In this special episode of Making Risk Flow, host Juan de Castro engages an esteemed panel of insurance industry experts during InsTech’s recent Beyond Theory event in London. The panel features James Wright, Head of Technology at Beazley; Henrik Bjørnstad, Managing Director of National Markets at Markel International; and Jason Howes, Chief Transformation Officer at Allianz Insurance. Together, they explore the evolving landscape of commercial insurance, with a particular focus on digital transformation and its implications for underwriting processes.
Juan prompts the panel to candidly discuss their transformation journeys, uncovering the distinct challenges each company encounters as they modernise operations and enhance customer experiences. James details Beazley’s efforts to centralise operations for better SME client service, emphasising the importance of tailored underwriting for diverse risks. Jason outlines Allianz’s ambitious strategy to digitise and simplify its services, underlining the need to meet rising customer expectations. Meanwhile, Henrik illustrates how prioritising workflow improvements has significantly boosted productivity and customer satisfaction in underwriting, especially in speciality lines. The discussion also delves into strategies for fostering a cultural shift among underwriters to embrace innovative digital tools.
Listen to the full episode here
Juan de Castro: Welcome to another episode of Making Risk Flow. This episode is slightly different. It is a recording from an event in London that we organised in partnership with Google and InsTech, and we invited three of the most forward-looking commercial insurers: Allianz, Markel and Beazley to share with us the transformation strategy and how they are deploying digital risk flows. We got deep into the details of their target underwriting workflows, their tips on how to effectively drive change and also the challenges they face along the way. We wanted the session to be informative and avoid generalities, and the feedback from the over 200 attendees was exceptional. Hope you enjoy it as much as the audience did on the day.
Welcome, everybody. I'm joined by an amazing panel. I'm going to let you introduce yourself, but I've known James, Jason, and Henrik for a long time. They are some of the people I respect the most in the industry, so I thought it would be a fantastic panel. The only thing they asked me when we were preparing for the panel was, Juan, what do you want me to say? And all I've told them is, I want you to be real, I'm quite fed up with these panel discussions where everything is rosy, everything is perfect. I think let's get real. I think that is the only way that through these types of sessions, we will make an impact in the industry. So with that, let's do a brief introduction. James, do you want to introduce yourself?
James Wright: Hi, everyone. James Wright from Beazley I look after the digital division. We primarily focus on SME clients.
Jason Howes: Hi, everyone. Jason Howes, Chief Transformation Officer at Allianz UK, responsible for all transformational change across all of our brands and across all of our locations.
Henrik Wærsted Bjørnstad: I'm Henrik Bjornstad. I am the Managing Director of what we call National Markets at Markel, which has regional offices across Canada, Europe, and the UK. And like James, we predominantly look after SME clients.
Juan de Castro: Thank you. We thought this composition of the panel with three global insurers, but focusing on different areas of the market, could be a good representation of the different objectives, challenges that different insurers face. So the three of you have some sort of transformation programs in your companies. Let's start first at the very high level. When you think about your transformation programs, what are you trying to address and what are the main objectives? Perhaps start with you, James?
James Wright : Sure. So we created a brand new division about three years ago, which is Beazley Digital. What we did was we centralised operations technology and underwriting into one team to better serve our brokers and clients for the smaller end of the market. For most of us who are in maybe large complex businesses, the irony is we were underwriting large risk or small risk, the same as large risk, which didn't make sense for anybody really. Part of that journey meant really two things. One is meeting our brokers where they want to trade. That was really important to us. And email, which we'll touch on obviously, is still a predominant channel in some countries for doing business. So that was point one. Point two is the fact that really to be digital, response times back to the broker, and delivering exceptional customer services via those digital channels, Also need to be first class. That was another key objective. Really what we were trying to do is improve access to the large complex products that are traditionally offered to large corporate clients, to the smaller clients that are still specialists, but deserve to have the coverage and service that we give our larger clients. That's what it was all about.
Juan de Castro: Okay, and perhaps just before we get on to the other two, just to bring a bit more to life of what you just said. When you say you were trying to avoid underwriting all types of risks in the same way, so basically different operating models for different types of risks. Correct. Can you give us an example? So we started working with you guys, for example, in the SME cyberspace, which is very different from your London market type of risk. So can you explain to us what are the challenges in underwriting cyber SME?
James Wright : Sure. So we'll start at what was the objective. Our goal, which we've not yet met so we're not speaking from a position of done, was to automatically quote an email submission coming to Beazley using technology, zero touch. And that application form that we receive can be from absolutely any market, Chubb, AIG, Beazley, on anyone's paper, receive that submission, and then obviously automatically quote it. The first part of that is to work out, is this for a complex risk? Or is it for an SME business? And to your point, Juan, that's where really the operating model starts to separate. The point at which that has to happen is the point at which you receive the application form. That was all being done by outsourced teams in low cost locations. And to respond within the SLA we wanted to respond to, that was not going to work. So that first decision point that we need to use data is understanding, is this a complex or simple risk, which isn't just revenue. There are other dimensions to that. It could have to do with the amount of infrastructure that we can see that client has. It could have to do with the location that that client is in. It could have to do with the kind of control factors that the client uses to protect their security. All those decisions needed to be made digitally upfront to kind of split the operating model.
Juan de Castro: So you're trying to do a triage of the risk as almost the first step in the process.
James Wright: Step one, yeah.
Juan de Castro: Step one in the process, in an environment where you're on the right, you don't have capacity to probably go through every single submission. And for those people in the audience who are not familiar with cyber, just give us a sense of what it takes to triage a cyber risk?
James Wright: So point one is obviously to understand who the client is, location, rough size of the client, a proxy for that's revenue. The second phase is then to understand, okay, what kind of technologies have they got in place? And a big part of it is an understanding of what type of data might this client be holding in their infrastructure, which is another big factor. So it requires a degree of investigation, I suppose, to make that decision. Some of that is still human, but we've obviously digitised the large extent of it.
Juan de Castro: Okay, we'll get into the details of that in a few minutes. So Jason, similar question to you. What are the main pain points and the objectives of the program?
Jason Howes: Exactly what he said. So, no, Allianz, and again, it’s in the public knowledge that is trying to transform across the world. Okay, it's been on a journey for a few years. The UK is an absolutely crucial part of it. And I mean, it's three main objectives. It's digitalize, modernize and simplify. Alison, quite a lot of what you said resonated with me because actually, I don't think it really matters to us whether it's a really small client, a retail shop or an individual buying home insurance right the way through to Google, who we were talking here today. Everybody wants similar things. And what I mean by that is they want clear communications. They want ease of use. They want things on their phone. They want things on their iPads. They want the ability to talk to people if they want to talk to people. And what's been really interesting, I think, for Allianz is what's driving our transformation now across the UK is putting the brokers and the customers right at the design thinking of that. And that's really different, I think. And a lot of the insurance industry, to be fair, is doing exactly the same thing. And that's not always been the way. And I mean, for those that have worked in insurance for a while would know it's quite a traditional market, particularly like commercial insurance. And I think it's really and you've seen a massive kickstart in transformation now across the UK and in Europe and the US. And it's really driving this simplicity point more than the other two even. It's just people want to do things quicker and smarter. People want to do what they do on Google and Amazon. And it doesn't matter whether buying insurance or buying commodities or whatever. So for us, that's what we're trying to the highest possible level design and deliver. And we're doing it across the whole value chain, as we call it. So from underwriting to operations to claims and everything in between. And that's where we got introduced to Cytora, of course, at the start of that journey, where obviously we had lots of pain points where people were sending things to us. And really, it was taking us a really long time to actually digest that information, ingest it into data sources that we could then move across different places, bring it back together. Sometimes we were offshoring and using our offshore colleagues as well to input data into other places as well. And then by the time that had done, the feedback from all our brokers and our customers as this was going on is you've missed the boat. And again, that's just, I think, the pace of the world and people's expectations changing. Because that was Ok 10 years ago. So this is the reason why Allianz is embarking on the transformation journey. And we're doing it on multiple fronts, from personal lines to commercial to pets for those people who've got their dogs and cats insured with us as well. So lots going on, lots of change. And I think the next sort of four to five years are going to be huge for us, really.
Juan de Castro: So again, very different companies, but response time. So being responsive to broker requests, things like quotes and around time is something really important for you.
Jason Howes: Yeah. And you've gone from a world of days was okay. Like for those that remember. Remember the fax machine again. It was ok to leave it on the fax machine for a while. Then email came along. It was OK to leave an email unread for a while. Now it's not OK. Now people want things within 30 minutes. And the broker market, if you just take commercial insurance, there are thousands of different forms of data coming in in different formats, things called different things. And what technology is doing is enabling all of us through some of the smart stuff that we're seeing. So actually now take that data, turn it around and do really cool stuff with it. Not just in. I mean, I'm a huge fan of AI, and everything, because that's what we're all talking about. But there are some brilliant technologies out there that are coming through from places like San Francisco and Silicon Valley at the moment. Some of that stuff is just amazing. And this is where I think all of us big insights into the future is like, I think the acceleration we're going to see now over the next two to three years on transformation will probably far exceed what we've seen in the last five years as well.
James Wright: On the expectations piece, it's gone mad. So on the SME cyber market, which is, I'm talking globally, predominantly the US, if a risk of ours refers, because it needs to be checked for something, the chances of that binding is next to zero.
Jason Howes: Yeah.
James Wright: This means that brokers have access to markets which are instantly quoting and they're taking it somewhere else. So in the last couple of years, they've shifted quite a bit.
Juan de Castro: It's gone from like weeks to days, to hours to almost minutes. And so on that point, Jason, can you share with the audience, have you made progress in reducing your quote turnaround?
Jason Howes: Yeah. And you guys have been a great help in that. And along with some other parts of our sort of IT ecosystem. But we've moved from sort of five days for a commercial quote turnaround. Obviously, it's very different on personal lines and other lines of business. But if you take commercial, five days was not unusual turnaround, put it in all your internal systems, make a phone call to one of your broker partners, and then just start to have a conversation. And that's now down to three hours, depending on the complexity of it. And it will continue to reduce because that's not happened overnight. It's a constant reduction. And the other massive thing which we've seen, which is really exciting, is that those five days before you landed with a presentation, let's just take the Google example of the building across the road. There's the building. There's the presentation. The broker sends it to us. We then start to read it. Now, you enrich the data that comes through on that presentation, with potentially hundreds of data points. So actually, by the time it hits an underwriter's desk or whoever's going to look at it, you're actually in such a much better informed place to actually have that conversation to start with. So again, it speeds everything up. And again, this is publicly available information. But in the past, to actually bring that together, to enable your machine learning to run its eyes over it and then start to give you some thoughts on it, it's quite amazing, again, as these machines are learning what we're starting to see.
Juan de Castro: Especially as these are like mid-complex risks, right? Risks with potentially hundreds or even thousands of properties.
Jason Howes: Yeah, I mean, yeah, if you take a large real estate, somebody owns 50 properties in London, you used to sit and go through every single property and work out against the postcode what the flood risk was, what the fire risk was. That land's done now. So, that's in the space of three to five years. And this is what I mean about the acceleration and the excitement that I think is coming to the industry for all of us. It will really help us to accelerate and give much better service and much better output to clients and brokers.
Juan de Castro: Makes sense. Henrik?
Henrik Wærsted Bjørnstad: I think just to give you an idea of our operations, we're quite a niche insurer. So, for example, we are one of the leading insurers on care risks. So think about daycare facilities, care homes, nurseries in the UK. So it's quite niche. We don't have huge teams, but we have around 800 people spread around a lot of locations. So that's our model. So, for one, we have worked with Cytora since before your time. So it's been a bit of a journey for us but two stories I want to tell. So when I go back to five years ago about that, maybe a little bit more when we started working with Cytora, it was all about data. And I wanted us to be so good at underwriting that people said, wow. And we got a lot of amazing data on these care homes. What type of customers have they had in the past? What did we know? Flood risks, all of that stuff. And what happened was after that project, the underwriters, what they really wanted was, I like that extra information, but actually just getting that report in front of me instead of clicking another link. So the benefit from that was more on the workflow side than on the sort of brilliant intelligence on the underwriting. So it kind of shifted the focus away initially from data to workflow. And I think that's really important to what Alison said, because we have since then changed. What we're trying to do is to be easy to trade with, which has a relevance component. It has a service level component. So what we say is that in the niches we play, we have to be leading across those. aspects. The other story I want to mention is that I came from outside the insurance industry. So I've been in it for about 10 years and pretty early on, and I didn't see it in my own eyes, but I heard about it. Someone said, you know that underwriter, he just photocopied his iPad to be able to store the information somewhere. And he sort of told me that, well, it doesn't always have to be that fancy to be impactful initially. So where we then took the next phase with Cytora in this case was that we looked at the smaller end of our business, that tends to come through some sort of web portal or email and not necessarily underwritten online. But the workflow aspect of that was quite poor. And Cytora really helped us with that. And not only did that improve how quickly underwrites could get to the risk, but it also gave us much better data. So it allowed the management of the team. We came with a totally different level, which was great. And then the next phase of that was to be able to prioritise the risk better. So for the underwriters, I mean, productivity went up significantly. It had a whole sort of cultural shift in that team, which was really great to see. And now we're trying to roll out all the parts of the business.
Juan de Castro: Well, given that you've mentioned the productivity uplift, by how much did it increase?
Henrik Wærsted Bjørnstad: Significantly.
Juan de Castro: In order of magnitude.
Henrik Wærsted Bjørnstad: Order of magnitude very significant.
Juan de Castro: So I'm sure everybody thinks this sounds fantastic, but how have you achieved that? So perhaps it would be useful to do a couple of minutes overview of what the underwriting workflow looks like in some of your teams that have enabled you guys to achieve this level of impact? And perhaps we'll start with you this time, Jason.
Jason Howes: Yeah, sure. So probably I'll choose two risks just to sort of bring it to life. So let's just take a motor fleet case. So a client's got a hundred vehicles and that motor fleet case comes into us. And like I said, the old world was very much, it would then get logged and data goes everywhere. Now, what we've done at the front end is really streamline that data ingestion, obviously using technology like OCR as best as we can to actually make sure that we're not touching where we don't need to touch it because that speeds everything up. Pretty simple. So it comes in, it then feeds the systems that we need or reaches out to the systems and starts to swap data. So for example, have we seen the risk before? Have we not seen the risk before? Because again, there's nothing more annoying than having a conversation with someone who said, I sent you this last year. It's like, did you? So it's just basic stuff like that, right? That you can start to overlay data points. And I know we're going to hear a lot about data today, but it really does drive then the quality of the conversations that come out of it. So in the journey then, our big, big sort of focus was getting it decision ready. That's what we call it. So that's because you want it to land with a human to look at it and to actually then be able to do something with it rather than, and this is an industry wide problem that's existed, is that there was lots of administration that that underwriter or person would do. So once it's decision ready, then you're into the realms of trading and you're into proper conversations with brokers meeting the customer where possible. so, it becomes a much richer experience for everyone involved so that's where we are now. And obviously then that flows through to quotations, premiums put forward, letters issued, contracts, certain letters, of course, as per the regulatory requirements. And then off the back end of that, we liaise back with our broker and we bind the risk or don't bind the risk. And then we update the various systems after that, which drives lots of MI and lots of data. And obviously everything now, that in the past you would have probably thrown that away. It would have just disappeared either into these disparate systems, but that data is then available then to actually make you much better informed decisions as you go forward. So even if you don't win the risk now as a new business case, actually there's a lot of learnings to be taken from that, that you just wouldn't have had before. So that's the motor side of things. I would say the property and casualty side of things is very similar, but obviously very different as well, because you have many, many more data points on the property and casualty side. Exactly what we're saying, just using the Google building as an example, from fire details through to how far away is the fire station, through to flood risk, ground flood risk, everything else that you can possibly imagine. And again, all of these data points, which weren't available before are now available to be able to pull into, to create that ability, to get that as onto an underwriter's desk, to then to be able to have a proper conversation. And that's the journey that we're on at the moment. I mean, there are lots of parts to that. And I won't talk about claims unless people have particular wants and needs to cover that today. But that sort of takes you through two commercial journeys if you like.
Juan de Castro: And to some extent, that actually links back to what Alison was saying in the opening session, which is how do you use data to inform a more granular risk selection? At the end, you talked about the first step is to be able to capture the data from the whole intake, which allows you to be much more granular in your risk selection, which, if I understood correctly, this is what Alison was referring to.
Jason Howes: Yeah, that's exactly right. I mean, the ability with data cloud now and everything that sits around it and the vastness of the data at our fingertips, if you get that right, it's really, really powerful. If you get it wrong also, because you have massive problems, and I think we've seen that in the industry as well. But if you get it right, it's really, really powerful. And that's why I think that probably my colleagues here and most people in the industry would say that the focus on getting data at the heart of what is coming out to create the right conversations with brokers and clients now is where everybody's driving their sort of quotations and their renewals.
Juan de Castro: Perhaps you, Henrik?
Henrik Wærsted Bjørnstad: Just a comment on that, allowing the broker conversations. What we saw in that first phase of trying to get all this risk information is that if you start at that end, it can be quite tricky to communicate to clients and brokers why your underwriting decision ends up in a certain place. What we see now is instead, if you start with the workflow, the underwriters understand how that works, and then they ask for more data, which is a much more customer-friendly journey.
Jason Howes: Yes.
Henrik Wærsted Bjørnstad: So with us, what we do, we get these submissions in. They appear in the CRM, which is then auto-populated by Cytora, and then we get the risks prioritised based on some external data and some of our internal data. For example, what type of risk are we going to say no to because we don't know much about them? It's much better to let clients and brokers know as soon as possible. Where do we see success in the past? And we prioritise those first. And that's really the reason when you talk about productivity. For me, it's not really the point. We don't do that to save money. It's really to come back to customers faster. So a more relevant measure is to see how the Net Promoter Score goes up from that. And as we mentioned that, you know, particularly on that responsiveness, that has gone up quite significantly with the type of clients and brokers we see success with. That just has such an impact on the teams because people want to think about new types of risks they can take. They want to have fulfilled jobs, and it doesn't feel very good to come home and have to go to the office and sort of have traded that with an admin task. So personally, that's the biggest, most fulfilling side of all this.
Juan de Castro: Yeah. Sometimes we think these need to be like groundbreaking operating models for technology. At the end, your underwriters are looking at the risk in the CRM. I think the difference is they've got the decision within the CRM within minutes of having received the submission so that they can get to the broker much faster.
Jason Howes: Yeah.
Juan de Castro: Right?
Jason Howes: Just going to add one point because you mentioned culture earlier and people as well. I think we should be really clear. This technology doesn't replace people. What enables you to do is the stuff that you want people to be doing, i.e. Going to see Alison. Just to be really clear because you said earlier, which is one example, the feedback we have from clients of all sizes, even if someone's spending 5,000 pounds, it's a lot of money to somebody. Actually, people want to be seen. They want to be heard. They want interactions. Brokers want more and more interactions. There was no flex in the old model to enable you to do. Everyone was working at 1,000 miles an hour with legacy systems. Mainframe, old technology. So actually the new technology is really crucial to do this point that you're saying. I think the people element, I will perhaps touch on it again later, is really crucial because I don't think people will go anywhere. It's then the quality of the people trading that's going to be the difference.
Juan de Castro: Great point. Let's hear James' point on the workflow, and then we'll get back to culture and people, which I think is a very important point.
James Wright: The service piece is real, but if there's anybody in the room that's not close to their customer, the baptism of fire for me was a US roadshow early in the journey, and I thought it was going to be great fun meeting brokers and having dinner, but I spent the second meeting a whole hour being shouted at, the diabolical service from our team. I think that was a real moment of going, okay, this is a real problem. We have to fix it. Anyway, I'm not going to draw to the flow diagram, but just to let you know, the way we do it is we have a classification engine at the front of the email ingestion, just in case we don't want to use Cytora anymore. We haven't quite given you the front door. It goes into the classification engine, and if it's a cyber new business submission, it extracts about eight key fields, business name, URL, address, et cetera. We extract those bits of data because it kicks off in parallel a vulnerability scan of the client's domain name. To give us back more data on assessing the risk, and because we only offer the quote to one broker, first in wins, we then reserve the market for any more submissions that come in. So we take the client name and we automate the clearance check. That's the first pass. Second pass then, once those bits are done, and worst case, it's three minutes for the vulnerability scan, we extract all the other data from that cyber application, and it hits, at the moment, a person in the loop. Boo, very disappointing. The reason we're doing that is, that what we didn't want to do is port quality data in front of our underwriters, only for them to have to go back to the original submission because they don't trust it. That would have been an epic fail, and it would have been a fail for Cytora and for the program. So we have someone quickly cross-check the screen in Cytora which gives them a confidence score on data, e.g., the business name is Bob's Golf, but the email address is something completely different. It's an obvious fail. That would be flagged as low confidence, and that person would correct it. Once that's done, that then gets pushed into the quoting platform, where the underwriter needs to do absolutely nothing. They can trust the data, and either hit quote or decline. We are at a point where we could automate that last step. There's a couple of reasons we can't. Without this process, we'd have no idea. Bits in the application forms are missing or always incorrect. Two major problems. One is identifying the business activity of the client. Brokers may not know, so they often leave it blank or put in something like technology company. That's not particularly useful if you're writing a risk because obviously there's lots of variation on that. So we're using some third-party products which are powered by AI to enrich that data. However, for those of us who are in regulated, controlled organisations, using AI-generated data to make an underwriting decision, is there any regulation? No, no. Might not be what your compliance or regulatory team want you to do. So again, that's why someone has to look at that and just press the button. But we are at a point where we've got the auto-decline and quote rules. What we need to do is, which is where you're going in the conversation, I think, is take a few more people on the journey so they can understand what are we doing between receiving this very bad piece of data and, of course, making it better, such that we can trust it to offer terms to a client.
Juan de Castro: So let's do a final round on culture and how to get underwriters on board, etc. That's often been one of the biggest hurdles. So perhaps we'll start with you again, just because you were mentioning this already. So what are you doing with your underwriters to get them on board through this process?
James Wright: They are super engaged now in looking at what they receive from the platform. And when it's incorrect, helping us correct that when that comes back in as learning data. They have had to up their digital acumen quite a lot. Because I think the thing we're losing in doing this is underwriters touch risks. And it's not a gut instinct. It's not gut feel. But they are learning machines. So they start to know what's risky and what's not. And they know what to look for. And as you remove them from the process, they're really anxious about how do I portfolio manage on that basis. To do that, you've got to up their digital acumen using data. And that is a journey. So we're still at that pivotal point. I think they're trusting the tech. They're now realising my job's changing.
Juan de Castro: I think one of the things you did, I thought was very smart is in the initial stages, you provide the underwriters with the decision ready data, but also access to where the data came from as an interim stage, right?
James Wright: Yeah. If they're curious, they need to see the data lineage and understand it.
Juan de Castro: Jason, similar question to you.
Jason Howes: Yeah, so what we found fascinating is similar in terms of the human side of things in terms of if we leave them to get on with it, the error rate we think was 40-ish percent. So 60% was about right between 60 and 70, and then the error rates would come through at that type of level. What we found with Cytora, we're structuring the data. And as you know, we've moved into 90% plus pretty much across the board in terms of accuracy of data coming in. But you still get scepticism from underwriters, even though you tell them that the percentages were here and now we're here. So I do find it fascinating that culturally we've still, I think as an industry, got a long way to go in terms of getting everybody onto the same page that we need to trust the data. We need to go with it. It's okay. To not actually have all of the reams of information and the 300 page document that backs up everything that they do. I think we've made a lot of progress and I think overall we will get there, but I still think we're a way off. I don't think there is much more as an industry we can do around education. This is about people now embracing the technology. But again, what's fascinating is people trust Google Maps as we all do, yet they won't trust some of the data. that is being fed to them from places like Google. So there's this bizarre industry wide scepticism on some of this. And I think it's just this shift than the acceleration that I talked about earlier of people really getting their head around where this is going and how fast.
Juan de Castro: And similar, but perhaps the angle that you could take, because I always love this story that you tell, Henrik, about the cultural shift in your lead's office. Perhaps you do want to touch on that?
Henrik Wærsted Bjørnstad: My favourite is Birmingham. I think what surprised me initially when we started working together was how much time you spent with the underwriters. Just not talking about technology, just understanding what they're doing. And it was so important because what happened was you got that feeling from the underwriters that you understood what the pain points were. And that allowed us to get that first win. So they understood what was happening and they understood that we as a company, by working with Cytora in that case, we could actually make it happen, which has led to several internal cases we haven't worked with you, but still made improvements afterwards, which I think was a really great partnership. And frankly, it was one of the things I'll remember for as long as I live, I think, that coming into that office one year after we had started and just feeling the buzz and seeing the head of that office won global award and had to go to our headquarters in Richmond at the shareholders meeting, sit on the first row in front of a thousand people. It was a very nice story. And that was the start of a cultural shift, and we have so much further to go on this journey. But that was a good start. So I'd say the one thing is really that in the beginning, just overinvesting in the engagement part was my learning.
Juan de Castro: Thank you. Should we take a few questions from the audience?
Audience: I’m really interested in what you were saying about the incorrect data that comes in from the brokers and how you correct it so your underwriters get better information. How about feeding that back to the brokers and the client so that we actually have our data enriched? Because we don't see that happening. And sometimes we're not aware that the broker is sending out faulty data either. So where is the role of the insurer in ensuring the client actually has better data to share in the future?
James Wright: We're always a bit anxious about data sharing in this industry. You probably know. We do send that data back to the broker. So it's not hidden. I think it's a point of commercial friction because we want them to do business with us. And we believe that's great customer service. So it's very difficult to sometimes have that difficult conversation of saying you've got to up your game on this. But yeah, we do provide it back to them and they'll see that adjusted.
Juan de Castro: It's the broker's fault, basically (laughter).
James Wright: Long-term, we need to get, you don't want to hear this, but we don't want email submissions, really. Not whether ones, But that's the world we're in.
Jason Howes: Well, just one thing on that. We talked about technology a lot, and there was some fantastic tech. We get a lot of pubs, restaurants, that kind of stuff. It's a great example. And a lot of times it just says pub. And actually, a pub can be very different in the UK. Some of them have got dance floors. Some of them do karaoke. Some of them do a lot of other stuff as well, and so actually what we found is with some of the technology advancements is that by putting them through some of the models now, they can scrape the web and we can actually, through the photos that people just post, we find ourselves going back to brokers and saying, you said it was a pub. It's actually a karaoke late night bar. And that's a very different risk, right, that you've got a bunch of people turning up to maybe have a quiz night versus a raucous dance floor with karaoke belting out on a Saturday. And it is completely different. Different rating factors. Different things. So, again, it's a positive example of where technology is really helping insurers to sort of exchange data so you can actually have those conversations. And this is all available data, right, on the web.
Juan de Castro: Thank you. Any more questions?
Audience: Thank you. I'm just interested to understand from the perspective of larger corporations that are going to be very heavily data-led, what lengths you've gone to to understand your pain points directly from the underwriters themselves, who are not always being a more traditional role. Naturally, they might not be as aware of the new technologies, as you said, sprouting up left, right, and centre to really change their roles, change the efficiency. So just keen to understand how you sort of approach that, how you keep their voices prioritised in your business, especially when you have so many underwriters. Thank you.
Henrik Wærsted Bjørnstad: So I can start to do spend and my management team quite considerable time travelling around. And when we do that, we always have at least one hour where we just take and write in, ask no agenda, just to tell me about your day, basically. And that is as fluffy as it sounds, extremely informative. And then where we have call centres, I personally once a quarter perhaps listen to calls. But a bit to my previous point, it is sometimes tricky. Because I don't have time and manage to often not have time to really get into the details. And that's where working with a vendor like Cytora was very, very helpful.
James Wright: My biggest piece of advice on that, well, I know this is not easy, get the right people in the room and come up with a shared objective that isn't talking about the solution. From an underwriting perspective, it might be a premium target or a quote to bind ratio or a quote ratio to underwriter ratio. Have that as the North Star and then warn them it's going to be hard, but we need to be empowered to experiment with different technologies. Because when we started with you, for example, LLMs were not around. And in that journey, that became a new way to achieving actually the key result we were looking for, for a specific part. So you've got to lift it up a level, get everybody's shed. Is it cost saving? Is it speed to client? Is it reducing costs? Everyone might in their head have different ideas about what the thing might do. And that's a very dangerous place to start. So start with that shared goal.
Juan de Castro: Time for one more question.
Audience: On your transformation journey, what sort of attrition rate have you seen of your staff that photocopy their iPads, that sort of thing? So partway through your journey, what sort of percentages are you seeing across your businesses?
Jason Howes: So in terms of transformation and the programs itself, we've actually seen a relatively low attrition rate, which has surprised us. And I'll tell you why it's surprising, because we thought that you'd get a lot more fatigue in these really large programs. But what's been fascinating is that the people that tend to really want to be part of these programs, the feedback from them is they're driving the future, they're driving the change, and therefore they're desperate to get it over the line. So actually, we've been on the journey for sort of three years already. And we're at two, three, 4% touch wood when I say this, that we don't have a mass exodus now. So we've been really, really pleased with that. I think in the wider business, we have seen spikes at different times, pretty similar to I think what the industry has seen, as some people are thinking that the digitalization, modernization that's happening at one insurer might not be happening elsewhere. I think that if you would speak to those people now, and we do, and we check in with what everybody's saying, I think it's pretty happening industry wide now. So I don't think there's anywhere to hide. Everyone's just in a slightly different place on their transformation journey. And I think people like to be in slightly different places, depending on where they are in their career. So it's a bit of a mixture between those two scenarios, but we haven't seen masses of change in this sort of rates over the last few years.
Henrik Wærsted Bjørnstad: Very similar to us.
Jason Howes: Yeah.
Henrik Wærsted Bjørnstad: Maybe, so certainly on the underwriting side, if I go back many years, maybe if you think about the sort of old style IT organisation, the way we used to operate that versus how we operate now working more with partners, there's maybe a change in skillset, but not really a sudden attrition, I would say.
Jason Howes: Yeah, we have hundreds of applicants for these roles. I don't know whether you've found it. If you advertise transformation roles and these change we have hundreds of applicants from outside of insurance now as well, which honestly was never heard of. So I think in a really positive way, so then in this particular point, there's a lot of excitement, I think, just simply because I think the bar is relatively low in the insurance at the moment. There's so much scope to improve it. And there are so many people that wanna jump on that exciting journey.
James Wright: Europe and UK stable, US SME cyber markets are barmy, and we've had a few ups and downs with newly short tech MGA's appearing, disappearing and various changes in that space. So that's where we've seen some attrition.
Juan de Castro: Thank you so much, Henrik, Jason, James. Thank you for accepting to join us today. Thank you.
