do strong individual underwriting, you need to have good tools and diagnostics and support to the underwriter to manage your cycle, you need a strong underwriting framework. And so that's one key driver for me. Another one for me is sort of around in general, increased recognition of the importance of data. You know, that there was maybe a bit of a boom in kind of, AI and ml and. People investing heavily in sort of data science teams a few years ago, which I think hit a bit of a, of a block or a hit of a road, a road bump, in that they realized that they've brought in all these smart people, strong academic backgrounds, but they didn't have any good data for them to work on.
And they weren't necessarily, they hadn't necessarily hired people who were good at data. And I think there's been a reset from that and a recognition that actually. A huge USP and a huge piece of the IP of an insurance company is their internal data, and therefore capturing it effectively, structuring it, thinking through the whole end-to-end flow. And from that, being able to do things like build better pricing models, build better steering frameworks has become really important, and therefore, moving onto a systematic. Way of capturing your data throughout the pricing process is I think another driver of change. I guess I'd say the, the last one I'll just call out is in general, it feels like there's a transformation happening right in the insurance market or about to happen in terms of digitalization that, you know, this sort of insane situation where a broker's got all the data on their, in a, on their database, they print it out. If you're lucky to something machine readable, email it to the insurer, who then pays someone else to key it back into their systems. I mean, I remember. 15, 20 years ago doing individual account pricing where? Where I'd got a scanned version of a print off of an Excel spreadsheet and I was like, somewhere there is a spreadsheet that they could have just given me, but I'm gonna type it in. So I think the move from that to actually that, that's insane. There's going to be much more of a move to sort of algorithmic underwriting, even if the underwriter's still there, but that the data will flow automatically. You need to have underwriting workflow, you need to have pricing tools that are digital to all link together. And then you need to have much more confidence in your models and an understanding of your models to be able to decide where you're gonna do things algorithmically. Where are you going to, refer up to an underwriter? And, you know, all of these things are, they're not just pricing topics, but pricing is a sort of enabler and building block of them.
[00:10:08] TEC: I agree with, with all of those. Actually,I'll just quickly make some points, going backwards across them, so your points around things still being a very manual process, you know, it's, it's probably. Even worse than that, right? COVID was supposed to have sorted out all of this sort of paper trails and everything had moved online, but brokers are still wondering about the market with slip cases, with things on bits of paper.
So the fact that you get sometimes an email with a PDF can, can actually be quite good. You know, it's, it, we, we are still, we're still in that industry where, where, where that happens. But, you're right. Where, why on earth should there be PDFs of Excel documents when that could just be sort of sent across?
You know, it, it, it wasn't that long ago that. I was watching part of the underwriting process and someone had, copied down, some information from a screen onto a bit of paper and then walked it over to operations who then keyed it into another system. And it, it's not so many years ago, and, and it's quite, it's quite frightening that, you know, we've been talking about automation, right?
Which is a part of the sort of pricing transformation. I mean, you and I have been talking about this. Since, I think, since we started our careers, and we're still talking about it now. So it's, it's fascinating that we all know we needed this transformation, and yet it, yet it hadn't happened. And you're right.
I think with, with the data side as well, it's, it's more critical than ever that, that this happened. You know, we, we've, the insurance industry has been using data for, Lloyd's has been using it for hundreds of years. Right? So a stories of the shipping, you know, that, that they relied on, on sort of face-to-face tales of what was happening to, to give them the data.
Uh, obviously we have a lot more and a lot better data than that. And I think this explosion of data, the a hundred or so zettabytes of, data that's consumed every year now, that's like, that's trillion gigabytes. So pretty, pretty high number. We, we have vast amounts of data and, and it, it is no longer feasible just to be capturing.
These sort of bits and, and forms and, and sending them in, you know, you need to have better ways of ingesting it all. And again, this is all part of that pricing transformation. So I can definitely see why now is more of a time, the, the first point around the sort of state of the market as well. We, we are for the first time in a long time in, in a, in a very hard market.
And at this point actually companies are making. Some money and, and doing quite well, take away some hurricanes and covid and and Ukraine, Russia wars and stuff. But the, the, the market is, is, is as good it is as it has been. And yet you'd have thought sort of as the market was softer in the past, that was the time when they really needed that technology.
But I guess, I mean that this is probably that the reason now why it's potentially changing is that there is investment to be had, you know, interest rates are. R n a much higher state than they were, which means companies do have a bit more money to invest and, and spend on this sort of technology. So perhaps that is also, a bit of the reason.
And yeah, doing this transformation now before the market sort of turns again, when you are really going to need that, that extra margin is, is kind of critical. So I, yeah, I, I'd agree with all of that. I think.
[00:13:09] James: Yeah, and I say it's that old cliche, isn't it? The best time to have done this is 20 years ago? The second best time is now. And that's not to say I. Think there are definitely companies who did do it 20 years ago, and you can see you, you do tend to be able to see that in, you know, that they managed to, they had a much clearer strategy around the cycle and they managed to outperform. So I think it's also been proven that that sort of data-driven approach does. Provide value and drive performance.
One other point I was just gonna make is, so my background is kind of Lloyd to London, so I would be quite sad to see all the brokers with their big slip cases disappearing completely.
But so, sorry. A more sensible comment is I do think, you know, relationships are a very important part and will continue to be a very important part of the industry. It's just that you don't necessarily need to have it all on paper, all of the data on paper
[00:13:57] TEC: A hundred percent, you know, the data could have been streamed directly from a broker into the pricing platforms and, and sort of been semi underwritten by the time that the broker has arrived on foot to then chat about the risk and actually understand the more subjective parts of it.
And, and again, that trading relationship. I couldn't, I couldn't agree more. And to your point as well, that some companies have, I found a lot of companies have done this sort of transformation some 10 years ago, some 20 years ago that's now repeating. This isn't a, you transform and you are done.
This is very much a, a, a journey in every, i, I forget the number. Maybe seven to 10 years is, is, is the usual part. I think if you wait 10 years, you've probably left it a little bit too long, but every seven or so years, A big privacy in transformation should be happening because technology has moved on, data has moved on, the world has moved on, and you, you need to be, you need to be at that cutting edge if you are to, remain competitive.
So all of those companies that have done well, we're seeing them repeat that exercise now again, at, at, at that right opportunity.
[00:14:59] James: As I sort of said at the start that, you know, there's lots of different types of pricing transformation. So even though you did a pricing transformation two years ago, say on your technology, maybe now's the time you collected some better data. Let's start trans, let's transforming the modeling. So it, it's also a cycle of the different bits of pricing, feeding each other.
[00:15:16] TEC: Yeah, absolutely. But then it's not even that easy though, right? We, we talk about it and say we should be doing it. And I guess if it was super easy, the market would just do their transformation and, be getting on with it. But that's obviously not the case. And there's probably quite a few of these, but what, what, what do you see as the sort of main barriers for change, and, and for this level of transformation?
[00:15:37] James: Yeah, absolutely. And that actually, again, sorry. It was an interesting thing for me about moving from industry to consulting is that rather than doing these transformations every seven to 10 years, I'm actually. Getting to do them repeatedly. So I'm seeing more both the barriers and the lessons learned, and that's really fun. And in terms of the barriers, you know, I think there is an interesting thing around transformation, around pricing and underwriting that you know, you tend to see the big transformation spends and a lot of it gets spent on things where you can clearly communicate the benefit in terms of financials, and it's often in terms of improvement to the expense ratio, and that's the fundamental of, it's predictable and you can very easily demonstrate it, whereas, you know, actually.
You look at combined ratio, by far the biggest part is the, the loss ratio. You look at the spread of companies performance over the market and the spread of the expense ratio is not huge, whereas the spread of the loss ratio is huge. And so you can really, by moving from like, you know, bottom quarter to top quartile, it's enormous the amount of change you can make. But it's very hard to. To quantify, it's very hard to say Next year if I do this transformation, my loss ratio will be X. 'cause you know, you could say that and you could then hope that there isn't a cat that year or something. It'd be like, but so, but I think that's been a real, that is a real challenge and it's something we really, when we are doing this work with our clients, it's something we, we try and support them on a lot because it, it is hard to quantify that and finding ways to kind of communicate and explain the benefits that are coming both quantitative and qualitatively.
It is really important and I, you know, I genuinely believe this is where the most value for companies comes from, but it, it's a harder, it's harder to communicate that. So that, that would be my first barrier. I think there's a second barrier about any sort of big transformation, right? That they're hard and they take a long time and. You know, maintaining that sort of, the momentum and the enthusiasm and the resourcing through throughout a kind of large scale transformation does take a lot of time, does take a lot of effort, and you really need to think through your strategy and your, that kind of change management piece of.
Bringing people on the journey because there's always an example of how this hasn't worked somewhere and people who've been burnt. So, you know, it is, I think it's right that people don't jump into this lightly, that people think through exactly what they want to get out of it and plan how they're gonna get there because it, it is a really, they are challenging projects to do.
[00:18:12] TEC: It’s the vision you mentioned earlier, right? It's about setting that at the start.
[00:18:17] James: Yeah exactly. And you know, I think that's also something about the way you lead, lead people through that project. 'cause there, you know, a project that takes a year, 18 months, there's going to be a bit in the middle where it's hard, not everything's gone to timelines. unless you go with us, obviously, then it'll be smooth. sorry, you, you could cut that if you like. There's always challenging moments. There's always that bit before you first deliveries where you know you, there's a bit of momentum or a bit of enthusiasm lost. And having people who can drive that change remind people of the vision is also. So, yes, the envision's really important, but then the leadership through that as well, and the change management's critical.
[00:18:59] TEC: Yeah, it's, it, it is, it is difficult and, you've obviously seen it done, done well and maybe done not so well as well. And it's, it's an interesting point about the, the time it takes, you know, this, this, this traditionally has been a long. Journey. We've been through these in-house and also from more the consultative side.
But I thi I think that that is changing a little bit now. I don't think transformation needs to be taking five years anymore. We've had customers and clients who have done this in less than a year. and when I say done this, I mean full on transformation. You know that this is implementing a new pricing platform, new underwriting processes, new front end.
Data out all of this interconnected and completed in, in less than a year for a very large scale and legacy insurance companies as well. It's, it's not just the ones who can do it. So there there is, there is light at the end of that tunnel a little bit to, to be able to do the change. It doesn't need to be, this long thing anymore, I don't think.
[00:19:55] James: No. Exactly, and it's also how you structure it, right? Of, you know, part of that not having a slump in the middle is to ensure that there are you, you are kind of accelerating the delivery of wins and value added to the start, so that you are proving value and keeping that momentum.
[00:20:11] TEC: I really liked how you framed earlier as well about the key levers on this with, the combined ratio and obviously with the expense ratio, the fact that there isn't that much margin, you know, it is a proportion and you could, if you chop a couple of points off it, it's just as impactful as chopping a couple of points off the loss ratio.
But as you said, the difference between what the top quarter and the bottom quarter companies are is not huge. And the loss ratio one is, is massive. And that, that is a clear, a clear indicator that you can make these huge differences. And I think it's, to your point about it is it is very hard to prove.
I, I would love to be able to go and say exactly what the loss ratio benefit of a pricing transformation is, that I'm sure you would, because then the cost talks go away completely. And it's, it's, it's really obvious why you do it, but there's a lot around. You, you can see these tangible benefits and you mentioned the sort of quantifiable thing.
So if you are saving your underwriters tons of time, if you are saving your actors tons of time and allowing them to focus on actuary work, to focus on, to focus on underwriting, then you kind of like that, that sort of time saving benefit is then I. That, that will drive that. You know, there's, there's plenty of reports out there.
I think a McKinsey report recently said it was between three and 6% impact on the loss ratio or percentage points, so of pricing transformation. And we also did a survey at, gyro last year for, it was over 150 actuaries and over two thirds believe that, again, pricing transformation will have. At least a 2%, if not more impact on, on the loss ratio.
So we know the, the case for change, is there sometimes boards are, are much harder to convince on that because as you said, with an expense ratio LED one, you, you very much see. It's like, well I can make FTE savings here. That's a clear impact on the loss ratio. And then you can see the cost benefit analysis.
So yeah, it is a little bit difficult, but companies are going through it to doing it successfully. So it's, whilst there are these barriers, I think, it's, it's a very positive outlook and with the right help, From a consulting point of view and the right technology, this is, this is how you do it. And I think both of these things are available in the market now and, and companies can, should take sort of full advantage of them.
[00:22:17] James: Yeah. And actually to, to a kind of earlier question you asked me of like why this changed now. I think part another of the reasons is that a lot of the solutions that were available to companies that are available now weren't available five, 10 years ago. You had to do it yourself. In the London market particularly, and doing it yourself brings a lot of challenge.
Yeah. I think we've both been at companies that has sometimes successful, sometimes not, and having better solutions and better options out there for different types of company is probably another driver. One other thing on just on the. Kind of benefit qualification and something else that we do ourselves, is that kind of on the quantification?
Because we've got quite good benchmark data, we can also do some of that sort of stuff of, you know, here's how you sit amongst your peers if this transformation could get you to being just median. Among your peers, this is the loss ratio. You know, there is some ways you can do different ways. You can quantify it with some assumptions underneath it.
And then, yeah, I also agree there's, there's a few sort of market wide analysis by kind of this, the strategy houses giving some sort of high, high level.
[00:23:22] TEC: Anecdotal. Yeah.
[00:23:24] James: Well, I'd love to know where they got it from.
I'm sure there is some, some…
[00:23:27] TEC: I think they surveyed a load of companies and they believe it could be between, it's, it's a, it, it is a difficult thing to measure as you, as you said, you know, you could have a cat event and all of a sudden the loss ratio for that line of business is completely destroyed, or you could not have a cat event, and actually the loss ratio improves.
And regardless of, of, of what you're doing, but the, the underlying trend should always be that this is there. But, I think we are getting better at quantifying and, and seeing that. And I think companies do see that, that, that, that that benefit whether however tangible or not, it may be. And I think we, we have, we are much better now at articulating.
While that is, and I mean you, you mentioned more kind of looking forward beyond the sort of price and transformation and you mentioned sort of algorithmic. Underwriting and under. I just wanted to, to quickly sort of talk about, how companies are sort of innovating going forward on top of then the, the transformation.
So what, what does it look like in that area? And actually maybe just to go into what, what actually algorithmic underwriting is.
[00:24:21] James: Okay. Yeah, so I mean, it's kind of quite a fancy title for something that has been around a lot in different places, but you know, it's fundamentally the idea that you, your underwriting is, it's a flow rather than, And a piece of paper to piece of paper, the data cut is ingested automatically and some of the underwriting involvement is reduced in that, you know, there's one extreme where every single risk is priced by an algorithm and you know, there's no underwriter involvement. There's another one where Some of the base level ones are priced by, the algorithm and then kind of referrals are kicked up to the, the underwriters as they wanted. And you know, you've seen this in the kind of s m e in retail that it used to be almost everything was referred and over 20 years it's moved to be, almost nothing is referred and the London market is a lot earlier in that journey. But you, you are seeing people. Do try and build that out more and more in different ways. You know, I think the big poster boy example is Britt's Key syndicate, which is with its automatic follow, which is doing one type of algorithmic underwriting. You know, it's, it's follow only. There's still always somewhere a human being doing the, the lead underwriting, but that that choice of follow and the choice of follow line is purely algorithmic.
You're then also seeing, I think, lots of other examples of companies who are doing it for their. You know you're doing lead follow business for smaller kind types of business, general aviation, pi, that sort of example where it's kind of a broker links direct to the broker. The broker feeds the data in, and then it's automatically priced and quote issued. So that's starting to build up, but sort of. Touching to the point we said at the beginning, that, that I feel it feels that that's gonna come more and more and it's gonna become the default more and more that you need to have those capabilities. And so to me it feels like there's two, two paths emerging.
That there's gonna be this more and more the, the simpler business, the business that you're following will become. Automated will become algorithmic, and then the more complex stuff still gonna be underwritten by, by, you know, by experts and there, you know, the, but there's still a transformation there of giving a lot more.
Well, it's, I think, I don't think I coined this, but this's kind of like concept of kind of, Augmented underwriting so that the un, you know, not that the underwriters become a Borg or something, but that the, you know, that the underwriter's giving a lot isn't just given, here's your risk, here's your technical price.
But it actually is given a huge amount more insight into the risk. Does a lot. You know, a lot of the data flows are still automated and that, and there's probably still quite a lot of algorithms behind that, giving him a better insight, pulling in external data sources automatically. So there's a sort of transformation on the underwriting side that either is to feed an algorithm that automatically prices the risk or to feed an underwriter with a lot more insight to make a better decision making. So I think it's both. Both is kind of, the structure behind both is actually gonna be quite similar in some ways. It's just whether it's an algorithm or.
[00:27:28] TEC: And, and all of this is basically enabled by that data flow, right? But both, both of those require that streamlined data. So either the algorithm or the human can actually, again, do their insightful parts and, and, and analyze the risk as opposed to spending half of it re-keying of the
[00:27:44] James: Yeah. data is like absolutely critical and I think all of this is based on the assumption that we get better at, we as an industry get a lot better at using our data and transfer, having consistent data standards and consistent data transfer.
[00:28:00] TEC: Absolutely. And I, I guess, it wouldn't be topical unless we sort of talked a little bit about AI and generative ai. I came across a phrase of the day, I, and I think this particularly applies to insurance, that, I don't think underwriters or, or actually are gonna be replaced, by, by either of these things, but you may well be replaced by someone who knows how to use them very well.
[00:28:21] James: Yeah.
[00:28:22] TEC: I was intrigued. 'cause, 'cause we have seen AI in the market, a bit in the specialty space, particularly, on some of the lines you mentioned, like PI, General Aviation, where you have a high volume of, low premium quotes and you need to do bits of risk triaging. So there's been some machine learning models applied to these already.
But with the, with the sort of emergence of, of the generative ai. So the more the chat GPTs. Do you see anything here infiltrating into the insurance space or is, is this, is this very much a, to, I almost thought I'd write out my intro on chat GPT, and then I was like, no, I'm just gonna do it myself.
But there, there's, there's lots of interesting use cases is my point. I'm just, I'm curious to, to what your thoughts are, whether they, apply well in the insurance space or, or indeed, anywhere else.
[00:29:05] James: Yeah. And kind of on the non generative first, I guess. I, I don't know how much of the kinda algorithmic underwriting is true kind of ML AI now. I think a lot of it's gonna be quite, I'd, I'd love to know what Britt's Key is doing. You know, I'm sure they are, have got, they have a lot of AI capabilities in-house and I, so I. I'm sure they are doing a lot there. A lot of the algorithmic underwriting I see there is, you know, quite simple rule based at the moment, which is fine 'cause that's all it it needs to be. I'm sure that will develop over time. I'm also think, I also think a lot of that kind of ML and kind of natural language processing. Also applies to the sort of the flow, the data flows of, you know, linking data, identifying gaps in the data, pulling in data sources and matching them all together. So I think there's a lot of, not necessarily the pricing, but the kind of flow of data and the bits around the pricing that will, that can be. Immediately helped with, with these models. On the kind of chat GPT I, I'm a little bit worried that by the time this podcast gets released, it'll be, everything I say will have been invalidated and I will be,
[00:30:10] TEC: It's moving quite quickly. Everything.
[00:30:12] James: Exactly. But, but I mean, I'm sure there, and I liked your comment of like, you know, not being replaced by ai, but being replaced by someone who knows how to use the ai.
'cause you know, clearly using something like chat, GPT is a skill. And my only use of it so far is I got it to write a, bedtime story for my daughter where I told it to make the main character her name. You know, I made it themed around rabbits 'cause she likes rabbits. And I made it, it asked it to have an underlying theme of, sharing being important. And it knocked me together a five minute story. Really good, but that's, not particularly related to insurance.
[00:30:47] TEC: Awesome. No, no, but that's. That's. What a great use case. So you are now, you are now the master of bedtime stories.
[00:30:58] James: exactly. I just got, lemme lemme think of a story, off the top of my head. but it, I mean, but actually doing that made me realize, you know, it took me a few attempts to work out how to
even something quite simple, how to type that in and how to ask the right, right questions. And it, to your point of knowing how to use it, I'm sure it will come and I, you know, I'm sure you know, there's some just super obvious use cases around, you know, you've got a slip under helping the underwriter understand that slip.
Understand, you know, some of these solutions are already in the market, just not using the, these technology and I'm sure it will supercharge that of really helping identify. The understanding of the kind of contract wording. Has anything changed since the last time you looked at it? All that sort of stuff.
You know, a lot more automated checking of that sort of data inputs and contract. You know, if you wanted to check if every one of your contracts has a silent cyber clause, suddenly. You've got probably got, but I mean that's, that's not necessarily a pricing one, but it's afu. It's a fun one.
[00:32:01] TEC: For insurance in general and I guess it can, it does sort of fit into the price and underwriting space, right? If you wanna understand slips and, and understand your risk profiles, then, then that's kind of all part of it. no, no one's getting all of that information outta slips at the moment.
And, I do remember having to troll through, lots of ones to find sub limits of various, bits and pieces and, and that that could easily be done through there. So it's an interesting topic. We, we shall see. I think, it will evolve over the over the coming months, you know, there's, there's still huge data privacy issues around chat GPT and people need to be incredibly careful about what they put in there, and particularly around company IP company data.
It's open on the internet, so yeah, you, you need to treat it with, the, the same, respect as you do when you are, when you're googling for, for information or, or transferring data. But, we shall see, we shall see where it, where it lands coming towards the end of the year.
[00:32:55] James: I was gonna say, 'cause I do see that there is. Is an opportunity to have almost like an underwriter assistant, right? That the underwriter could sit there and you know, it's pricing a risk. And you could ask it questions and it would, you know, how many mines like this? Mine have blown up in the world. And it could give you answers and it could give you sources.
And as you said, you know, it could help you interrogate the slip and think about the risk you're writing. But as you say, it's got to be worked out. And yeah, you've got to be not uploading your documents onto the, to the. To be available to everyone as you do that. So I'm sure that, I'm sure it will come in some form or other, and it, it seems very powerful as a kind of strong way of kind of processing and understanding text. Will we really, we'll be watching it within just what, what comes of it?
[00:33:39] TEC: Me too. Me too. James, this has been absolutely brilliant. I have one, one last question, which, I'm sort of asking to everyone and I think given our topic, this, this is probably a sort of easy answer really, but what do you think insurers, reinsurers, MGAs sort of need to be doing now to prepare for what will inevitably the, the softening market to come?
[00:33:59] James: Yeah, and I mean, it's everything we've been talking about, but for me it's particularly building that framework of, you know, that kind of pricing and underwriting framework of you are seeing that rates are starting to slip. You, you, you've, first of all, you've got the MI to understand that, but then you've got the mechanism to be able to.
To enable your kind of senior underwriters, your portfolio underwriters, a way to steer their portfolio and to both kind of set out a strategy for how, how they want their teams to be renewing, their books, and then to be able to monitor it in real time and react. I think that's gonna be super critical when you are kind of moving into this soft market because it is such a, it's gonna be moving into such uncertain times with so many changes at once.
[00:34:45] TEC: Great. Yeah. Good advice. So yeah. Brilliant. Thank you very much James for joining and, for discussing incredibly interesting topic. And, I'm, definitely going to be stealing your idea of chat, GPT for bedtime stories. Much, appreciated. So, thank you for joining.
[00:35:02] James: Yeah, no. Thanks Tom. Great to be here.
[00:35:06] TEC: That concludes another episode of TEC Talks. If you enjoyed today's show, wanna find out more about the topics discussed, head over to hyperexponential.com to gain access to a range of resources related to this episode. The link is in the description, and of course, wherever you're listening to this podcast, make sure you like, subscribe, and leave us a comment or review.
Thanks so much for joining us and see you next time. Bye.
