.webp)
This a shortened version of Making Risk Flow podcast, episode: “The roadmap to deploying end-to-end digital risk flows. In this episode, Juan is joined by both Neil Peters, the National Director of Branch Network at Arch International UK, and Arvind Drubhra, the Director of Operations for the UK and Ireland at HDI Global SE. During the course of their conversation, Juan, Neil and Arvind discuss how to leverage data to drive growth, the steps and target states to achieving full digitisation, and how technology might allow businesses to accelerate the onboarding of new underwriters.
Listen to the full episode here
Juan de Castro: Welcome everybody, to another episode of Making Risk Flow. I’m joined by Neil Peters, who’s the National Director of the Branch Network at Arch Insurance, and Arvind, who’s the Director of Operations in the UK and Ireland of HDI Global, today we’ll go through the visions for the businesses’ priorities, and target operating models. I think we want to get real in this discussion and really answer the question of where do you start deploying the Digital Risk Flows. Like, how do you start capturing value?
Perhaps starting with you, Arvind. Let’s start with an introduction of yourself. So if you can introduce yourself, give an overview of HDI Global and the different geographies where you operate.
Arvind Drubhra: My name is Arvind Drubhra I’m the Director of Operations for the UK Island branch of HDI Global. We operate predominantly in the large corporate space, writing the industrial risks. The HDI Global is headquartered out of Hanover with branches around the globe. I’m an operations change transformation professional, been in the insurance industry since I graduated in the mid 90s and worked at various technology companies as well as large PMC companies of a similar profile to HDI, the previous one being in Allianz Global Corporate when I was there 16 years in a variety of different transformation roles.
Juan de Castro: Thank you, Arvind. Neil, do you want to do the same, give an introduction?
Neil Peters: My name is Neil Peters. I’m responsible for nine branches across the UK. Geography is very important to us, they’re not particularly big branches, so we have that challenge. It’s actually harder to run a small branch than a big branch, in my opinion. So you’ve got kind of that dynamic going on. Our target market couldn’t be any different to Arvind’s. It’s UK SME, we write predominantly commercial line business, property owners. We do fleets and motor trade. So that’s predominantly what we do out of the branches. Those are our core products. Technology is becoming increasingly important to us because it’s all about customer service. So that’s the main focus of what I do on a daily basis, really making how those nine branches trade with brokers in their local geographies.
Juan de Castro: That is one of the reasons I thought this was interesting to have both of you in the same panel, because your types of business are different from each other. Let’s start with you, Neil. You lead the branch network. What are your objectives? What are the pillars that are going to allow you to drive growth in those branches?
Neil Peters: We’re a relatively new entrant in a very congested marketplace. Differentiation is quite difficult. Certainly, there are bigger names and more recognisable names than us in the marketplace. So we’ve decided that we wanted to focus on two specific things. We wanted to be as local as possible to the brokers that we traded with, hence the nine offices. Of course, those brokers are big brokers. You got the household names, the Gallaghers and the Tower Gates of this world. But we wanted to trade on a personal basis with the people that dealt with the client. So we’ve invested heavily in the nine branches, and I said differentiation is really difficult. So customer service has got to be it’s building those relationships, but also being really responsive and trying to be as accommodating as possible to those people who we choose to deal with and who are supporting us.
Juan de Castro: When you say that customer service is a priority and it’s going to help you differentiate and compete with the larger competitors, how would you define that customer service?
Neil Peters: Probably the single word that describes it best is to be responsive. We’re not going to be all things to all people. We’re not like any insurer. We have our appetite, our competitors have their appetite, and they don’t match up. But I think a quick in and out. So get on the phone very quickly and establish that we can do what they’re asking us to do and that we’re interested in doing that, or tell them very quickly that we can’t and take us off the list and move on to the next thing. So we’re quite pragmatic in that respect. So customer service can be misinterpreted as somebody that is over accommodating. That’s not the plan at all. The plan is to be really clear in articulating what we can do, what we want to do, and that we want to do it. On this particular risk on this submission with you today, I guess it’s being responsive and being clear.
Juan de Castro: When you talk about being responsive, would you measure that in terms of first contact with the broker?
Neil Peters: So I think first contact is really important because you lay-out your store very quickly as to establishing that first contact, whether you can and can’t do it, and then after that, you establish what the broker’s timelines are. So I don’t think service standards are particularly helpful because an internal service standard defines what you think you can do, not necessarily what the client or the broker wants you to do. So that initial contacts, can we do this for you? Are you really interested in doing it with us? And what are your timelines that we now need to work to? What additional information do we need from you? And how are we going to be very important that we do what we say we’re going to do and try and stick to those deadlines so that first contact creates the service standard, if you like, sets out the timeline.
Juan de Castro: Arvind, does this also resonate in the type of industrial larger business, or is it something different?
Arvind Drubhra: It certainly does resonate. I think the key thing here is HDI, probably alongside many other similar carriers, is facing that challenge of how do you become more responsive to the brokers and the clients. So how do you raise your game, effectively? Yes, everybody’s striving for growth, but growth needs to come alongside being able to service and deliver the right appropriate levels of service to clients and brokers. So our focus, for example, HDI, yes, we’re on a strategy to grow the business and grow this certain market. If I look at it from a UK and Ireland perspective, my focus is around delivering efficiency into the organisation so that we can free up capacity on underwriters, so they can focus on the core risks. Now, how do we free up that capacity? It means that we’ve got to invest in the right solutions that allow those historic legacy manual type activities to be automated to a degree or to be taken away. And that’s really where our priority focuses, looking at those efficiencies and how can we deliver those efficiencies into the organisation in a reasonable time frame. Because these are things that we can’t sit on for years and years. And I’ve been in the industry long enough to see initiatives that go on right, in sort of similar companies to ours, where you talk about this area around delivering efficiency and delivering better service levels. But actually, for a lot of carriers, it takes a long, long time to get there.
Juan de Castro: I think you really need to be more efficient on one side to drive productivity, for sure, but even more importantly, to drive better service to your brokers and therefore support your growth, right?
Arvind Drubhra: Exactly. And that’s why I look at it from both lenses. Right, it’s about driving efficiency internally, but by driving efficiency internally, it means that you’re actually providing a better service externally as well. So you’ve got to be very cognizant of if we’re striving for growth, we’re striving for efficiency, we shouldn’t lose sight of the fact that what does a broker want from us in terms of service? What does the client expect of us and how can we react to those needs in a timely manner and actually improve and try and beat our peers who are also doing similar things?
Juan de Castro: And you mentioned some of the underwriting activities that can be automated. Can you give me a couple of examples of activities that your underwriters are doing today but in the future, you’re looking to automate?
Arvind Drubhra: Without a doubt. I mean, the obvious one for being the underwriting cycle is the initial risk intake process. The risk intake process of in terms of capturing onto our initial systems, doing the initial draft entry onto our internal tools. Those are the sorts of things that we should be leveraging new technologies to help us try and automate that process. So there’s a lot of new technologies that are out there in the market that can allow us to use things like NLP and AI to help streamline these so that we’re not building something bespoke to ourselves. But actually we’re in a marketplace where things are evolving, actually driving that type of transformation through the market so that other carriers, other brokers also in that same game to also change. So the obvious one, I like to say is that risk intake process, how can you automate that in a way where that data entry today, which is either done by sales distribution underwriting type people, which you effectively remove that so that underwriters are then focusing actually on the core task, which is reviewing high quality, relevant submissions and focusing on delivering that in a time frame to service the needs of the broker design.
Juan de Castro: Does that Neil resonate with you too, automating the intake?
Neil Peters: It’s the same challenge. There’s an awful lot similar. Despite the fact that the business models are completely different and the target client is completely different. There are so many synergies, aren’t there? We’re already using a lot of third-party data that comes in. The obvious one is flood mapping. It’s a hot topic in the UK in the last ten years. We’re using third-party flood mapping, we’re using financial checks, stuff like that. And all of that currently is being dragged out of some app or some system somewhere manually. So the ability to log, that’s quite important. So actually absorbing that submission into the business, but then also being able to do those checks and have those presented to the underwriter at their desktop without them having to go off and do searches on third-party equipment and sites, that’s really important because all of that is really informative. We can’t live without it anymore, we’ve become totally dependent on it. But that’s not a great use of a skilled underwriter’s time. That’s just admin. So being able to chunk that out and deliver that in a more timely fashion, in a simpler fashion off of the information that’s provided to us from the submission, that’s kind of what we’re working on. So we want as little admin being done by skilled people as possible, and we want them to spend their time underwriting and assessing risk and negotiating and talking to our customers and understanding our customers’ needs better. If we can get to those two things, that’s absolutely fantastic. I think it’s a long way off in certain areas, but some of that kind of dragging data from other places that’s actually available today. So I think it’ll happen in chunks and we’ll build on the platforms that we’ve got and then we’ll find better ways to use the platforms that we’ve got, but really getting data to underwriters so they can make those assessments. That’s really useful.
Juan de Castro: When you think about it, is it just purely the automation of the ingestion of the data or are you also thinking about once you’ve got the data in the system, how do you then accelerate the analysis and evaluation of a case?
Neil Peters: We have an in-house data analytics team. It’s exactly the same problem as we were just answering. We work off a two axis really. One is the quality of the submission by our criteria. So how we assess the risk, to give them an indication, it can’t do the job for them, but it gives them an indication that this is one to pursue and then we have a likelihood to win as well. So how good is our relationship with that broker on that product, for instance, in very simple terms, but that’s got to be dragged in as well. It’s the speed to be able to assess that you want to make that first contact with some confidence and that you’ve got all of the data that you need to be able to have a meaningful conversation as quickly as possible with the broker to establish finalise some information and move forward, or to take yourself out of the game. As I say, our rates range around the 50%, so half of what we see is not acceptable to us. Making that clear to the broker and getting away from that quickly is as important as to the point Arvind is making. We want to deploy our limited skilled resource in the best, most efficient way possible. So killing stuff off is really important. Working out what you’re going to work on is really important. And all of this data that’s coming in from other places is informing that all the time, right? So the slicker it can be delivered, the quicker we can just get on with the job really.
Arvind Drubhra: And I fully echo that because I think it’s getting the risk intake but it’s that enrichment process through that risk assessment that happens. So external data sources which underwriters today are logging into various tools and systems externally to get those pulling that all together and presenting it in a visual way but then actually applying what would be considered, let’s say the internal appetite rules, the prioritisation getting that. So actually what you’re presenting to the underwriter on the top of the screen then are actually already the filtered set of submissions that they would want them to focus on and it’s that sort of prioritisation view which really then starts adding value to an underwriter.
Juan de Castro: Often we talk about ingestion and automation there’s a component of just efficiency, of not having humans doing that. But isn’t the more value to the organisation really by doing that you can do a pre-evaluation of the risk early on before it gets to an underwriter, to make sure that the underwriter first of all they are working on the most attractive risk first. They can have that first point of contact with the broker that you mentioned, Neil, as early as possible for the ones you really want to win. It’s automation to enable a slightly different operating model where your underwriters are getting really much quicker to brokers on cases they really want to win, right?
Neil Peters: In the SME space, I don’t know if it’s true of Arvind, this is where we might diversify, but in the SME space there’s a lot of volume but the exposure is still there. So you could be dealing with a relatively small case and still suffer a 15 million pound loss. Right. The job still needs to be done, but you’re shifting through a lot of stuff and that sifting process is the more help you can get with the sifting process, the better because we still need eyes on it. It’s still a skill that we aren’t able to automate through our digital business. This is a step up from that, but it’s not really big because it’s not really big, there’s loads of it. So you’re constantly balancing, how much time can I spend on this without exposing the company and still doing a good job?
Arvind Drubhra: Certainly on our side, we’re seeing an increasing number of submissions because of the way the growth of the business is going. But then what we want to do is make sure that the underwriters are focusing on the ones where they see the priority focus should be on. So yeah, it’s sifting those out and that’s hence why the things that are already considered out of appetite, they shouldn’t go in front of an underwriter. Let’s get them in, let’s get them logged onto the system, let’s get the appropriate review and assessment done and then straight out the back door say, right, this is something that we’re not interested in. But the insights, I think going back to Juan’s point is once you start building up that repository of data and knowledge, it’s the insights where are actually other areas that we can potentially start generating, which historically we’ve been declining, right? So should we look at reviewing some of our appetite rules and some of our areas of focus to say there are certain potentially profitable accounts that in the past we would not have written because we would have been declining? So that’s for me then, the next level of insights that start coming through the sort of evolution of a digitisation approach.
Juan de Castro: You talk about prioritising risk and guiding the underwriters to the most attractive ones. One of the challenges historically was about defining the high priority ones. Do you define them as the ones most likely to convert? Do you define them as the ones that best fit your underwriting strategy? And often those are based on, I don’t want to say gut filling, but do you want to say on historical beliefs? But I think one of the things we’ve seen, for example, is as you start really applying kind of consistent rules to the way you prioritise, you see that often clients are prioritising higher risks that they never end up winning. So it’s almost like you need that real data really move from beliefs to like okay, you might think that a certain Tower Gate branch has a higher conversion version rate, but perhaps it’s a higher conversion rate on certain types of risks.
Arvind Drubhra: Fully agree, because when you look at the prioritisation there’s going to be various criteria and every company is going to have their own criteria around what would be considered as part of that prioritisation rule set. So things like treaty exclusions, for example, right? You’ve got certain CAT appetites. So every company is going to have their own rule set. And I think the power comes from using the correct tooling to actually automatically filter out things that you know are not going to be applicable in your business. Then start focusing on the ones that are relevant. So that to me is what the power comes from, because otherwise, we have underwriters today I’ve got things coming in, but I can’t quote that because that automatically goes against what my treaty allows.
Juan de Castro: Any thoughts on that, Neil?
Neil Peters: Amongst those multi-site brokers you’ll have different quality relationships, inevitably, because despite all the talk of technology, this is still a relationship-based business and we might have a fantastic relationship and be able to get rate conversion with one set of people, poorer conversion and less likelihood to win from another set of people who work for the same organisation. And that’s not always clear. Whereas if you get your data analytics right and if you get that the submission is absorbed and chunked and then put in front of the underwriter, you can warn them. Now it’s quite interesting because underwriters are very good at convincing you that they are going to pay something with you because they want a quote. So we’ve had various conversations where we’ve gone back to a component of the data and said well actually you’re not winning from this guy, this girl. You think you are, but you’re not because they’re really good at making you feel good about quoting to them but you never actually win anything. So it starts to challenge some of the norms and some of the perceptions of what is good, what is not, who is good, who is less supportive and it really refines your thinking and clarifies it as well. There have been some genuine surprises that people find. I thought our conversion would be really good with it. No, it’s not. And you are kissing a lot of frogs not to find many fruits and I think increasingly, although it’s really hard to find really good people as well. So we are trying to squeeze more out of the people that we’ve got and this year seems to be particularly bad. There does seem to be a real sort of glut of good quality underwriters in our marketplace. So again, that efficiency comes even more under the spotlight again. So what are they spending their time on? Because we’ve got the opportunity, but we just don’t have to necessarily have all the resources we need to do it. So can we get fewer people to look at more stuff and that comes into play. That kind of lays the light insight into, yes, you really are likely to win this and you really do want it early on, provides a different level of confidence that allows you to act with being more direct, if you like, with the broker “I really want this one and I’m ready to go”. In the old days you’d have to read through this and you’d have to get really comfortable with it. And then there’d be lots of questions. If you can get to a place where you’re already feeling very confident about something, that confidence transfers and then that broker wants to deal. So it really helps if you’re providing more certainty earlier on in the conversation.
Juan de Castro: In this environment where it’s becoming increasingly harder to recruit good underwriters. So one lever is to get more and squeeze more value of your existing team. But do you think this type of technology that allows you to prioritise and help underwriters will also help you accelerate the onboarding of the new ones?
Arvind Drubhra: I think there are a couple of things there. I think, yes, the role of an underwriter is inevitably going to change in terms of what is the core role that we want an underwriter to be doing in the future. If we’re talking about automating parts of processes today, focusing on the core value, which is that subjective risk assessment, understanding the risk in detail and doing a level of price assessment, price negotiation to get the best for your organisation, that’s the crux of the role. But actually what that means is the insights then start to be automated and really do part of that role of underwriter then become more, let’s say, a Data Analyst, Data Scientist type approach. So therefore, when you look in future careers, well, do organisations then start focusing less on, let’s say, the underwriting side as well? That’s always going to be required, but actually start supplementing it with other types of future skill sets as well. So I think that’s part of what I see, the future sort of challenge coming through the industry as well.
Neil Peters: Yeah, that was a great question that caught me outside and I was thinking, how does that work? But the answer is simplistically yes, because when you hire an experienced underwriter, they know how to underwrite. So that’s done. So what is the onboarding process? The onboarding process is your clunky system versus their clunky system that they dealt with somewhere else, your appetite, and then all of your admin that they have to go through around the edges. So if you minimise the admin and the system’s simpler and everything’s in the one place. Really, you’re just teaching them your appetite. Maybe there’s some introducing them to your protocol as well that might differ from where they were before. But actually, the onboarding process is really kind of trying to fit into norms that have developed and systems that have shot-off from something else. But the simpler an underwriter’s job is, as I said, if you can get it down to negotiation pricing and risk acceptance, if you like that kind of risk selection and you’ve got less admin and you’ve got simpler systems, then actually onboarding is going to become much easier. Yeah, and I actually genuinely think the retention, it will become much easier as well. The last thing you want is your underwriters sitting at traffic lights outside the office, dreading coming in because you’ve got antiquated systems. There’s 17 of them in one office. There’s a ton of admin to do. They want to come in and just be underwriters. So boiling it down to, I’m going to free you up from the admin burden and just let you get on with being an underwriter got to be a more attractive place to work. Right? And that environment is very competitive as well as the marketplace with brokers, the recruitment environment is very competitive as well. So, yes, undoubtedly technology can make it just a better job, a more enjoyable job, and an easier job to absorb. Yes.
Juan de Castro: Because what you’re saying is, one, less people will leave, the potential for them to stay will be higher if they’re working in a more modern environment. Second is if part of your boarding is understanding the appetite. Okay, so if you’ve got digital risk flows that are already prioritising and making those decisions for the underwriters, they will already have a recommendation. So they will ask, going back to your previous point, they will feel more comfortable making a decision, or is this in or out of appetite? And then the other point you made was about getting to know the brokers. Right. Again, if the system can hint them and say, well, focus on this one, because we’ve historically had a really high conversion rate with this broker, again, this is something we see across our client-base. The onboarding gets massively accelerated.
Okay, all this sounds fantastic, but, how each of you have started in this journey of deploying digital risk flows. And obviously with both of you, we’ve got a phased approach of deployment. So perhaps if both of you could give an overview of what a phase one looks like in this case with Cytora, deploying its risk flows and how do you see that journey to start making progress, I think that would.
Neil Peters: it’s been quite useful for us. I think it’s definitely quite easy to stage that approach. So we log an awful lot of inquiries. We receive thousands of inquiries every, well, probably month even and across multiple sites. And we try to centralise logging. We’ve tried different things to make that as efficient as possible. So partly it’s a volume challenge for us, but also it’s a speed challenge. So we want that done as quickly as possible and we want the submission in front of the underwriter so that we can start to engage with the broker positively or negatively. So that’s one chunk and that’s a big chunk of work, and that’s pretty easy to identify. Building naturally from that is that subject that we were talking about, which was kind of third party data sources that are coming in, other tools that the underwriters are using that is very time consuming. And also if you don’t do it, it can go horribly wrong. So it’s really vital as well. So I think the second part for us is that kind of deployment of, okay, so where’s the workflow come? Where does the data come from, how can we automate as much of that as possible right now? Are we in a position where we’d be able to totally rely on Cytora? Probably not, but it’s just a natural progression. We’ll work out how to bring data in from other areas and then linking it with their own data analytics as well. So that I think that could be slightly more complicated because they live in a complicated world of their own. So working with them, slotting in our appetite into that. So submission absorption, getting all the data that we need, and then dealing with our own kind of internal risk selection process in those three chunks, when we get to that, well, frankly, I think we’d be in pretty good space if we can do those three things over the next three years and then happy days, really.
Juan de Castro: That is definitely the roadmap. Submission setup, enrichment of data, triage. And one of the things we discussed also is about eventually straight through processing. Once you’ve done all of that, then can you identify the simpler submissions and start auto quoting some of them?
Neil Peters: Right, you’re actually right. That’s not necessarily my world. I sit in a world where we still kind of look through each inquiry and do it by hand. But certainly that you’ll get to a place where you can triage it. Right. So simple stuff goes automated, more complicated stuff, that’s another trick altogether, really. And again, great for utilising your resource to its best effect. That can become very powerful as well. Yes, you’re right.
Juan de Castro: Arvind, do you want to share with us also?
Arvind Drubhra: Yeah, from outside, obviously, we’re very early on in the launch with Cytora. So for us it’s around digitising the sort of submission intake, aggregating the external sources of data like sanction screening, credit checking, resolution to sort of external partner databases like Moody’s or Dun & Bradstreet and bringing that all together in along with other sort of relevant risk information as well. So whatever external repositories that we believe are relevant to that line of business. So, for example, property, it might be going on to Munich Re agreed to get flood zone data and liability side. It might be other liability relevant data sources, but pulling that all together, applying our actual prioritisation and appetite rules and then routing it back to the underwriter and actually having it integrated into our internal CRM system. So the data is all there, it’s being prioritised in our internal CRM, which the underwriting login to, it’s already there automated and with an initial view on what happens then. So that’s our sort of initial focus of activities getting into the risk flow integrated into our CRM. But actually, if you look beyond that, the next steps are actually you start looking further downstream in the value chain. So can we start pre populating out underwriting pricing tools, the data there on the submission, in most cases it’s about pulling that out and working towards that. So actually then you start removing a lot of the manual administration efforts that underwriters do say, and you’re presenting them with a package which they’re going then and using their expertise to actually perform sort of a proper assessment and review without having done all that manual workload, which we’ve now automated. So that’s our sort of focus. This year we’re piloting this in the UK. If we can prove the business value out of the UK, which I’m very confident that we will do, then we look at scaling this across other geographies, across HDI global. So that’s a sort of vision and roadmap, really.
Neil Peters: You mentioned a CRM actually triggered something as well. And that’s the other thing that this provides us with because you’ve got that data in absolutely the right format. The ones that get away, our conversion rate is not 100%, so the ones that snip out, we’ve got all that data for next year, so we can start to become much more proactive around targeting brokers and saying, look, we looked at this previously, we’d really like to look at it again. This is where we think we are. I think that proactivity, it’s really hard work. Pipelining has always been really hard work for sales forces and for development underwriters that brings in a whole new ballgame because it’s there on a plate. You’ve done the work, you’ve just kind of got to tweak it. And I think that makes a fundamental difference as well. So it kind of allows you to project forward a little bit as well and be much more proactive. A lot of what we do, and that’s a lot of what this conversation was, is how we react to what brokers send us. If we could get on the front foot, we’d probably give ourselves a little bit more time as well because everything feels rushed and just in time, the way the industry works now, if we could draw out the timeline slightly, that would help as well and it would help with the service provision too.
Arvind Drubhra: If you look beyond this, actually brokers are talking about automating a lot of their data intake processes as well. So rather than us getting sort of PDFs and PowerPoints of data on a submission, it’s actually receiving it electronically from broker systems. And therefore the play in future then is really right to Cytora start aggregating that from different brokers and pushing into an already existing pipe that has already been configured and built. So that’s a sort of future evolution of this whole risk flow process.
Juan de Castro: Almost regardless of the channel you receive the risks through, you still need to do all the activities you mentioned, right? You still need to augment it with third-party data. You need to make sure that you’re augmenting on the right risk and the right client. You need to be able to triage. You still want to make sure that the kind of underwriters are focusing on the most attractive risks first, right? So I think this has been a fantastic chat. I think it’s been quite clear that even though we started by saying, both of your businesses are very different, but I think then we converge. In the end, it’s all about creating underwriting capacity, making sure you’ve got accurate data in front of underwriters as soon as possible so that they can have a first contact with the broker quickly? And then how do you make sure that some of the more mundane kind of steps of underwriting are automated? So the rights were really properly focusing on risk selection underwriting and negotiating with the broker? Thank you, Arvind. Thank you, Neil. I think this has been a fantastic chat.
