Transcription:
Patti Harman (00:06):
Welcome to this edition of the Dig In Podcast. I'm Patti Harman, editor in chief of Digital Insurance, and today we're discussing how the insurance industry is using AI as part of the claims process. So probably no technology has seen as much hype and adoption as AI, which has been around for decades, but became more mainstream with the launch of chat GPT in 2022. As the insurance industry integrates this technology, it's creating a variety of opportunities, changing interactions with customers, and changing many different processes. One of the areas where AI, Generative AI, and even Agentic AI are having an impact on the claim space. And here to chat with me today is Lee Elliston, chief Operations Officer of claims and UK operations at Aspen Insurance. Thank you so much for joining us today, Lee.
Lee Elliston (01:05):
Thanks for having me, Patti. I believe it is the second time, so I feel like an insurance celebrity that's making their second appearance on Saturday Night Live.
Patti Harman (01:15):
You should definitely feel that way. So I don't think you can look at any aspect of the insurance process and not see some sort of impact from technology, but AI seems to be dominating much of what we're covering these days. So could we start with a look at how insurers are incorporating AI into the claims process? And maybe you could give us a high level perspective on the different ways companies are using AI in claims at this point.
Lee Elliston (01:49):
Yeah, sure, Patti. I think it does relate to different insurance companies in different ways, depending on whether it's a retail business or specialty business that you are supporting within your underwriting portfolio. But specifically the first notification of loss process, the ingestion of data and documents as part of that process, how we triage claims within that process once we've received them, and then how we typically screen documents historically in quite a manual way that can now be digitalized a lot more, but also how we can use it to screen and provide risk alerts around fraud and sanctions. So there's lots of different parts of the claim life cycle that I think people are kind of utilizing AI and parts of AI technology within.
Patti Harman (02:50):
Yes, and that's a great segue to my next question because I was going to ask you, how is the adoption of AI changing the claims process? Because I feel like when we're covering it, we've been doing a major series on this topic on digital insurance, and it is just fascinating to me to see how it permeates every aspect of claims. So how is it changing it from your perspective?
Lee Elliston (03:18):
I think this question's really interesting around the change that creates. I feel like the different elements of AI tooling and technology have quite a powerful force, and I relate it to the powerful force that smartphones have that enable us to do some really smart things in such a small device, but have the power of us doing things in a very different way to the way that we did them before. I feel like in the insurance lifecycle we have moved from analog to digital, but a lot of that move is just digitized. What we did before not fundamentally changed the way we did it or how we did it. I think AI enables us to digitalize our processes and our practices, and I think that's why AI has got such a support around it from different people, technologists, business people, operations people. Everyone seems to have quite hype around ai, and I think it's because they've seen the power in the tool that can re-enable the way they work and it can make it smarter for them.
Patti Harman (04:39):
It really does, and it's been interesting to me. I feel like I have a very different perspective. I get to see the insurance industry from that 10,000 foot view, and I've really been fascinated to see how companies are using it in different ways. And we talk about AI as the major catchall for technology, but we're seeing even AI is evolving and new versions seem to emerge with different focuses. We had AI and generative AI has the ability to capture and verify images. Are you seeing carriers use this type of AI as part of the claims process as well?
Lee Elliston (05:21):
Yeah, we're definitely seeing it be used around the more interactive elements of AI that you may have used chatbots and others for in the very early days. We're seeing it replace the automation and some of the robotics that people used in the early days as well. I think the imagery and the aspects that AI can be utilized for that, especially generative AI is interesting because it requires some credibility around the actual data, then the documents that it's referring to and the images it's referring to, and it needs to capture the source information and the audit trail and the data around any imagery as well. And that's quite critical to the use of generative AI based both on imagery and unstructured data in documents to give people the confidence of what the source of the information is, what the audit trail of that is, and how trusted it can be to be a real powerful tool for people to use in their business process.
Patti Harman (06:37):
Yes, we're trying to figure that out from a journalistic perspective in terms of how can we utilize AI as part of our processes without it creating any heartburn for us, so to speak. So you talked about using Generative AI, will it be able to help insurance carriers identify fraud claims or at least maybe fly claims that might require a little bit more investigation then?
Lee Elliston (07:05):
I think that's a really strong user case, which I think people have started to develop and use already. A number of carriers, including ourselves, have risk factors around fraud. And if you've got risk factors and rules that you can plug into the AI technology, then it can really, really provide you a powerful insight into a scoring or a rating mechanism around the fraud or not. And my big thing around the AI tooling is that it just provides intelligent assistance, each part of the process and the cycle. And that intelligent assistance is really powerful when you are looking at a quite volatile environment where claims volume is going up and the complexity of claim in the quantum of claims is increasing. So when you've got intelligent assistance that can cut through that and support your decision-making, then it's really helpful.
Patti Harman (08:10):
I love how you call that intelligent assistance because it's such a great way to describe how companies can be using AI. We're going to take a short break now and we'll be back in just a few minutes.
Welcome back to the Dig In podcast for chatting with Lee Ellison of Aspen Insurance about how carriers are implementing AI and other technologies into their various claims processes.
So we've talked about AI and Generative AI and the latest version of AI to emerge is Agentic AI, which has the ability to help adjusters as part of the decisioning process, such as helping to confirm coverage for losses and certainly other uses too. But are you seeing companies integrating this type of AI and how could it change the claims process going forward?
Lee Elliston (09:08):
So I've not seen anyone adopt this element of technology and use it in this way as yet, or that's not to say they haven't. I don't believe that too many people would necessarily be vocal around adopting the technology in this way. I think there's some fear factors that come with it. I think that you've got some policyholder and customer fear factors and risk factors that need to be considered, but you've also got the fear factor of people feeling like the AI is doing their job at making a decision on a claim and an assessment on a claim. But the use case for it for me is that intelligent assistant. So how can it provide intelligence from unstructured data and structured data harmonizing those data points to give you some recommendations or insights and those recommendations or insights move the intelligent assistance and information into the hands of a technical specialist.
(10:07):
And the technical specialist then has the power to make the decisions and that they need to and do today, but they're just doing it in a more effective and efficient way with the tooling that's provided. I think this is a really powerful way of change and practices and processes. I think too many, too many claims departments have turned into data entry teams historically and moved away from being technical specialists. And I think this moves the pendulum back and swings it back to being technical specialists that are closer to the product, closer to their policy holders and add real value in the cycle of the product and the claim.
Patti Harman (10:48):
Yes, I would agree. I've interviewed a number of executives and one of the things that they always stress is the value of still having a human in the loop that you can't just assume that AI will be able to do anything. And I know as we're working trying to figure out how to use it here at Digital Insurance, that is one of our key tenants is that no matter what AI does, what it summarizes or what it thinks it knows, we still have to have a human review whatever's going on. What are some of the challenges for carriers that are integrating AI into their various claims processes then?
Lee Elliston (11:30):
I think the challenges are around the risks around it. So the perceived risks around customer or policy holder harm or litigation. I think they are risks that can be mitigated and managed with regarding the tool and how it's utilized. Back your point, Patti, around human in the loop, human in the loop is a risk mitigation to certain decisions that may be made and how it may become discoverable in a litigation process. So there are a couple of the risks. I think the capability point is a risk as well. I think we haven't necessarily supported the capability of claims teams historically, and I think that's because we've wanted them to be technical specialists, but now we are in a world where we want them to be technical specialists with the ability to read and write rules that come out of the technology, be able to use the technology and tooling that the AI is operating within really seamlessly for the benefit of themselves and their customers. And I think we need to help support and upskill the adoption of AI through our people's skills and the development that they need around that.
Patti Harman (12:59):
I think the upskilling piece is particularly important because it removes some of the concern or fear that people might have about using it, and it really helps them to see that, hey, this can help enhance what I'm doing. This can save me time or it can flag things that I need to pay a little bit more attention to. And I think that's a really important aspect of the adoption process. There are projections that close to 400,000 insurance professionals will be retiring in the next few years and claims is expected to be one of the more impacted areas. Do you see AI and maybe some of these other technologies helping to kind deflect some of this loss and automate more of the processes? And if so, where do you think it could have the greatest impact?
Lee Elliston (13:55):
I think the upfront processes around ingesting a claim, assessing a claim, and processing the data, whether that's structured or unstructured, I think it can really help within, I think it will allow the apprentices and the graduates that we've got. We've got programs within Aspen for apprentices, graduates school leave, and it allows those people to start really focusing on becoming technical specialists for the future and not doing the learning on the ground kind of processing activity that they may have done historically. So I think that that helps with the colored talent gap that we perceive that we've got from the technical specialists leaving the industry in the world of claims. I think the more that we look to innovate around the use of technology and the use of ai, the more people we attract to a role, insurance sometimes gets a bad reputation, is not necessarily looked at as a career of choice for generation alpha, but I think the use of technology in insurance can make it more sexy for that generation. And I do think that they will be able to see roles that they can relate to within insurance because of that. And I think that's really important for us to sell about how we using AI and how people can develop in the world of insurance as part of using their data skills and their technology skills.
Patti Harman (15:36):
I totally agree with that assessment. I think people underestimate, well, I know that they underestimate how incredible this industry is, but I think the technology aspect is definitely helping to attract another generation of insurance professionals and they're already used to going to their phones to get information. It's like when I was raising my kids, I had to check books on parenting to learn things. Now you just type in a question and the answer pops up and you have an answer in 10 seconds. And I think it's the same thing in the insurance space, that there's a lot that you can learn online and from Google and from AI even. So how are policyholders adjusting to the digitization of insurance? Are they finding that adoption is easier because they basically use it in so many other areas of their lives at this point?
Lee Elliston (16:32):
I think it's more relatable. I think that the point you just made around we've got this, a lot of us have these powerful smartphones and iPads that are available to us and the things that we can do on them all make our lives easier. And I think the journey of how we can explain how technology is used to benefit the customer and the policy holder and how relatable that is to a chat GPT tool or other tools that are on their phones day in, day out, I think is really helpful. It also provides some transparency as well when we're talking to policy holders around the differentiation that we can offer in the Aspen service from our technical specialists, and that's what we pride ourselves on. We can be more transparent around the process and what we are doing in it and how committed we can be around what kind of response or information or data you'll get at different points because we know that we've got trust in how the trust and confidence in how the technology's been deployed.
Patti Harman (17:42):
What excites you the most at this point about the adoption of technology in the claims space?
Lee Elliston (17:49):
I think what excites me the most is that even though it's technology, people have stopped talking about how a system could help them. So I think that point around digitizing before is all around, well, I've seen a system or a platform that can do this for me. And actually it has very marginal impact on what a platform or a system can do. I think this technology is more game changing because it impacts behavior, process, practice, the tooling that we use. I just think it's way more powerful than what we've seen before. And we are building change management streams of activity Aspen around our projects because it is more than just deploying technology. It's a real game changer and impacts so many elements of what people do and how they do it.
Patti Harman (18:43):
And that kind leads into my next question, which is how is Aspen incorporating AI and other technologies into your claim strategy at this point?
Lee Elliston (18:54):
So we've run a couple of proof of concepts, which I think we referred to when I was on with you the last time. One of those proof of concepts has turned into a formal project where we are looking to deploy generative AI within our reinsurance claims process. This is looking to take our data processing away. It's looking to provide insights during the notification of loss and subsequent claim update process. And it's looking to do that, looking at historical data and existing data that comes in within a human in the loop process. So it's a powerful platform that sits on top of our legacy technology. So we're now getting our legacy technology to not work as hard, but still be an underlying platform that benefits how we operate within Aspen with the technology up here that enables our people and the interaction our policy holders have to get more benefits in what we do in the way we do it. So it's got real potential to show our first real user case across the claim life cycle for reinsurance claims. It's a test bed for what we could possibly use it for within insurance claims, and it's got everybody really excited about having a powerful tool at work as well as in their hands when they get home and they want to use chat GPT.
Patti Harman (20:34):
Wow. It's exciting to see how everything has kind of evolved for you. Do you find that there's more collaboration between the different departments as Aspen integrates AI across the company? And is this making adoption easier for your employees too?
Lee Elliston (20:54):
I definitely think it drives a lot of collaboration from a operational perspective and an IT and data perspective. I think we've got really strong principles and processes around a joined up data strategy and a joined up architectural strategy and a joined up operational strategy. And we're starting to see how we can have cross-functional projects and opportunities from using AI in the same way or using the same platform or tooling in the same way. So it's definitely bringing us closer together from an operational perspective, I think it definitely makes adoption far easier because it will allow the value chain at Aspen within somebody that sits within claims or underwriting or finance or operations to all talk the same language, all understand how the same tool operates, and there'd be kind less friction in that process of collaboration between the different functions. So I think it has a lot of business benefit, operational benefit and strategic benefit.
Patti Harman (22:13):
That's what I've seen as I've covered the adoption in other companies. It tends to bring down those silos because no one team or no one department can implement AI because it's going to affect everybody else that has anything to do along that value chain. So it's been exciting for me as somebody who's covered the industry to see what an impact it's having all across the board for everybody. So we have covered a lot over the last few minutes. Is there anything I haven't asked you that you think is important for our listeners to know about claims and technology adoption?
Lee Elliston (22:51):
I think the one thing that I would say is that claims has probably been an underinvested part of the insurance value chain for some time. I think platforms have come and gone and they've never had the business benefit that was socialized at the front end. I think that the rate of investment and adoption of AI now within the claims function is really positive, both for the employees within claims teams, but also the future talent that could be in claims teams. And I think that people should really look to regenerate their ideas around innovation. The business cases that they had around it to generate a solution and a platform or a workflow would support their claims process and dust them down and think again and think again with a new technology and a new mindset around how you might use chat GPT at the weekend because it can take you somewhere. And there's a lot of simple solutions to quite complex answers and processes.
Patti Harman (24:08):
Yes. I think there are a lot of possibilities still ahead for this industry as it figures out how to utilize AI and other technologies across their different platforms. Thank you so much, Lee, for sharing your insights with our audience. Thank you to our audience for listening to the DigIn podcast. I produced this episode with audio production by Anna Mints. Special thanks this week to Lee Elliston of Aspen Insurance for joining us. Please rate us, review us, and subscribe to our content at www.dig-in.com/subscribe. From Digital Insurance, I'm Patti Harman and thank you for listening.