Within specialty and commercial insurance, precision is paramount. In these markets, where each risk is unique and every policy tailored, pricing isn't just a matter of number-crunching. It's an art, a science, and a strategic lever. But as the digital age continues to unfurl, those who overcome the industry’s stubborn resistance to change are the ones who thrive.
Pricing innovation doesn’t need to grand, all-encompassing transformations. Incremental, data-driven changes can deliver meaningful improvements to underwriting workflows more cost-effectively and with lower risk.
In this episode, TEC speaks to hyperexponential’s Head of Pricing and Innovation, Jamie Wilson about the current state of Specialty Pricing and the opportunities ahead. Together they examine how innovators in the field are leveraging cutting-edge technology and a deep understanding of data to reconfigure their pricing models.
The softening of the market looms on the horizon, and the companies that will endure are the ones that think ahead. We'll discuss why it's essential to engage in forward-looking conversations now, exploring how to fortify your operations for the coming challenges. From KPIs that reveal where you stand in the market to the imperative need for strategic investment in innovation, learn what the path forward looks like.
In this episode
Unlocking the business value of data
The challenges posed by technical debt
The role of machine learning and NLP
Fostering a culture of innovation
Preparing for market shifts
Never miss an episode
Subscribe on your favourite podcast platform:
Full episode transcript
[00:00:00] TEC: Hi, I'm Tom Chamberlain. That's absolutely everything you need to know about me. I'm delighted to welcome you to the latest TEC Talks. In this episode, we're talking to Jamie Wilson, who's the Head of Pricing and Innovation here at hyperexponential. So let's get straight into it. I'd like to welcome Jamie Wilson to the booth today, and Jamie is a qualified actuary and has many years of experience in pricing specialty insurance, as well as being a pretty decent coder, and most importantly, an exceptional innovator. In this episode, we'll be discussing a topic close to Jamie's heart, innovation in pricing and underwriting.
The focus will be on specialty and commercial insurance, and we'll be discussing what's already here now, so what's coming in 2023 and beyond from innovation and what lies further ahead. So welcome Jamie, and let's dive into it with the first important question. Why should companies really care about innovation in pricing and underwriting?
[00:01:04] Jamie Wilson: Yeah, it's a great question, Tom. And I guess firstly to start, thanks for having me here. Really glad to be talking about such a fun topic. Innovation in pricing and underwriting. I get it's kind of there's always this classic case that people talk about right with innovation. They kind of if you're not doing it, someone else will be.
Kind of people look to things like, you know Blockbusters used to be the big thing until Netflix came along And kind of cds used to be the big thing until Napster Spotify came along and in reality There's always this risk within industry within kind of any form of business that if you're not looking To innovate someone else will and they could either come and disrupt the market And sometimes that happens in like big fancy ways like, you know, Spotify like Netflix But in reality, it doesn't always have to be so disruptive I think kind of innovation can be on lots of small things and just getting one good piece of innovation in your industry or in your company can give you an edge over your competitors.
So you don't have to completely disrupt the market with innovation. It can be something kind of small that allows your underwriters to outperform your competitors slightly. And that will kind of lead to loss ratio improvement. And I think that's the challenge is if you're not doing it. Someone else is going to be doing it.
And if someone else is doing it, their loss ratio is going to improve. And chances are your loss ratio is going to deteriorate. So if we're not investing in this space, we're just putting ourselves at risk.
[00:02:35] TEC: Yeah, makes sense. And, the Blockbusters is a good analogy of not keeping up. But do you think with with those sort of industries versus insurance is insurance right for this kind of thing? Or I mean, it's a relatively slow moving, slow paced industry. So is innovation really, really a possible thing here?
Does it really give you the edge do you think?
[00:02:55] Jamie Wilson: It's a great question. And I think this is the thing, right, is, there's, I would kind of break innovation into those two areas, that disruptive area. And in my mind, insurance isn't ready for a disruption. But then again, everyone in an industry often thinks that industry isn't ready for disruption. I think it's that kind of small innovation changes that I think the insurance industry is completely ready for.
Maybe not all companies. To be honest, there are going to be some companies that are kind of more pro innovation, more look forward looking that are kind of trying to find out, okay, what can they bring in that will give them the edge over their competitors? And I think that's why kind of there's such a focus on like small things that can happen.
Within the innovation space that can really impact the underwriters lives. And it can also the pricing space as well.
[00:03:40] TEC: Yeah, I think we have some bits of this in more personalized type of insurance in the small business market where there's a lot more data. There's much more kind of scope for looking at the data and innovations in different ways. I guess specialty is still catching up a little bit, but What's really been happening in the specialty market?
What, what kind of big leaps forward have you seen in the past few years?
[00:04:02] Jamie Wilson: Yeah, I think kind of in the last few years, I think there's been this general trend of looking to kind of the retail market. I think, especially in the pricing space, I think kind of this started maybe 10 years ago in the pricing space of looking to the retail market and saying, okay, what are they doing there?
And what can we adopt into the specialty market? And some of that's been more successful than other parts. I think there's this realization that when you're looking beyond your own area and kind of looking to other sectors or other kind of parts of the market for innovation, you have to adapt it. So I think kind of some early adopters may have kind of said, okay, fantastic.
This is how pricing. is done within the retail space, let's do it like that in the specialty space and quite quickly realize that we don't have that level of data. I kind of, you know, our risks are not as homogeneous as they are in the retail space or kind of our claims are a lot more volatile. And suddenly things that work in the retail space don't actually work in the large corporate specialty space.
So I think that in the pricing. Part of the insurance market, there's definitely been this kind of trend towards, okay, well, how do we actually adapt those approaches now kind of using far more kind of advanced machine learning approaches that allow for the lack of data or kind of the volatility we're seeing in the data and pulling out kind of signal that can be transformed into kind of a really valuable kind of pricing assets, kind of a pricing model that actually helps kind of our underwriting teams to differentiate themselves.
I think that's. One thing that we've seen more and more, especially over the last five, three to five years, I'd say there's been this real change, not in all insurance companies, but definitely kind of more and more in the specialty space. I think the other, the thing that I'm probably one of the areas I'm most excited about, and I feel like, you know, I, I head up a predictive analytics team in a pricing space, so I should be most excited about what I've just talked about.
Um. But, and I am, don't get me wrong, the, I think what's happening in the submission space is something that I, I think is going to be an absolute game changer. And that's effectively kind of, everyone's looking at kind of how submissions happen within the London market or kind of within the specialty market.
And they kind of say, okay, well, you know, we're sending PDFs, we're sending word documents, we're sending Excel spreadsheets, and you know, it's happening by emails now, right? So we're not kind of doing it on paper. So that's a, that's a leap forward in, in and of itself.
[00:06:22] TEC: Is that true? Is, is it brokers that's still not coming in with big, big, big files or papers and, ready to quote at the…
[00:06:30] Jamie Wilson: Yeah, I mean, I think that does happen still, but kind of it's becoming more rare, I hope at least, or kind of from what I've seen, but you know, it even but I think kind of people say, okay, fantastic, you know, everything's digitized. And in reality, you know, in my mind, that's, that's great. That's the first step, but it's still not digitized.
Right? You know, if I've got a PDF in an email somewhere, And that effectively doesn't go anywhere, maybe other than kind of the underwriter typing in one or two pieces of information out of that PDF into some underwriting system into a pricing tool. Is that submission really digitized? And I think this is where, for me, this is kind of one of the areas that is seeing kind of some of the biggest transformation in the London market or kind of the specialty market at the moment.
And I think we'll continue to see that change. I think it's coming in two places for me. The first is an area of OCR NLP. So this in reality is probably one of the main applications of AI that we're seeing is kind of AI that is being used to read submission data. so kind of look through the PDFs, look through Word documents, pull out kind of the relevant data items and process it so that kind of we go from saying, okay, actually, we've got a load of PDFs and Word documents, Excel spreadsheets to having a lot of structured data.
That's kind of one of those areas. I think on the same time, I think what we're starting to see in some of the kind of the larger companies, or maybe in some of the breaking spaces. is this realization that, well, why are we sending PDF documents? Why are we sending Word documents? Why not kind of structure that data, come up with a common data model, and then send that data in a structured format to kind of our insurance partners?
And so kind of, you're seeing these kind of two, like, tackling the same problem from two different ends. Kind of one is saying, okay, when I've got unstructured data, when I've got PDFs, how can I read them? And then on the other side saying, well, why are we using PDFs? Why don't we structure that data? And I think those two things coming together.
I always like to avoid using the word disruptive, but I think it really could be one of the most disruptive things that's happened in the specialty market for a long time.
[00:08:35] TEC: Yeah, I use a slightly different word sometimes for this, which is transformative because this, this is going to utterly change the way, that the business is done and cut out this huge admin burden on underwriting, which quite frankly is, is, is not needed in this supposedly digitized world. There's an interesting kind of concept.
You had the two sides of it there. The first, the first one being, Sort of be able to extract this and turn unstructured data into structured. But then the second side is actually, can we get the market to bring in structured data now, having, having worked in underwriting and having received submissions from brokers, you could, you could receive a general aviation submission in the morning from, from a broker and in the afternoon from the same broker, a completely different set of documents and Excel sheets and stuff.
And, It's a tall order, I think, to try and get that market at the moment to at least to bring in this heavily structured data because different insurers are going to want different inputs. There's no kind of common wordings. there are in some markets, but no common question sets that people want to know the answers to.
I do remember a small business commercial moving this way. It must have been 15, 16 years ago now where iMarket came along and, there was a many, years of arguments with insurers over different question sets, asking various different questions like, how many windows does this building have?
But someone else wants to know, how many windows does it have on the front? And, and that's just an important question. So, I would love to believe the market's going to rally round and, have a common data set, and that's definitely a utopia. So, Then, then the answer probably in the shorter term lies in the the first one.
So, I mean, you mentioned OCR and NLP, that that's gonna allow you to sort of pull PDFs into some, some form of, of structure. But then how do you then take that into a submission? Because again, unless you're getting it in a common format, how do you extract that information and, and sort of make it sort of into a submission that could be either automatically entered into pricing tools or, or, or policy admin systems.
How does that work?
[00:10:41] Jamie Wilson: No, that makes sense. So it's effectively around this idea of saying, okay, well, you said every insurer has their question set, they have their data points that they want to know about a risk. And so kind of, I think kind of talked about aviation there. So it's, you know, how many aircraft are there? Kind of what type of aircraft are there?
So kind of every insurer is. Every kind of underwriter is effectively saying, okay, I know what data I need to extract from this. So kind of that NLP OCR is effectively structured to say, okay, well, what are the data elements that we want? So we effectively kind of, as an insurance company, would, you know, develop our own internal data model for kind of, Any part of the specialty market.
And so this effectively means that data model gets populated. We have a database with every single submission that's come in with the data and a nice structured format. The beauty of that is that then obviously we can then push that data. repeatedly. It doesn't matter which broker has come from because we've then taken out the data, put it into the same format in the database.
We can push it into our pricing processes. We can put it into our pricing tools. We can push it through sanction screening. Any kind of part of the underwriter process, which involves manual wreaking should be able to kind of then be automated. So like you say, the underwriter is kind of, the burden of administration for the underwriter is kind of reduced, and I think that's why I think often when people look at submission automation, it's that part that people get focused on and it makes a lot of sense, right?
We look at how the underwriters are spending their time and we know a lot of it is on the less value added work, when actually there's a lot more value added. work that the underwriting teams can be doing. So it's saying, okay, well, let's take away the administrative burden. But then you've also got to think, actually, kind of, the, kind of, the next levels of, kind of, transformation from that will come from that data that we're actually capturing.
So, kind of, historically, we didn't capture all of the submission data. What can we start doing with that? And I think that for me is where the true disruption will come once we start getting into the space of actually what kind of advanced analytics, predictive models, machine learning can we apply to that new data asset that no insurance company has had historically in the specialty market?
[00:12:45] TEC: I mean, it sounds easy, right? We should, we should automate the upload of data. So why, I mean, automation, we've talked about, I've been talking about for years. You've probably had the same thing. Everyone's been discussing automation because it's been around other areas of insurance. But we're still talking about it.
And why do you think that is? Because again, it just sounds like a relatively straightforward, well, we'll just, we'll just take the data, put it in a structured way, and then we're all good. Well, what's, what are the barriers to this?
[00:13:15] Jamie Wilson: I mean, yeah, why are we still talking about it? I mean, sorry, it's a slightly ridiculous answer because it hasn't happened yet, I guess, and everyone, you know, there's such a demand for it. We all look at the systems and we all look at the processes and I think everyone is agrees this needs to be automated and we still haven't done it and I think you're right you've picked up on it Tom it's it's because it's not easy it is challenging you know and artificial intelligence NLP a kind of natural language processing that's used to extract this data is moving on leaps and bounds every year.
So kind of the, the, the, we are actually using kind of the most advanced approaches in this space. Maybe kind of NLP probably wouldn't have been sufficient five years ago to do this. I think it's, it's that type of thing where we realize actually in order to automate some of these processes, some of it is just going to come from using the most advanced data science techniques are going to be coming from using the most advanced technology.
And. In reality, the insurance industry is often quite far behind when it comes to those types of things. We're not necessarily known for being on the cutting edge of technology, let's say. And so I think that's, and that's, I think that's one of the key problems. It really slows us down.
[00:14:27] TEC: I think you kind of, you kind of mentioned it earlier that, it's, it's allowing the underwriters to get on with the job of underwriting. And, I saw some, I saw some stats recently was saying something like 40 percent of an underwriter's time is, is, is on data admin. rekeying and triple keying and quadruple keying, eliminating even half of that is going to free up a day for your underwriters to then go and actually do some more stuff that they're good at, which is the analysis of the risk.
But then what, what, what on the other side that I mean? You're kind of getting, getting the data in and you're able to now ingest it. And it's all that data burden admin is taken away. Why then is it so important to have this data for underwriting and, of course, pricing as well?
[00:15:15] Jamie Wilson: Yeah, I mean, thinking about the data needs for on the submission side, they kind of, there's so much stuff that could be done with that data. So if I have all my submissions in a structured format, then, you know, I can build models, which will predict which of these risks is going to be better priced, like, kind of worse priced within the market.
So which risks are actually going to be most profitable? And saying, actually, Underwriter, you know, we've removed that 40 percent of time that was spent on administration, so you've now got a lot more capacity. At the same time, I'm now going to be pointing you in the direction of the risks, which the data is saying are going to be the most profitable.
So these, and these aren't tools to say, okay, let's replace the underwriter. They're tools to then augment the underwriting experience. So not only taking away the administrative burden, but actually empowering our underwriters by using these types of models and saying, okay, well, which risks are going to be most profitable?
Which risks are you most likely to bind? Which risks, are fit best within our risk appetite. so there's all sorts of models we can build there. Even things saying like, if we were to capture all the submission data. One of the, one of the biggest concerns, within pricing model development is that we don't have enough claims data.
Um, you know, often kind of, if I've only been writing in a specific industry or a product for the last two or three years, I might not have enough claims data to build a valid pricing model. If I'm capturing all my submissions data and that submissions data contains claims data, suddenly I can start building pricing models.
Based purely upon the market and again, augmenting the underwriter experience. So I think that for me is the exciting part around the data capture piece. But it's not just on the submission side, right? This is true about the entire data journey through an insurance company's infrastructure. So it's okay, let's capture better data at the submission stage.
But what about the data we're capturing within the pricing stage? What about the data we're capturing within the PaaS? Like, how well connected are our various kind of different infrastructures?
[00:17:11] TEC: And is, is there, is there something to be said for sort of other third party data as well, if you're able to ingest all of this or all of the submission data and pull in other stuff that you have around your company, that there, there are 70, I think it's something like 70 zettabytes worth of data, which was digested, ingested, analyzed last year, just that, that year alone, so that's 70 trillion gigabytes.
It's, it's an extraordinary number. yeah. And a vast amount coming from, from IOT devices, and a lot of which you could think, actually, that this could be useful for insurance as well. And so do you see sort of innovations around being able to get this data in and kind of crunch it up and again, analyse it in a different way?
Because once you start getting this vast amounts of data, I think that the human head starts to explode a little bit, and, and knowing what to do with all of that.
[00:18:07] Jamie Wilson: Yeah, no, definitely. If you tried to give me 70, that was how many zettabytes of information. I'm pretty sure my head would explode. So yeah, external data is, you know, again, one of these areas where there's lots of excitement about people know that in order to be market leading, I need to be using external data and it's because of, you know.
If we're not doing it, other people will do it and we will be selected against. But I think it's also a place where people really struggle. Often kind of the use of external data is underwriter goes to websites and searches something up manually. And in reality, external data becomes more pain for the underwriter.
And rather than saying, okay, well, this external data is actually contributing to the value add, actually augmenting the underwriter experience. So I think with. External data. Yes, it has to kind of be part of the package, but we also need to make sure that we've got the technology in house to kind of bring it in to the underwriter kind of experience automatically.
And we've also got to make sure we've got the technology and also kind of the data science capabilities to actually process that data. Simply, you know, looking to an external data provider and saying, ah, yes, let's say these are the risks that are good. These are the risks that are bad from the hundreds of data points that external data provider.
gives us. Simply, that's not going to be good enough. You are going to have to this when we start moving into space, you're getting more and more data, it becomes harder and harder for individuals for people to pick out trends. And we do need to start moving towards kind of using more kind of advanced analytics approaches.
[00:19:38] TEC: And is this, is this, um. Is this where now, because everyone's been talking about machine learning and AI for quite some time, is this one of the spaces where this is we're going to see more of this now? Or are we already seeing it? And is this is this the only place? Or do you see other places for that kind of application?
[00:19:56] Jamie Wilson: No, it's a great question. I, I think external data is obviously there will definitely be a place for machine learning with an external data in part to process it. And it depends entirely upon what the use case is, right? am I using external data for pricing purposes? Yeah, machine learning obviously fits the bill.
Am I using external data to simply say. I want to know about one very specific data point about a company to say, actually, do I actually want to even look at underwriting this risk? You know, maybe it's their credit score and saying, I, as an insurer, do not want to insure anyone with a credit risk below certain X, in which point I don't need machine learning.
So it depends upon the use case, but it's definitely external data is definitely a space where machine learning is a necessity when you get to the kind of the larger data sizes. And I think this is the thing, right? Within insurance, there's lots of different areas. You can use machine learning. I think often the challenge is identifying the most beneficial areas or kind of the most profitable areas to use it.
And I think that in reality, in my mind, I think we saw kind of this surge of interest in machine learning and data science, maybe five. He may have been even up to 10 years ago within the insurance space. It started kicking off. And I think a lot of people came in from outside of the insurance industry and try to apply data science, machine learning to everything.
And there was a lot of kind of, I think a lot of hopes were dashed because I think everyone's like, this is going to solve all of our problems. And in reality, you need to have a real understanding of the business. And a real understanding of, okay, what is the pain point I'm trying to solve? Or what is the kind of the edge that I'm trying to obtain from building this machine learning model?
So for me, I think kind of machine learning is happening, but I think we've matured a lot as, as kind of an industry in knowing that it has to be combined with real business understanding in order to get that profitability uptick.
[00:21:7] TEC: Yeah, and then this, this probably brings us back to the difference between the sort of small specialty type of insurance and personal lines where personalized. You've got a lot of homogeneous risks. You've got a lot of claims data. You are finding those little marginal, Edges, which extra data sets and a bit of a I plugged in probably will help optimize your portfolio.
Whereas with specialty lines, you either have no data at all for new classes or data that's utterly relevant, like we're within cyber, which is constantly evolving and even for more established. lines of business, the risks are just still so dissimilar that it's very hard to do that. And you're ending up with very sparse data sets and applying AI onto those, even with augmented data, just through the sheer lack of claims is just not going to help you understand that.
So I guess, I guess, I guess that's one of the, one of the barriers then for, for that type of innovation as well.
[00:22:40] Jamie Wilson: Definitely. And I think it's also knowing, I think insurance companies need to know what data they actually have. And, and say, okay, well, I'm going to apply machine learning where I have the data to do it. Like you say, if I've only underwritten a risk, or if we've got something like cyber, it hasn't been around long enough necessarily, I might not have enough claims data to do something.
But do I have enough data on the pricing side to try and identify which risks are going to be underpriced or overpriced according to the market? Are there things that I can do to say, actually, well, I, you know, I don't have claims data. I do have pricing data for every risk that I have bound within cyber.
Can I try and use machine learning to help me identify where my pricing model is overpricing or underpricing compared to the market? So it's, I think often it's this case of saying, actually. Being a bit innovative in how we're looking at the problems we see and saying okay Well, I don't have this data.
I do have that data. What can I do with it?
[00:23:32] TEC: Yeah, that makes a lot of sense. And well, what's what's your what's your thoughts on the market at the moment? Our companies, did they have a good grasp of their of their current data? I mean, certainly, certainly years back, it was it was definitely not the case. I'm hoping there's been some improvement, but maybe you tell me otherwise.
[00:23:50] Jamie Wilson: Yeah, I I would say I it's probably not as optimistic as as you'd hope it would be my answer I think you know The insurance space often suffers from a real technical debt, you know, but larger insurers have the joy of, you know, kind of larger, more well established insurers in the specialty market, you know, sitting on a wealth of data that they can't get at because of the technical debt and your kind of work with teams and find out that actually no one has ever accessed this database or no one knows how to access this database or we, you know, you really have to kind of spend a lot of time digging into: where is our data? What data do we have? You know, it's there. You just can't get to it. And you kind of like contrast that with, you know, smaller insurers, kind of newer insurers who probably could get access to the data, but they have only been, you know, if they've been kind of set up in the last couple of years, they don't have the same wealth of data.
So it's this kind of wonderful kind of contrast to say, okay, well, you know, those who have can't get it. And those who don't have could get it, but they don't have it.
[00:24:55] TEC: I guess that's, that's another barrier, right? It's the, the, the, the legacy side of it. And, yeah, that, that makes sense. And, are there any other barriers? We've, we've talked about a few, right? We, we've talked about sort of legacy systems and the data side. What did you see any other barriers to innovation?
Are companies just not investing in the right areas?
[00:25:13] Jamie Wilson: It's a, it's a, it's a good, good question, I think, kind of, you know, and I've, I've actually spent quite a lot of time talking about this internally about kind of what are the barriers to innovation, I think, like, technical debt, like I mentioned before, is often one of those saying, okay, well, the insurance industry is, you know, significantly far behind on kind of technology.
You look at the banking space, you look in obviously the tech space, we're using our data technology and we've been very slow to transform that. I think that's kind of often a big, I don't know, a hurdle or barrier to change. I think there's a lot of cultural issues with innovation as well.
I think a lot of people talk about wanting to do innovation, but when it actually comes to it, they, you know, people kind of start kind of. Like clamming up a little bit, so kind of you'll say you'll have board members talking about across the kind of insurance space about how we're going to be innovative.
We're going to change things and then actually kind of you hear these horror stories of people saying, Oh, can we try this and being told? Yes, but it has to succeed first time or kind of yes, but you've now got to kind of fill in this many forms to even just kick off a proof of concept or even just engage with a third party vendor on a potentially potential innovation that could be applied.
So I think there's definitely pain points there. Yeah. I think it's improving, or I hope it's improving. And I think in part, there will be some companies that do it better than others, and they will effectively force that improvement in the companies that are slower, because, you know, when innovation happens, when companies start to do things and they start to take the lead within the market, they will become more profitable.
Other insurance companies will either have to follow and kind of pick up the same level of innovation, or they will fall behind.
[00:27:01] TEC: I think you picked on a really, touched on a really interesting point there, there as well, which is, it's all very well. Say you are going to be data-driven and say you'll be innovative and actually investing the money, but then you actually have to have the right culture to, to want to be data-driven.
And it's, it's very easy to continue to do things in the same way and even when presented with more information, not actually use it. So, you do need a big cultural shift in these companies to, to, to realize these benefits. Otherwise it's just, money sunk into some cool looking tech and, that, that doesn't necessarily help you.
Yeah, that makes that makes sense.
[00:27:35] Jamie Wilson: I think kind of the, the other thing I'd add to it as well, I think there's a certain level of cynicism that people need to kind of try and shed from their skins. And I think to a certain extent, a lot of that cynicism is sometimes well deserved. You know, people have been told about all changes coming, you know, ML machine learning is going to solve all of our problems.
Like blockchain is going to solve all of our problems. AI is going to, and I think kind of people get exhausted. They kind of get innovation fatigue, but in reality, the kind of the reason I think kind of. That innovation fatigue or kind of that cynicism can be useful if people are going to be slightly more questioning and say, actually, which what if this happens, how will actually impact my business?
Okay, if you build this machine learning model, what will actually change? There's some really kind of useful things that can come out of You know, healthy cynicism or kind of healthy skepticism to kind of challenge these things and say, okay, well, what value is actually going to add? But then I think kind of often in the insurance space, we're stuck, not a healthy skepticism, but just kind of kind of cold hearted cynicism, kind of complaining.
And, and I think that that is something that that does need to change. Unfortunately, I don't know how to change that, but it does.
[00:28:45] TEC: It's an interesting one. There's there's a lot of project fatigue in companies as well. And, you and I have definitely both been through those with projects which take years and years to implement and then don't quite have the same, quite the big delivery that are expecting. And there's there's these expectations that innovation needs to be this huge Great big jump in in capability, and these are the ones that tend to pretty much fail and fall over because you're trying to do too much all at once.
And the way that I've seen it work well is where you're trying to do incremental changes. And you can you mentioned this right at the start? Don't try and go for this big bang. You know, everyone Talked about blockchain, as you said, and expected it to do this big, wonderful thing, but actually just just trying to apply it in that way just doesn't just doesn't work.
You need to think about the problems you're solving and incrementally solve them. And then you kind of continually innovate towards this bigger goal. And before you know it, you've kind of got to where you you needed to be. Plus the fact that if you try to innovate for something which is coming in a few years time, By the time you get there, you're going to, you're going to be behind someone else who's innovating on something different.
So it's, it's, yeah, you need to do these things slowly and steadily. Otherwise, yeah, you just, you just keep getting caught out by these massive projects, which don't always work. Maybe just one or two more questions then. So And this is probably one of the harder ones, but what do you see kind of coming potentially in a sort of more future horizon in the next few years and get out your crystal ball and, tell us where to invest in the next tech
[00:30:17] Jamie Wilson: yeah, I think, you know, the next few years, the horrible thing about saying something like this and being recorded is that someone in a few years time can look and like come back to me and say, Jamie, you were horrifically wrong. Well done.
[00:30:28] TEC: or completely correct.
[00:30:30] Jamie Wilson: yeah, and actually, what actually what I'll do now is I'll list, you know, 20 things so I can at least get one or two right.
And I think, What's going to change in the next few years? Better technology. I'm sure that's going to, you know, we're seeing it happening. I think there's a demand for kind of more cloud based technology technology that can be more agile. I think there's this realization with it, broad realization within the industry that we simply don't have the technology we need to innovate.
Um, we can't be quick. We can't be agile. if we've got a significant amount of technical debt. and I think we'll see some companies do that better than others, but I'm sure that is something that has already started to happen. And I think we'll start to see kind of real benefits coming out of that in the next few years.
I think the next few years, this submission piece, I think kind of this, you know, we'll start to see progress there. Will it be perfect in a few years time? Probably not. I, I really doubt that, but I think we'll start to see some, hopefully, products start to see the benefits of that. Hopefully, some underwriters start enjoying the benefits of that as well.
Um, I, for me, I think I'm slightly more optimistic that the industry will come around some kind of common data model. I probably am overly optimistic about that because, you know. Often that's the way I go. So I, yeah, I think kind of that submission piece. And then I think we'll start to see machine learning, you know, actually lead to some benefits.
And it's never going to be that kind of exciting. Oh my goodness, machine learning has transformed my life. But it could very well be, you know, machine learning has, you know, created something that helps my augments my underwriter experience. So it gives an underwriter an indication about which risks they should focus on.
It. gives kind of our pricing models more accuracy or allows us to kind of identify where are we overpricing, underpricing or maybe even on the claim side, you know, it's allowing us to identify where should our claims handlers be spending more or less of their time. so I, I think kind of there's, there's lots of things to be optimistic about.
So, you know, hopefully that's all shed our, cynicism and, get on with it.
[00:32:40] TEC: No, I genuinely would love to see common data sets across the industry as well. That would be a happy, happy day. but it's good you mentioned things which basically touch on sort of two key areas, which is sort of the operational efficiency. So helping free up the time to go and do better jobs for actors to be actors and underwriters to be underwriters.
But then also innovations around helping. Actuaries also understand the data and helping underwriters also understand the data. So you've got these They're both sides because I think both are both equally important for for companies coming up right now That's really great. Jamie. Thanks. Thanks very much I have one more question, which i'm which i'm asking everyone and you've probably touched upon a few bits here and there uh in the last half an hour or so, but What do you think insurers, reinsurers, MGAs, what do they need to be doing sort of right now to prepare for when the market softens?
We're in a nice hard market now, but as the cycle begins to go the other way, what are some of the, well, one or two key things that you think they should be doing?
[00:33:40] Jamie Wilson: Yeah. I mean like hoping that the market doesn't soften. it's probably, you know, Probably a terrible idea, but it's a lovely one. no, I, the short answer in reality probably would be more than they are doing, is my guess. I think kind of, you know, it's human nature to kind of enjoy things while they're good and to kind of, you know, not think too much about the negative side of things.
But we all know that the market is going to soften. And I think it comes down to the simple things from my side of things. I think companies should be talking about it more and not in a kind of like, Oh, the market's going to soften, the market's going to soften, but simply say, actually, you know, what is our plan?
Like how many CEOs are going to kind of their underwriting heads and saying, okay, guys, great work. You we're doing really well, you know, we're profitable. This is fantastic. We know the market's going to soften. What are we going to do about it? Like actually, whilst times are good, let's prepare for when times are worse and actually come up with a plan.
Because in reality, I think a lot of insurance companies wait until they start to see that loss ratio deteriorate before they start doing things. And at that point, it's much harder to come up with a plan. So I think kind of coming up with a plan when times are good, having actual like grown up conversations about that whilst also enjoying the good times I think making sure you've got the right kpis in place to identify where you are in the market You know, it's no good using your loss ratio Um or kind of you know The profitability of your portfolio to kind of identify where you are in the market Because quite frankly other insurance companies will be using more forward looking kpis And they will be selecting against you while the market is softening and you will kind of Not perform well.
So I think kind of getting yourself in order, getting, getting set up. So you've got the tools to track that. And yeah, I think kind of in, you know, to kind of come back to the, the theme of this talk, investing in innovation now, like, I think often it's, it's easier to invest in innovation when things are going well and investing, because if you don't invest, your competitors are gonna invest, you will be selected against over the coming years.
And that will just add to kind of the, the misery of a soften market. Whereas if you can invest now, While profits are good, while there is the opportunity, hopefully you can actually soften the fall of the softening market for yourself.
[00:35:55] TEC: Yeah, makes sense. And it also brings a background to what we discussed earlier, which is having the data and understanding the data better than your competitors so that you can recognize when these things are coming and make the right decisions. Brilliant. Well, thank you very much, Jamie. That's, that's been a really good chat and really informative.
So thank you very much for joining us.
[00:36:10] Jamie Wilson: Thanks, Tom. Really glad to be here. Thanks for letting me join in and yeah, as always, a pleasure.
[00:36:18] TEC: That concludes another episode of TEC Talks. If you enjoyed today's show and want to find out more about the topics discussed, Head over to hyperexponential.com to gain access to a range of resources relating to this episode. The link is in the description, and of course, wherever you're listening to the 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!