Jason Chambers on the Impact of AI on the Insurance Industry

Mark: To discuss the rise of AI and its impact on the insurance market over the next five years or so, I'm joined here in the studio by Jason Chambers, Director of Innovation at Aviva Commercial. Well, Jason, if we start with today - how much is AI already being used?

Jason: It's used significantly in our underwriting and our claims processes. The focus that we're trying to deliver on is how we're supporting faster, more accurate decisions as it is in every other industry, really. It's enabling us to remove repetitive or simple tasks that our underwriters or claims were previously delivering and ultimately give our people the ability to make an informed choice to provide the high levels of personalisation and to recognise the customer far more than they have before. Really good examples of that would be how we're delivering summarisation content. So using large language models, we're able to summarise the previous conversations that we've had with our brokers, and in some instances customers, and very quickly then respond to the customer as if, in essence, they'd have just picked up the phone the day before. Or how we're adopting information very quickly to enable underwriters to access. So instead of looking up, say, 15 or 20,000 pages of underwriting content, they can simply just ask the technology, what is it they want to know and we bring back that information. These are all areas that we're looking to focus on in order to support our people, give better customer experience.

Mark: But from the data you've already amassed from that experience so far, are you able to give a sense of how much more effective or efficient it's made you as a business, or is it all anecdotal at this stage?

Jason: We measure, as you can imagine, a number of critical success factors. Response time is absolutely vital. We're seeing our customers and brokers demand far quicker responses from us, not just at new business but across a range of services. We measure speed of response and we're certainly seeing that reduce. As we're able to get to the issues in hand far quicker, as we're able to visualise the work far, far better and how we're able to give our underwriters the tools and assets that they need, as I say, in order to make faster decisions. But speed, visibility and how we execute our underwriting in a flawless way, it's now incredibly difficult to be an underwriter. In some ways it's easy because you have data and technology but the number of things we're asking our underwriters to remember, particularly as some of the changes in the agility that we see from customers. So the more that we can do to support our underwriters, see and understand the customers at risk and support them with more tailored solutions that those customers will recognise the better.

Mark: And how easy is it to predict where AI is going to go in the future? I mean, is there a point where you as Aviva have to make a big strategic decision on how you're going to use it or is this more of an iterative process?

Jason: There are occasions where a bigger and perhaps bolder decisions need to be made. At the moment the focus certainly is on iterations, understanding where and how the technology can support our business. I want to be clear though, the focus for us really and you'd expect most businesses to say this, is ensuring that we're solving real customers' problems. So the benefit and the business case are certainly ensuring that we can deliver a focus for our customers. I mean, more of them are now digital customers in one way or another, whether it be digital accountancies or how the customers range with their services. Their expectations are high, the feedback that we constantly receive is they're looking for seamless, integrated and consistent experiences. Technology can enable us to provide that and through the measurement of services and how quickly we get through the work and obviously the net promoter scores that we receive as well as the retention rates that we obtain - we're able to see the value of that technology and how it's powering our people to essentially be the best versions of themselves.

Mark: What would you say are the main potential dangers of AI?

Jason: Transparency and the lack of it and bias are probably the ones that immediately come to mind. The focus for us and I suspect many of our competitors in the market is ensuring that we place data and technology in the right hands and at the moment it's supporting our people, make faster, bigger, better decisions and see our customers far more. It isn't about allowing technology to start to think for itself at this point and we take bias and transparency absolutely seriously and have high levels of controls and governance to review the impact that it's having for our business and the impact that it's having for our customers. I expect probably to see more legislation come on to that in every other industry including financial services, but at the moment the focus is on augmented capability which is how we're putting data and technology in the hands of people rather than allowing data and technology to start and run and think for itself.

Mark: So how many people would you employ whose job is for example to look at bias? How do you measure something like that?

Jason: So we have teams of data engineers and data science as well as pricing and actuarially expertise. We review that data at appropriate intervals and in some cases that could well be weekly or less often. We look to understand and what we expect that data to deliver for us, and then obviously are actively interested in when that data deviates from the expectations that we have and of course we do sense and reality checks to ensure that we aren't discriminating against the cohort of customers that we would expect to insure. So we do have appropriate controls that will continually need to be reviewed and challenged to ensure it keeps up with the investment and the technologies that we are using.

Mark: I know it's very early days but has AI at any point thrown out something that was just completely wild that you all looked at and thought, my goodness, I'm glad we've got a manual override?

Jason: Not at the moment but I suspect the opportunity for us to see constant challenges in the technology. I think for me probably the big thing is just how fast the evolving landscape is in terms of its capability. One area that I see growing very rapidly is something called autonomous agents and what is the agentive landscape that we will be entering into. So AI was essentially invented for agents if you believe the hype, and these are autonomous co-workers that will start to work alongside our people increasingly undertaking simple tasks. The speed of change is rapid and we constantly need to perhaps recalibrate our expectations around what it does for our business to meet the expectations of our customers and our brokers who are constantly wanting more and more from us. So it's something that we take quite seriously for sure.

Mark: I've pushed a little bit on potential negatives. But what would you say are some of the most significant benefits that are coming out?

Jason: I think personalisation is absolutely massive. The stats from our own research highlight something like 76% of all customers expect service providers to recognise their individual wants and needs and anything that enables us to provide a more tailored solution for our customers is what we focus on. We are freeing up our underwriters' time to focus on customers that perhaps need more support. So recognising the individual nature of those customers and where they will need more advice and more recommendations. But I think the biggest thing for us is how we make our industry, in terms of the products and the service we provide, just a far more simpler environment. Complexity challenges our industry at every opportunity and I think data and technology has a massive opportunity to simplify the way that we work, how our customers engage us, the range of products that we provide, and enable them to recognise the services that we provide that really deliver for them every day.

Mark: Final question, Jason, what can you do to ensure that clients feel comfortable with the fact that Aviva is using AI and is probably going to be using a lot more in years to come?

Jason: Trust and transparency I think are absolutely critical, I mean from our own experience we recognise that when you are able to highlight how you are using the AI and you bring that to life to the customer in a way that enables them to resonate and understand. So it could be populating a form simply on the basis of what we know about the customer or pulling information that perhaps they struggle to very quickly access. So if we are able to highlight where we are obtaining the information from and how it supports our business and how it supports them execute what they want from us far quicker, then we see significant levels of adoption but I think it is about trust and transparency.

Mark: We have to leave it there, Jason Chambers, thank you.

Jason: Thank you.