The human side of AI-powered pricing 

Robin: Welcome everybody to this week’s InsTech podcast. I’m joined this week by Dani Katz, who’s the founder of Optalitix. Dani, 

Dani: welcome. Thank you. Good to be back. 

Robin: Yes, very much back. You were with us about a year ago. How’s business at Optalitix in the last 12 months? 

Dani: Business is doing very well, thank you. We continue to grow revenue and customer base, and we’re innovating more than we did at the start of the business, which is the thing that gets me most excited. We’re innovating more because our clients are asking us to do more things, and I think that, to me, is where things start really getting quite exciting. 

Robin: So let’s talk about those clients. What are they asking you to do? Where are you seeing most interest, traction, growth in your business? 

Dani: So I think when we last spoke, I told you about our entry into the reinsurance market, and we’ve launched that product last year, and that’s been very successful for us. [01:00] We now service five reinsurers globally, and we’ve significantly expanded our reinsurance functionality, and that seems to be what people are very interested in. Now, the thing about reinsurers is that they’re quite different from primary insurers in some fundamental ways. Firstly, their data requirements are so much more complex. They need to understand large treaty experience. They need to do simulations, multiple different pricing methods. They also suffer from a wide range of different models in lots of different locations that they use for their pricing, and they have underwriters that really wanna understand all the complex calculations and how they work for the price that they’re doing with their brokers. So we adapted our product to help them with some groundbreaking new technology being used for them, for their data requirements. And the good thing is that our Managing General Agent (MGA) and insurer clients are also growing, and they’re benefiting from these types of product improvements, which they get as almost automatic upgrades for their existing systems. 

Robin: There’s no such thing as a steady state these days, so you’re constantly developing that product. Where are you with product? [02:00] What have you been doing the last 12 months? 

Dani: So some of the key changes that we made recently are the introduction of Optalitix’s Data Lake. That completely removes any restrictions for our clients on data storage, the speed of analytics. It’s a completely new user interface developed on a really modern technology stack, which is fully API enabled, and other enhancements that we’ve made for our actuarial customers are around our Python hosting and analytics capability, so it’s all getting very exciting. We’ve also just developed an underwriting AI feature and it’s due to launch in H2. At the moment, we’re testing it out with our own systems and with a few clients and really starting to get very exciting for us. 

Robin: So you talk about AI enabling your product set. You had an existing product, it was selling successfully. What does AI enabling your product mean? What does it involve? What can you do now that you couldn’t do before? 

Dani: I guess the starting point for us on AI is to ask why we were implementing it. Everyone’s talking about AI, but [03:00] the real question is, what is it gonna give to our clients, and how is it adding value? And for us, it’s really the underwriters that have always been the key to success in any pricing transformation project. We conducted extensive interviews into pricing transformation over the last few months, and what’s fascinating is to see the difference in views between the stakeholders. Underwriters seem the least excited about the progress that they made, with only 34 percent seeing significant progress in their company’s pricing transformation efforts, which is pretty fundamental. So what we wanted to do is to build AI that underwriters wanted, that added value to the underwriters in terms of their ongoing quoting process, and really it’s about automating the manual part of their quote process. So for our first AI implementation, what we wanted to do was improve the underwriting experience and do it in the parts that we felt were the most important, which is really around how underwriters interface with the quote process. [04:00] We developed an AI feature that allows underwriters to talk to quotes in natural language to get their jobs done, so they can do everything that they’d always done, but do it using natural language in the AI agent of their choice. Again, this is due to launch in H2, and it’s exciting because it really removes the barrier between humans and the system. We’re making it much easier for the underwriters to do stuff on the system, to learn it, and improve the speed at which risk data can be analysed by a non-technical person. 

Robin: So that’s how underwriters will benefit, but your customer base is partly underwriters, partly the actuaries they work with. What have you done for actuaries and how will it change their lives?

Dani: So what we’ve given actuaries is the ability to analyse the data far more easily. So AI’s helping automate all the data access that they’ve always wanted. It means that they’re able to use Python if they want to interrogate the data in one way. They can use natural language via the AI to query it in other ways.[05:00] And I guess the important thing there is that they don’t necessarily want magic. They want simplicity, transparency and control. They want to know where their numbers are coming from, what assumptions are being used in their pricing and whether they can trust the output. So for us, the best AI for the actuaries is gonna be almost invisible. It’s gonna be giving them all the data and pricing information that they need and allowing them to do far more with it than they ever could before using both the tools of their choice, which may be Python, SQL or something else, or using natural language via AI. 

Robin: We did a survey earlier this year with actuaries and MGAs. So how did MGAs of various sizes use actuaries and some didn’t, and then what they used them for. One of the things that came back time and time again was that in the MGA world, actuaries were spending 50% of their time sorting the data out before they actually got round to doing their job. That’s presumably the bit you’re tackling, and as a result, they’re gonna end up [06:00] with lots more time. Is that less actuaries or are they gonna learn to do smart new things with their time? 

Dani: Good question. I would say for actuaries, there’s always more stuff to do. They need to learn more about the insurance market, learn more about the business that they’re operating in, learn really more about the human decision-making element that is involved in the pricing any type of risks. We know that actuaries are incredibly good at building models, analysing data and really understanding what the underlying risk is. But historically, they spent way too much time cleaning data, reconciling spreadsheets, checking versions, moving models from one environment to another. AI and these modern pricing platforms should reduce a lot of that manual work, and the question then becomes: What is the higher value role of the actuary? What can they do in this market that really lifts them up and even another level above where they are at the moment? And I think it is for them to get much closer to the commercial decisions that are being made. And actuaries should spend more time with underwriters, with portfolio managers and [07:00] execs, helping them understand what the models are saying and what decisions they should be following, and really understanding their feedback on the market and what it means for their pricing models. Because I guess my view is a pricing model is only valuable if it changes over time based on what the underlying human behaviour is. So I guess my advice to actuaries, become better translators of risk. Understand how underwriters think, how markets behave, how the various incentives work, and really how humans make decisions under pressure, under uncertainty, in changing markets, and that is where I believe actuaries can add enormous value. And in my opinion, we probably don’t have enough actuaries. What AI is gonna do is create more jobs, more responsibility for the actuaries that are involved already. 

Robin: I like it that for the first time in a long time, I feel like we’re talking about actuaries. It comes up in events in my world. There’s much more sense that the actuarial world is changing, that they’re embracing it, that they welcome that change. I quite like it. [08:00] 

Robin: Yeah, look, on that topic, you guys did a survey about what’s coming down the road from an underwriting and actuary point of view. What were the main findings of that? 

Dani: I guess our survey looked at pricing transformation in insurance and really trying to understand what is holding organisations back. Why are they not doing more on the pricing transformation side? And I guess one of the clearest findings was that insurers know pricing needs to change, with over 70 percent saying they had been modernised or in the process of transitioning to a modern pricing system. And that change has been driven by things like competition from portfolio performance, and now more recently from AI. [09:00] There’s this pressure to move pricing systems onto a modern framework. And I guess companies will always rely heavily on spreadsheets and manual processes, and that means that is a barrier to that digitisation. I guess the key thing that the survey also highlighted was the major difference between actuaries and underwriters. Actuaries tend to be much more open to new tools. They wanna use the latest coded tools like Python and R. They wanna use new types of Generalised Linear Model (GLM) models, and that is really because their work is already analytical and model-driven, and they’re completely comfortable with those tools. Underwriters, from what the survey was showing, are much more cautious. They need tools that are fast, transparent and practical because they’re really operating the operational flow of a business. They are busy during the 1/1 renewals. They’re gonna be busy for pretty much three months flat just working on quotes, and things need to be much faster. So for them, a new tool that isn’t obvious to them in how it works isn’t what they really want [10:00] The big message, I guess, that we learned from that survey is that pricing transformation is not just about getting better models. It’s also about the user adoption. It’s about the humans that are actually running the pricing underwriting process, understanding the tools that they want and need and making sure that they’re available to them. So whether it’s Python and AI for underwriters or for actuaries, or it’s spreadsheets that the underwriters really understand and can collaborate with the actuaries, it’s saying that the tool must think about the user, not just implementer because it’s the best tool for the job. The best tool is the one that the humans like and can use easily. 

Robin: Is that why underwriters remain firmly wedded to Excel, do you think, or spreadsheets? They’re used to it. It’s a theory that you’ve had before, and I can moan about it as much as I like, but I think you’re highlighting a difference between the evolutionary process that actuaries are prepared to go on and adopt and be open-minded [11:00] and underwriters being happy with what they’ve got, and that therefore the rest of us have to accept that and evolve around it. Is that the thesis? 

Dani: Yeah, I think it is. Our company from the start, the first project we worked on was converting Excel into a system, and we realised then how, A, how important Excel was for underwriters and for the whole pricing process, but B, also how complex it was to convert it. And once you convert it, providing an interface that people are comfortable with. Excel’s just an incredibly familiar, immediate and flexible tool that everyone understands. If an underwriter wants to change an assumption, test a scenario, want to understand a calculation, Excel just lets them do that quickly. It gives them a sense of transparency. You can see the numbers. You can look into the formula. You can see the logic. Everything is right there in front of them. It gives them some control. And in our survey, underwriters rated Excel above all the other pricing tools that they had in everything but one category. They were saying for in terms of flexibility, pricing, speed,[12:00] collaboration, all of these things, underwriters felt Excel was the best. The only place where they felt Excel was not at the same level as other coded tools is really about sustainability for the future, ’cause what they were thinking about is that Excel doesn’t yet connect all the systems that they want. But these days with the technology that’s developing and the products that are being offered, including from us, Excel can be embedded in modern systems in an easy way. And as we’re proving, even going as far as to link it up to AI, so you’re going from an Excel model into an AI-driven pricing process just in the space of a few minutes is quite extraordinary. 

Robin: So it sounds like companies like yours have got to reflect the fact that actuaries are going off on the journey, are going to have multiple tools for multiple jobs, and are fully embracing AI, whereas the underwriters know what they like, trust Excel. So suddenly you’ve got to work in a pricing environment involving underwriters and actuaries and fundamentally [13:00] different tech. Is that a complex? What does that mean for a company like yours? 

Dani: So our system allows the underwriters to work in Excel and to get all the information in Excel, as well as in normal forms, while the actuaries are able to use the tools that they’re most familiar with. Our platform takes all of those tools and treats them the same. It treats them as tools that can be hosted in the platform, allows for the complexity of those tools, allows for the data to move from one tool to another, and also ensures that the underwriter gets the tool that he or she wants while the actuary’s providing the more complex tools into the system, which gives them the pricing accuracy that they need. It’s complex, but it really works well when you see it in action. A lot of our clients are finding it’s very useful because it bridges the gap between what the actuaries want and what the underwriters want. 

Robin: Let’s change topic slightly. I’m keen to know where you stand. You’re a thinker in this space. There are a lot of people, and then [14:00] some of them outside the industry, some in, who take the view that we’re on the threshold of some incredible change in a world where you can now automate many of the jobs that were for trainees and involved a lot of manual intervention. Are we on the threshold of a job Armageddon and jobs are going to be lost all over the place, and there’s no question of any graduate ever being recruited again? Or do you see the world a bit differently? 

Dani: I think I see the world a lot differently. In my opinion, AI’s definitely gonna come and it’s gonna impact on the jobs that are available in the market, but I think there are gonna be more jobs rather than less. I think people are gonna be doing more things because what AI’s gonna do is it’s gonna take away some of the more manual, more less thinking jobs in the market. It’s gonna make those jobs a lot easier. But now we’ve got humans that are actually thinking about business, about what it means, how they can expand it. It’s gonna mean more products. One of the big things is that the world is completely under-insured. When a natural catastrophe happens, around 70% of those claims are not actually covered. [15:00] Now, the opportunity that comes in when you bring in AI is that all these people that are working in the industry can now start thinking, well, how do we expand the people that are insured? How do we get more people covered at more of the time? And that just means we need more people. Ultimately, insurance is a people game. Expand the insurance cover, and we will need more people in the market. I guess I don’t see a job Armageddon at all. I see a shift in skills and a focus on more interesting jobs. I heard an interesting talk from a CEO last week at Reinsurance Outlook. What they were talking about is the fact that AI is removing the drudgery of the standard insurance job, and what that then now means is we can bring young people into the insurance world, and they can learn stuff which doesn’t have that drudgery. That has been removed by AI. They can actually do interesting work, and we may find that suddenly insurance becomes exciting again [16:00] for young people. We need more of those people in our profession to grow it, to really expand on what we do in this industry ’cause we really add value to the world by covering these catastrophes. We need to cover more of it. 

Robin: People don’t have dinner parties anymore. But if they, that old dinner party thing of you and I turning up and they say, “What do you do?” “I work in insurance.” But you had the additional problem of being an actuary, and you could see people glaze over before you even start. The next generation might be lucky enough to be able to go to a dinner party, if they ever have them, because they’ve all got so many food preferences that you can’t possibly sit round a table and all eat the same thing these days. But now, that aside, yeah, no, wouldn’t it be nice to work in an interesting profession? Just a little bit too late for us. I like that answer, so I’m gonna test you again. What does all this mean for the future of the London Market, do you think? How do you see it evolving, and what role do you expect your business to be able to play in helping it evolve that way? 

Dani: I guess, I think already the London Market is becoming more digital, and I think that’ll just continue to happen. I think it’s gonna be more data-driven. I think we’re gonna see a [17:00] lot more technology being brought in. I think it’s gonna grow. I think it’s gonna be a lot bigger than it is at the moment. The network effect of being in London and having all these insurers within such a small distance of each other, you just cannot understate that. London is already the global hub of the insurance and reinsurance world. London is an exciting place to be already. I think what’s gonna happen is because you’ve got data now flowing through those systems, you’re gonna find suddenly things move faster. And when things move faster, you need people to spend less time on administration and more time on the risk selection, negotiation, portfolio strategy, growing your book. Those are exciting things. The role that people play in that market is just gonna expand. People are gonna have to lift themselves up. I guess where we see our role as Optalitix is we wanna help insurers make that transition practically. We want to sit in the middle of the problem everyone faces. We wanna be able to take their models and tools that everyone is using, especially Excel, and turn them [18:00] into cloud-based platforms that can use AI, use other systems, make them technologically advanced and able to do pricing on the world stage in the most advanced tools available. Key thing I wanna say is we’re not gonna force the market to abandon what works. We just wanna modernise the way people work so that they can actually adopt these new tools. We wanna make our tools things that people actually want to use and to get involved in. 

Robin: I do think that’s the key to where we sit now. It’s all very well being able to articulate the extraordinary things that AI is going to enable us to do. But if you don’t take the humans on the journey with you who have the wit, knowledge, experience and capabilities that AI will feed off, then I think what’s the point, frankly? Yeah, I always enjoy our chats. I hope that you’ll be back next year to give us your updated [19:00] thinking. We will certainly be testing you to see whether your theory about jobs and the London market is holding up. And thank you, too, for your support for everything we do. 

Dani: Thank you, Robin. It’s been an absolute pleasure always working with you. Every time we do these podcasts, it makes me look back on our business and see how we’ve changed, and I really appreciate you pushing me really to explain what we’re doing and how, and I guess also to put a human face to everything that we do. Thank you. 

Robin: No, it’s our pleasure. I never thought I would get so involved in the world of the actuaries, but thank you for forcing me to do that, too. Dani, thanks for joining me. 

Dani: Thank you.