How the Insurance Industry Is Leveraging AI - with Peter McMurtrie from West Monroe

Peter McMurtrie, a partner at West Monroe, stops by the show to discuss the evolution of AI in the insurance industry. Peter says the industry is currently focused on agentic AI and workflow orchestration, and he highlights key challenges such as data quality, governance and workforce readiness. Peter urges firms to avoid getting caught up in "AI Theater" by having a clear business case and ensuring the right use of AI tools. The conversation also touches on the potential for AI to improve decision velocity and quality in complex claims and underwriting.

More Insights from West Monroe

• Insurance Modernization Under Pressure
• Powering the Future of Insurance with People and AI

Transcript

(Note: This transcript was creating using AI. It has not been edited verbatim.)

Sean McMahon  00:06

Hello everyone, and welcome to the Modern Money SmartPod. I'm your host, Sean McMahon, and today we're going to be talking about AI in the insurance industry. Joining me for that conversation is Peter McMurtrie, a partner at West Monroe. Peter, how you doing today?

Peter McMurtrie  00:23

Hey, Sean, I'm doing well, great to see you.

Sean McMahon  00:26

Yeah, it's great to have you on. You know, you've been in this industry quite a long time, and so you got your finger on the pulse of what's going on. And obviously, among many hot topics, AI is one of them, you know, at least for the past couple of years, we could say. So let's start at the top. Let's look at things broadly, you know, what are some of the latest trends and how insurers are leveraging AI?

Peter McMurtrie  00:48

Yeah, it's, you know, Sean, it, you know, it's a great question. And you said, I've been in the industry for a long time. I somehow I went from being a young upstart in the industry to being the sort of the old veteran. But as it comes as relates to Ai, ai has been in the industry longer than I have. We've been using AI for almost 50 years in insurance, around, sort of modeling our pricing, helping work with claim flow. So really, when we think about AI, and what's really changing, I think the newness is really around agentic AI, and the power of that. And as I said, the insurance industry was an early adopter in AI, and it's continued to evolve throughout I think what's exciting is the trends are really what existed before continues to exist only faster, and then you're really seeing sort of new opportunities to help transform the industry. And so what I mean by that is like, where there's already automation or transactional capabilities in place, AI is really helping improve the efficiency and the effectiveness of the models and the capabilities that were there before, and then, where you really have a lot of power coming in is around workflow orchestration and sort of interconnecting, sort of the AI models and capabilities into the workflow of the underwriter or the producer. And it's really this co pilot, sort of solutioning that is really creating a lot of energy and excitement of, sort of the the art of the possible, of what's to come.

Sean McMahon  02:26

Yeah, I like that, the art of the possible. And, and you mentioned how it's actually in some way, shape or form. AI has been in the industry for quite a while. Obviously, recent years, it's really just taken off. And, you know, seems like endless possibilities. And you outline some of them right there. You know, with underwriters or producers. What are some of the hurdles right now amid all this buzz?

Peter McMurtrie  02:45

Yeah, and like all things, you know, everything is easy, and everything that is easy is difficult. And you know, it's that holds true with AI. And so the the capability of the models and the things that can be done is sort of the possibilities are infinite. But you're back to this idea of what needs to exist for the enablement of AI to exist within organizations. And I kind of think about this as these building blocks that have to be in place. Part of that is organizations have to have defined, sort of the guardrails around how they want to use AI, and I think this is like around sort of the, you know, what are the, you know, the ethical uses of it, how are they defining that? What's the governance around data security and things of that nature that they want to have, and they need to have those clear policies in place. They also have, need to have sort of a strong data environment, both structured and unstructured, because your AI is being trained off of off of your of your data, and sort of the the availability of it, the granularity of it, that all matters as well. Then you have to think about the environment you're deploying your AI into, and both in terms of the technology platform, and are you sitting on a legacy tech stack, or are you on a modern tech platform that can really leverage automation? And then it's also the standardization of of your workflows and your processes. You know it's always this, don't automate a bad process. Well, AI makes that even more challenging, in particular when you're leveraging AI to work alongside of your underwriters or your producers or whatnot. If they're working in different workflows and environments, you're having to create custom solutions to align up with how everybody's doing things. And the more variability you have, the more challenging that is. And then the last piece is the readiness for it and the and this is back to because AI is interacting more and more directly with your workforce, having you know your workforce ready for adoption, thinking about what does the future of work look like? What are those the skills and capabilities that your workforce needs to leverage and adopt AI effectively and that. Has to be an intentionality of how that's going to work. So again, a lot of those building blocks that have to be in place so that you can take advantage of, you know, sort of the exciting things that are available, but you got to do the work.

Sean McMahon  05:14

Yeah. So I love how you outlined all those things and but that last one workforce, I want to ask you a little bit about that, because, you know, I've recorded some other shows with some experts, some from other other areas, and there's a lot of talk about, you know, employees, like, How comfortable are they, you know, not just using what's kind of, what they're kind of quote instructed to use, but also, you know, toying around the capabilities and expanding the capabilities. So, so what are you hearing from that like, how are companies out there, a assessing their employees, you know, comfort level with AI, and also as a group, you know, if you've got half the people are comfortable, half not. How are they kind of leveling everybody up and and, you know, increasing the, you know, the IQ of the entire organization when it comes to AI,

Peter McMurtrie  06:03

Yeah, and I think, I mean, it's, there's really, like, two broad things there that need to be addressed. I think one is, you know, this idea of future ready workforce and exposing your workforce to the ability to interact with AI and understand how the different models can create value in the work they're doing sort of more broadly. And part of that is you want your workforce continuing to think about and imagine new ways to leverage the models and resources that are available now. To do that, it requires, again, that safe environment, and by me that is that structured environment where they can operate, say, with it, with an internal chat GPT solution, or something along those lines, that allows them to experiment, learn, grow, get comfortable with that. And that's that really just, again, creating more of an AI native organization. The other part of this, though, is in the work that they're doing, and under, you know, ensuring that they're using the tools and solutions in the way they're intended to be. And what I mean by that is there are some solutions that, you know, sort of being deployed where it really is intended to do full do the full job, in place of having a human sort of take care of that task, and in that instance, we want to make sure we're letting letting the the models and the tools work the way they're designed. There are other tools that are intended to be decision support. It's bringing relevant information forward to you. It's giving you context you know might be a client like this typically needs this type of coverage at these this limit range. That's not intended to be a prescription of write it this way, but it's intended to say, take that in consideration and then add your brain and your perspective to it to make those decisions. Same thing on the underwriting side or even on the claim handling side. And so it's important, again, from a workforce readiness that we're helping our teams understand the intended use of the tool and how they're supposed to interact to it. And that's really I think, then it's their leaders observing, monitoring and coaching to ensure that there is that that appropriate given, give and take. The risk is if I fully ignore the tool and just keep doing what I was always doing, that's not a good outcome. If I just follow what the tool says and don't apply my own knowledge and experience, that's not a good outcome. And it is that that dynamic of leveraging the resources helping an Augment improve the decision quality that I make, and the if you're doing it in that way, the models will evolve and get smarter as well along the way. 

Sean McMahon  08:46

Yeah, and I think everyone still likes to hear about the human still being involved in some in some way, shape or form, and kind of guiding it. So, I mean, I know you talk to a lot of folks in the industry, and you know, want to kind of speak a little bit about who's kind of kicking the tires on new capabilities for AI? Have you heard of any unique or something that might have even surprised you, that is a way that an organization is using AI that is either flying under the radar now, or when you look at it, you're like, Oh, wow. In a year, 18 months from now, everyone's going to be doing it that way.

Peter McMurtrie  09:22

You know, I wouldn't, you know, in this stage. I mean, there's not much flying under the radar. I think, if anything, you've got, there's a fair amount of organizations that are doing a lot of AI theater. And what I mean by that is, they're, they're, they're talking about what they're doing, but it may be in a, you know, sort of very controlled, isolated environment. There are really not a lot of players in the insurance space today that are truly fully leveraged, leveraging AI, where I think about the opportunities and so not as much. Under the radar, but just under realized yet, is in the complex claim handling, the complex underwriting, complex selling from a producer perspective, and it is that, as I said, that AI copilot of partnering and leveraging large sets of unstructured data and patterns around what we've done with these types of accounts or these types of claims in the past, and bringing that forward in a just in time knowledge approach to allow that decision maker to have the right context in the moment, not having to go search for it, swivel chair, if you will, into other areas to find it, but Have it served up in a way that provides the broader context as well, that allows them to make better quality decisions faster than they ever have before. And I see that as sort of like under the radar, in the sense that that's like the highest value use cases, if you can improve the decision velocity and decision quality of your highest paid and your most impactful associates.

Sean McMahon  11:07

Okay, now I want to just shift gears a little bit and talk about governance, right? So, you know, obviously, in the last few years, there's been all this buzz about AI and I, actually, I like your phrase from a few moments ago, AI theater. Yeah, that's a good one. But how has governments, excuse me, how has governance evolved to keep up with the rapid pace of adoption?

Peter McMurtrie  11:28

Well, it's a nice Freudian slip, because you said government, I think, and in all seriousness, there is from a regulatory perspective, that's where it hasn't kept up. But I fully expect that we are going to see an increase in regulatory guidance and control around how carriers are are using AI. That hasn't come yet, but I absolutely guarantee we will see more of that emerging in the in the coming, you know, certainly in the coming months and by year end, I would expect to see NAIC and, you know, and others weighing in. In that perspective, I think internally for organizations, that is where they it's it's been a guardrail and and, and a little bit of an impediment, or a governor, depending on, you know, the how optimistic or pessimistic you want to be around, keeping organizations from moving too fast, because there is the concern around, you know, disparate treatment, how we're using the how we're using AI solutions, and the types of decisioning that they may that may be leading to, I think there's also the data security concerns related to sort of exposing data to the to the models, or the potential for a model unintentionally exposing data into the environment. And, you know, with the with, you know, through the rash of cyber attacks in the insurance industry that we had last year with a lot of the data exfiltration that created caution as well. So I'd say right now, it's a bit of organizations are realizing, you know, their their governance controls, are not able to move as fast as, sort of what you know, as fast as the what the models could potentially enable right now. So it's creating a little bit of that, that governor or slow down, but they're certainly going to need to speed up. But the regulators are going to bring more sort of controls and perspective that carriers and others are going to have to have to respond to. So it's going to be a challenge state here for the next, you know, certainly 18 months, that's going to create a little bit of caution in terms of what people decide to do.

Sean McMahon  13:56

So in terms of that caution, how does that impact, you know, not just the governance piece, but like, overall, the value of AI, how is, how are people feeling about the ROI right now?

Peter McMurtrie  14:07

Yeah, I'd say, I mean, most, you'll, you'll hear a lot of the ROI is low right now. However, the ROI is low where there's not a clear business case that's driving the decision of how they're going to the organization is going to use AI. So it's not a clear business problem that led to the adoption. It's low where there's not good underlying data quality. It's low where there's not well defined, consistent workflows that the solutions are being deployed into. It's low where organizations haven't been intentional around the change management and the people readiness, components of it, when those things have been addressed, the return is very, very high. It's like 13 multiples of what it is on average. And I think that's the distinction of back to if you're building blocks are in place. And. Been intentional, and you're solving the appropriate problems for your business. It's not about running out and say, I've done something with AI or I have an AI strategy. It's about I've got a business strategy. I'm solving tangible business problems, and AI is a tool that's fitting into the solution when done right, high ROI, when not done right. That's what you're seeing. Is the the lack of real benefit realization. Yeah,

Sean McMahon  15:29

it sounds like, you know, my next question I had ready for you was about, you know, what differentiates firms that are that are really at the forefront of AI and some of the laggards? And sounds like you kind of partially already answered that question in terms of…

Peter McMurtrie  15:42

Yeah. I mean, our industry is a I mean, we're made up of the haves and have nots. In particular, from a carrier perspective, when you look at the top quartile performing PNC carriers, the gap from top quartile to bottom quartile is the broadest of arguably any industry out there, and it's and those top quartile carriers that made sort of disciplined investments in their technology, disciplined investments in their data over time, they have structure and rigor in terms of their operations. They know what problems they're trying to solve, and they're consistent about their approach. Those are the organizations that you're seeing making meaningful progress with the adoption of AI, because they're very targeted with the problems they're solving and how they're using it, and that discipline is rewarding them. And then you have a whole bunch of folks looking for a moonshot or a CEO that wants to be able to tell his board how they're using AI, or be able to talk about it at a cocktail party and and those are typically in your your laggard group,

Sean McMahon  16:48

Alrighty, well, you've, you've kind of already walked us through a lot of the the key components of a successful AI strategy. But, you know, I'm going to ask you now to just sum it up in kind of one or two points. You know, if you're giving advice to firms that are, let's just say, at the midpoint of the adoption process, right there. They've got some good ideas. Haven't, you know, gone all in yet, but they're not completely dragging behind. You know, what are one or two pieces of advice you'd give them?

Peter McMurtrie  17:12

Solve meaningful problems. That's number one, find the right use cases, because that's going to the value is going to be there if you're solving the right problems, and the second is all around your people. And if you've got the right problems, and you've got your your people aligned with how you want to solve it, the rest is the rest is going to happen. The the models, the solutions are phenomenal, but it is the environment that you're operating in, the problems you're solving, that's where the breakdown occurs. And so focus on getting those two things internally right. And then you know, whether it's the the leading model or one from a year ago, it's gonna it's gonna work wonders for your organization. But it is about getting those two things right first.

Sean McMahon  18:02

Well, that sounds like some pretty clear and concise advice. Yeah. So Peter, listen, I really appreciate you sharing all your insights with us today. So thank you very much for your

Peter McMurtrie  18:11

time. Yeah, Sean, I appreciate it, and this was great. Thanks a lot.

Sean McMahon  18:14

Well, that wraps up our show for today. If you enjoyed this podcast, please share it with your friends and colleagues. Be sure to follow us on Apple, Spotify, YouTube, or wherever you get your podcasts. The modern money. Smart pod is a production of SmartBrief, a future company.