CIO PODCAST: THE AI ADVANTAGE | EPISODE 2

Barbara Call 00:06

The AI era has arrived, but there's no guarantee of success. Industry estimates say anywhere from 70 to 95% of AI pilots fail to get off the ground. The question is, how can you move beyond POC prototypes and feasibility studies through real ROI and measurable business outcomes. What does successful AI adoption look like, and what are the steps to move from potential to pay back? Welcome everyone. I'm Barbara Call, Global Director of Content Strategy at cio.com, and this is "The AI Advantage: Navigating Risk, Reward and Real World Deployment", created in collaboration with cio.com and Vertesia.

Barbara Call 00:49

In this episode, we'll be talking about how AI adoption often stalls when the C-suite and the board aren't reading from the same playbook as everyone else. It's becoming clear that IT and business leaders need strategic consensus, governance and enterprise value around AI to find success. Today, we'll be exploring how to successfully bridge the communication and priority gaps between it the executive team and the board, and what is the shared language needed to discuss AI risk, define clear ROI and establish a scalable operating model that moves beyond siloed projects. But first, let's introduce our three speakers. First up is Andrew Robinson, CEO at skyward specialty insurance. Welcome Andrew. Tell us a little bit about your role and responsibilities at Skyward.

Andrew Robinson 01:43

I'm the CEO of Skyward. I've been with the company for a little over five and a half years, and been the CEO and the chairman of the company as well. But my role, as you would expect as a CEO of a company, is really very much about driving the sort of a strategic direction of the company in all regards. And I think notably for us, at the beginning of this year, we just completed a transaction which involved the acquisition of a quite sizable Lloyds of London business called Apollo. So today, my mandate, obviously, as of sort of a broader view with not only our US specialty business, Skyward Specialty, but also with our Lloyds of London Business, Apollo.

Barbara Call 02:38

Excellent. Nice to have you. And our second speaker is Casey Kempton, Nationwide, President, Personal Lines. Nice to have you with us today, Casey. Tell us a little bit about yourself and your role at Nationwide.

Casey Kempton 02:51

Absolutely, very happy to be here. I lead Nationwide's Personal Lines business. I took on this role almost two years ago. Next month will be two years. I have responsibility for all of our underwriting product, claims, operations, service, etc, for our personalized business, which is predominantly homeowners and auto insurance, in addition to umbrella and some additional products. We have made an investment across Nationwide in technology over the last five years, but going forward, a pretty sizable investment, and a good portion of that relates to AI and generative AI, and it's becoming a really critical part of how we are running our business, managing our business, and where we see future opportunity, particularly in Personal Lines.

Barbara Call 03:41

All right, great. Thanks for joining us. And our third speaker is Keith Schlosser, multi time CIO in the insurance industry. Welcome Keith. Tell us a little bit about yourself.

Keith Schlosser 03:52

Thanks for having me on the podcast with Casey and Andrew. I really appreciate it, and I think this is an important conversation. I've been in the insurance industry for 35 years. I was a broker and an agent. I was a field underwriter. I spent the last 25 years as a CIO. Currently, I advise AI and cyber security firms, PE firms and VC organizations.

Barbara Call 04:14

Excellent. Nice to have you, Keith, and so let's kick things off, Andrew, my first question is, for you. As CEO, you're the ultimate decision maker regarding where company capital and focus go when you decided to move beyond pilots and commit to an enterprise AI blueprint at Skyward, what was the single most compelling business case? And secondly, was it purely cost savings, speed to market, or something more fundamental that drove those decisions?

Andrew Robinson 04:45

Okay, well, it's a big topic, so I'll try to try to simplify, maybe at least the experience at Skyward. So I may be rather unique in that even though I've spent a quite lengthy part of my career as an insurance executive. The four years prior to joining skyward, I entered the VC world as an operating partner, and then ultimately, led an InsurTech company through a series C financing round before, before leaving, and the company was an AI company. So I would describe my experiences being on the front line. And so when I got to Skyward, I came in with a very robust view about the importance of technology and how its going to play in combination with great talent for us to win where we're going to compete.

Andrew Robinson 05:41

And I would describe our approach as we really tried to open up the funnels and let the businesses lean hard into the use of technology, advanced analytics, predictive analytics, and prior generations of AI particularly, around machine learning and so forth. And we were very much about supporting them in a way that they could fail but fail fast. We would celebrate success, which would create a sort of innate competition across our organization to try to leverage what others were finding as being sources of success. Now we're way down the road in certain places, like, for example, in our A and H business, we're five years into deeply using AI in our underwriting processes and that certainly is an example that provided a great reference for other business leaders. But many other businesses took longer for them to develop their view, become comfortable with it, and I would say only really in the last two years has it sort of shaped into a clear view of what we're trying to do, both on the underwriting side and the claim side and that's a substantive conversation, but I would say that there are some things that we were very lucky about, which is, we decided to control the desktop ourselves, for the underwriters and the claims professional. So our ability to deliver new capabilities, let's say a new risk signal, that may be an AI driven kind of risk signal to our underwriters, is something that we could deploy really fast and so in that, the speed in which we built capabilities today, because we're controlling that desktop ourselves, what we call Sky View, visual underwriting experience, is really central to the speed in which we're working. And I wouldn't say that there's this macro blueprint, per se, as much as a framework for which we can deploy new capabilities really fast and this is an AI arms race that I believe we're way ahead of, but the moment we stop, clearly others are going to catch up, and I'm sure that three months from now, there will be a whole bunch of things that we will be doing that aren't even envisioned today, largely because we've created the environment and the framework to do it, as opposed to, like, hey, there's some brand strategy that is very clear on everything that we want to deliver to our underwriters.

Barbara Call 08:34

Oh, great. Thank you, Casey. I want to ask you, sort of the same thing. Have you seen the same kind of thing or something different. Can you tell us a little bit?

Casey Kempton 08:43

I think as we we look out into the world and we track the trends and we see where technology and data are evolving, we've been on this journey in insurance through advanced modeling, analytics, machine learning, that I think it naturally aligns for our industry to leverage as much intelligence as we can. I think, at Nationwide, for us, moving from pilots really helped us focus on three key priorities about our associates, our customers and in general, our operations. So we think about human machine collaboration. We're a very people connected and people oriented company, and we think of ourselves as being machine enabled. A lot of our pilots focused on how we could become more efficient, but we saw a lot of innovation opportunity and how at the desk level of all of our associates, we could enable more insight as an example, and faster problem solving and decision making. We also apply AI tools to solving real customer problems, whether the customer to us is both our distribution partner and our member, our policyholders, and giving access to tools that enable us to better serve our customers.

Casey Kempton 09:58

And then finally, we've been of all our operating model for years and AI allows us to really integrate agility and velocity into everything that we do, and moving to a global 24 by seven operating model to drive results. So when we anchor ourselves on those three priorities, the innovation that we've been able to do in this space of human machine collaboration has involved things, initiatives that we call everyday AI, where we've got the tools in place that we can get onto the desktop and into the workflow of every one of our associates. And we're seeing great enthusiasm for those kinds of tools. I think the biggest opportunity that we saw is really moving to solving customer problems or developing customer opportunities. And we developed 18 flagship use cases across our businesses and shared functions. These are all senior decision maker led, and they're driving very specific outcomes, whether that's in the sphere of underwriting, whether that relates to how we analyze claims and adjust claims, where we additionally can bring those forward to solve operational opportunities for us. So for us, it has been, experiment and then very quickly we came around. We've got to keep human in the loop in all of the AI initiatives, but taking them from pilot to executive sponsored flagship opportunities has been a pretty material shift for us in the last year.

Barbara Call 11:39

Wow, that's great. Very exciting. Both of you. Let's talk a little bit about challenges. So Andrew, let's start with you. Did you run into any challenges when trying to execute on your vision?

Andrew Robinson 11:51

I don't mean to stand us out as unique but I do think that we're rather unique, and that when I arrived, I challenged our CIO/CTO to not think about any constraints in terms of what what we're willing to invest in. But the ideas have to be powerful. They have to drive competitive advantage. We have this whole notion where we're made up of a lot of very distinct businesses that are very niche focused, and with this notion of rule your niche, which is about building like legitimately defensible positions and technology and talent together, are just as Casey had said, are really sort of that, that sort of that source of how we think about these defensive positions that we're creating.

Andrew Robinson 12:48

A big part of this though, is you have to get your business leaders, to really start to understand and envision what's possible. And that's something that's a really hard thing to create, but once you do, it's a snowballing effect. So I'll tell you a few things that we did to really affect that. I mentioned that very early on, we just started to introduce things our in age. Business was a great example of where we had early success. Twice a year we would get all of our employees together for what we call the Tech Activation Workshop, two hours each day over two days, and the business leaders, in combination with their technology partners, would present things that we've deployed that were really about advanced technology, that have been operationalized and effectively it was a stimulus to raise the awareness business led. It created a lot of competition, because people would say, oh, person X is doing that. I'd like to replicate that.

Andrew Robinson 13:57

And that started to snowball, and really became very much part of our organizational culture. We don't believe in a lot of things that other companies do around having separate organizations that deal with innovation. Organizations have been known to set up venture units and all that. We're like, "No, our business leaders need to own that." And that became a core part of our fabric. And I'll just give you a sense as to how far along we came. In 2023, we brought an early version of OpenAI over our firewall inside our organization so that we could experiment with it. And we use OKRs, objectives and key results, as the sort of goal setting mechanism. Anybody who's read John Doerr, or familiar with tech companies such as Google and Amazon, know that that's a mechanism they use to do goal setting, very tech centric. And I set out an OKR for every single one of our employees to build a use case in their role for the way that we are going to use generative AI, use GPTs. And in 2023, we had a version sitting inside our firewall. And so the challenge is always about how you basically get the organization both competent and comfortable with new technology. And we just went right to every employee in building that kind of understanding and trust. And it's part of the fabric of our company. It's a huge advantage for us. And so innovation doesn't sit with somebody else, it sits with everybody.

Andrew Robinson 15:35

I find that the byproduct of that is the number of super high quality ideas that then get surfaced, that then we want to invest in scale happens at a far larger quantity and at a faster pace than any other place I've been at any point my career, and that includes, by the way, 20 years of strategy consulting, and running a global insurance practice where I got a chance to interact with great, sophisticated global insurers. And I think we just simply do it better because we built this bottom up approach, and that really is the way that we've been able to not let the challenges stall our efforts.

Barbara Call 16:24

All right, that's great. Casey, tell us a little bit about your challenges.

Casey Kempton 16:29

Yeah, I think at Nationwide, maybe we weren't as early as Andrew, but I when joined there was a big push towards AI. Really, in every level of our organization. I don't know that there had been a lot of debate in our C-suite about, "Is this real? Is this the future? Do we need to harness it?" Instead, we leaned in around educating our associates on a pretty broad basis, to drive a level of comfort with AI as an enabler, as well as to work with our business leaders and our operations leaders to define specific use cases where we thought we could create real advantage, such that we could show what the benefits are, of leaning into these tools. So we have a program inside Nationwide that we call "Future of Work". And it's a requirement for Continuing Education Learning type courses, whether that be through LinkedIn or other sources, that we have to satisfy a certain number of hours. And it's a requirement that we lean into data and technology understanding, as well as specifically AI. And this is tracked, measured and monitored, and it's been that way for a few years.

Casey Kempton 17:44

We have AI questions on our pulse surveys that we do with all of our associates to really gage every year, every six months, are we improving people's comfort level with AI? And then bringing the entire employee population together for AI summits, and making information and access to resource available has been a big push for us.

Casey Kempton 17:44

So I think we're definitely making progress across a very large employee population base in and I'd say, recognizing that comfort with the tools and early experimentation by the time it gets into the rhythm and flow of our work is very important. So in terms of where we wanted to go from experiments to actually supporting and building it as part of our initiatives, one of the areas that we first got started in terms of application, was in our claims organization, and with claims, when we're on the phone with a member who's experienced a claim, through that whole life cycle of claim, from start to finish, there sometimes are various handoffs in the process, depending on the complexity of the claim, or the customer might call in for different reasons and talk to different people, and we take pretty robust notes at every customer interaction. And so it's a really simple use case, which is use AI to summarize across a very large file all of the interactions, such that as soon as the next time we talk to that customer, every interaction has been kind of bulleted out for them. We know where they are exactly in the process, and the context of what's been most important about those interactions is summarized. Now, what that does is it allows an efficiency related to how long the person on the phone has to go back and read about the case. But more importantly, right now, we can spend more time delivering on that human service that is so important to us.

Casey Kempton 17:44

So that's a simple thing that we were able to show the organization. Now, depending on where you build AI tools and applications, if you do that in an innovation area, then you still have to solve for your run system integration. So if we're running a claim center, and that's the operating system, if you will, of our claims associates, any tool that we build on day one, you may not say, let's build that into the core application. Let's build it off to the side. Well, then you have to be able to fully scale whatever experiments you run. And I think that's the place where we continue to get better in terms of understanding the early value and how a particular end-to-end solution will work, and then how we build it into the flow. In addition to that, we took on a flagship product in the underwriting space, in our agriculture business nationwide, AG, and that wouldn't be the first place you might think of, I think it's maybe similar Andrew to some of where you're really evaluating that underwriting process and saying, Where can this be helpful? And that team immediately came up with the idea on their own, and then nominated to represent PNC is one of our first flagship products, with the idea that if we could solve how AI is helping us make better underwriting decisions in our agriculture business, then we have an opportunity to expand that to other areas of PNC and so having that executive sponsorship buy in, the steering committees that we have, we're then able to see how one success, or the beginning of a successful pathway or successful tool build in one area of our company can actually be pretty quickly applied to another.

Barbara Call 21:44

All right, excellent. Thank you. So my next question, when driving towards an enterprise, AI mindset, you need every C-suite leader from the CFO and the CIO the heads of business units to be fully on board. Casey, I want to start with you. In your opinion, what are the best practices for getting all levels of management, from the C-suite to the supervisor to buy into a potentially massive and long lasting change to the business strategy and overall execution of the business goals and objectives?

Casey Kempton 22:18

I think we have to challenge ourselves on the financial side of it, that meaningful goals around what we think the benefits can be long term, and as you think about projecting forward the future, and then layering that into a business case, it gets pretty challenging, but when you set a high bar for what you expect the benefit to be, people start thinking differently and more creatively, and I think, become more willing to challenge the status quo with any kind of transformation. Which is where I've spent the lion's share of my career. in the intersection of business process, business opportunity, customer interaction and process experience with emerging technology and a very technical business of insurance. And when you think about how we get our associates, our managers, our process engineers, to evaluate a current state, what do they deem necessary and critical to the outcome we're trying to achieve, and where are they willing to innovate and push boundaries around human-in-the-loop assisted AI solutions, and so when you set a pretty high bar around the financial benefit that you see from these different AI projects, then it gives you permission to challenge that status quo and the architecture of how you're going to use AI to solve any particular business opportunity or challenge. I think at Nationwide, just being forward-looking as we are, and really leaning into both our heritage as a mutual and becoming a modern mutual there, there again hasn't been a lot of pushback that we need to do this. We have senior leaders at the SVP level and at the manager level adopting an AI-first mandate for their areas. So it isn't that we're having to continue to drive this as edict, it's becoming more organically embraced, and so then the ideas are mushrooming. While we had those 18 enterprise wide flagship projects in my business alone, we counted over 35 initiatives mid-year, 24 that were actively being worked on, some very small and some larger. And so when managers, leaders and supervisors are more organically applying the power of AI to how we do our work and what we end up delivering, I think the levels of the organization are just much more synchronized, and we have to challenge ourselves to satisfy a pretty high hurdle on the return on the investment that the whole system has to work together.

Barbara Call 25:11

That's great. Andrew, what are your thoughts?

Andrew Robinson 25:14

Well, first off, I love Casey's words when she talks about organic and how it sort of becomes self actualizing. I do think we're a little bit different just because one, we're 1,000 employees. We're $4 billion in managed premium. So it's just a different scale. And our organization is immensely flat. And as the CEO of the company, there's not a lot of distance between me and nearly every employee. And so our ability to be clear about what is strategically important tends to reach all parts of the organization very clearly at speed. And so I think we're fortunate that we set in motion an orientation towards this fantastic marriage of talent and technology. Our business is about expertise. Technology is about amplifying and widening the aperture of our people, and we keep reinforcing that and having successes. And it's to the point where we don't really talk about or converse about anything on the governance stuff. It just happened. It's infused in the core ethos, thinking strategy, the way we work as a company. And I think the critical thing is just have the organization in a position where you can execute on good ideas really fast. It has a huge amplifying effect. In terms of people saying, "Wow, being part of Skyward is an immense advantage", because the amount and the impact and the speed in which technology is being deployed feels unlike any other organization, and it just continues to reinforce. And since we're probably in our third year of cycling things so fast, I think its about terms of engagement of executives as a distant past.

Andrew Robinson 27:26

I'll tell you a really quick story, which is Jim Moore Milo runs our professional liability business, who's been one of our most effective leaders in envisioning what's possible and then having that vision embraced by others. McKinsey just did a recent report on AI in the insurance industry, and there were three or four other reports, and he's the guy who basically uses AI to generate a podcast that consolidates these four reports for my consumption. And that's kind of like daily fare around our organization when you have people running multi-million dollar businesses who are thinking about the tools that they can use in their everyday life to take their learnings and share with somebody like me. I just think that that's a unique feature, certainly for an insurance company, it's not for a technology company. But we're so far past the thought about how to engage the organization. It's really now just a question of, how do we ensure that we're able to maintain the speed of deployment on the really powerful ideas, such that people really feel that they're getting fulfilled as they continue to see these opportunities.

Barbara Call 28:57

My next question, what's your advice for CIOs and how to frame the AI investment as a strategic priority? Andrew, let's start with you.

Andrew Robinson 29:43

I feel like we're well past this. I will point to one thing, whether you just call it dumb luck or tremendous vision by our technology team, is that very early on, we expressly decided that we need to control the window pane for our underwriters and our clients, professionals and so the legacy systems, which for us aren't necessarily all that old, cannot be the constraining factor of deploying new capabilities. And so we effectively created Sky View. We've been winning a lot of awards off of this visual underwriting experience that is really the envelope that our underwriters work in. And I think that the thing I would point to is speed, speed, speed, including failing fast. And we don't think about operationalizing as something else we can do tomorrow if we discover a new risk signal in our credit insurance business that we want to deliver to our underwriters desktop. We can do that, and we can do that without having to deal with the API application of the back end systems or overhauling some operational processes. And we can measure or monitor the effects of that. And we like to try to do that with some sort of thought, but we know that the things that we're deploying are going to have a positive impact. But the truth is that if you're thinking about the most important thing that the organization can do, is the CIO organization able to operate at speed? And fortunately, early on, we built our tech stack in a way that allows us to deploy new capabilities, new services with very little work. And that's an immensely valuable asset that we have as a company. And I can't tell you how we ended up there, but I can tell you we're there which is a huge advantage for us.

Barbara Call 32:06

Excellent. Okay, Casey, what are your thoughts?

Casey Kempton 32:08

Yeah, I love what you're describing, Andrew, and as I think about the future for a large company, such as Nationwide, and what you're describing, how embedded AI is into how you think about how you work every day. It's really encouraging that you've arrived there so quickly. When I think about the CIO role, just in general, the more connected that linkage between the technology, innovation and development is to the business and the business goals and how business challenges the technology and technology can challenge the business in very constructive ways. I think is essential. I see these advancements in data in AI, 100% just blending with business strategy. I don't know how you disaggregate them. How you build a coverage form in our world, and insurance isn't as impacted as that as other areas. You need the technical experts to know how to write the legal documents, but the process that you use to do that can fundamentally change. And so it's no longer a question of, is it sponsored by technology or sponsored by business? I think this is a place where the two are 100% in sync. And so as I think about CIO organizations, and I personally, in my career, have worked in technology and have worked leading businesses and supporting businesses and operations as well, it has to come together as one. I think that's the real power of it. The dynamic in organizations have to be fully partnered. And each part of that technology equation, the business equation, it has to be mutually reinforcing. And I think the best organizations already recognize that they've broken down that proverbial wall that we maybe experience and less of an agile software development approach and more of a waterfall approach where there's that hand-over to technology. I think the organizations who have fully overcome that legacy approach to achieving business value, which is what really everyone's here to do. You talked about OKRs, Andrew. That's a critical part of how we deliver business value and achievement of that always is business and technology together. So just encouragement to keep pushing that conversation from the CIO angle and make sure that you can garner that support.

Andrew Robinson 34:51

I think the one additional thing I'd add, it fits with what case you just said, is that for a long part of my career, technology would frequently be measured in terms of very large deployments. Even modern day, there's been big iron deployed, whether it's, GuideWire or Duck Creek, or what have you and even today, there's new InsurTechs that recently got $100 million in funding that effectively is the entire front end process from the moment a submission comes in the door to handing off to the policy administration systems, which, again, is just a modern version of big iron, and things are oftentimes measured in these large deployments. I think the big difference, and it's really about a change in mindset is you're never done, even when you've had a successful deployment. This is really about having lots of things that are happening all the time, constant deployment. To Casey's point, there's no big waterfall and so forth. But even the idea of agile, it's just like this is just a never ending thing. This is an arms race. Today, I feel like we're ahead but if we stopped for six months, we would no longer be ahead. And your technology organization, along with your business, needs to share that understanding. It's pretty foundational. You can't rest on your laurels. The technology is moving so damn fast that even to the most informed, it can be overwhelming. So I think that's a big change in mindset, and your technology organization has to be up to that.

Barbara Call 36:50

All right. Thank you. Andrew Keith, would love to hear your perspective here.

Keith Schlosser 36:53

Well, I'll underscore much of what's been said by Casey and Andrew. I think the first thing that a CIO can do, and be successful at doing, is demystifying, in this case, agentic AI, for all involved. I think people go into it thinking that they understand it. I think they're a bit worried about it. Spending some time to help people understand what it is, is really important. And then a follow on, is that a lot of technology leaders will jump right to the tech when discussing the opportunity, and sometimes they use language that they understand, and maybe their colleagues understand, but business leaders and our colleagues across the the enterprise or the organization don't necessarily understand it, and that puts up unintended roadblocks for progress. And as we think about making it a strategic priority, it is really about a common language, common goals, common vision. And I think the CIO plays a big part of that, breaking down the opportunity into business terms is also really important. Less focus on the technology and the stack, and more focus on what it can deliver, I think, is super helpful. Something that I don't think has been mentioned yet expressly, but I think both Andrew and Casey hit on it, is showing how previous investments have enabled the opportunity to invest in agentic AI, and the benefits that that'll deliver. That could be around data prep, that could be around just the infrastructure in general to accept this and bring it in. I think its really important, and that helps, in my opinion, having sat before the board many, many times helps the board understand that this isn't all net-new. It's a continuation of building a continuous improvement of the ecosystem, if you will, stressing, think small test and learn versus big bang. Andrew talked about hard bringing in big platforms like GuideWire and Duck Creek. Couldn't agree more. If you start small, you get some wins. You build on them. I love the concept of the business leaders sharing what they've done successfully and hopefully, internal selling to other parts of the organization, is how I heard it.

I would avoid point solutions. When we think about strategy, oftentimes, there will be pressures on the business leaders and to deliver something quick. I'm all about that, but when you select a solution that may be too focused on a particular outcome, you miss the opportunity to have a solution that can handle an end-to-end type approach. I believe, oftentimes the back office is overlooked either in the budgeting cycle, sometimes dollars are taken away from back office improvements, much needed if you pick the right technology, agentic AI technology, you can handle distribution, you can handle elements of the underwriting process, and you can handle the back office automation and efficiency gains.

So the last point I'm going to make is that our colleagues, I believe, are scared of AI and what it could do to their careers. And if we're talking about strategic priority, strategic investment and the future of an organization, I think helping them get past that is really, really important. I love what both Casey and Andrew have talked about in regards to making them part of the process educating them, making education a mandatory part of the educational process, I think that all helps. They certainly have it figured out at their respective organizations, but ignoring the fact that people are concerned, I think will lead to program failure.

Barbara Call 41:16

All right, thank you, Keith. My next question, how can CIO is ensure that the entire leadership team is aligned on AI, ethics, compliance and security. Casey, what are your thoughts?

Casey Kempton 41:28

We take this all obviously very seriously, and because we have such a cross functional and centralized innovation approach to how we think about AI, the deployment, the security, the protections around the data, obviously compliance, we really established two purpose driven teams that were initially exploring AI capabilities pretty early on, a Blue Team and a Red Team. Our Blue Team is creating and testing the benefits of AI like where we see the efficiencies and customer service opportunities, while the Red Team considers all of the compliance risks and vulnerabilities. They're asking questions and having balanced discussions. It's really helped us make some significant breakthroughs and define the governance that we need so that we really can innovate quickly as we move forward, but also very responsibly as we deploy the tools. We know that human-in-the-loop is really important. And in our risk business, there's emotion and judgment and reasoning that we've got to bring to all of our customer centric strategy. And so in some cases, we even consider that part of the risk. We don't want to, in personal lines, or any of our other businesses, take anything away and for all of our businesses, whether it's through telematics and anywhere we're collecting customer data, always utmost importance is the security of that information and how we manage it effectively, and how we are always looking at ethics and compliance as a lens through which we execute all of our business solutions.

Barbara Call 43:06

Okay, thank you. Casey Keith, what are your thoughts?

Keith Schlosser 43:10

I feel the tone is obviously set by the Executive Committee, the CEO, the board. These are table stakes and non negotiable. The biggest piece of advice I would offer is getting the CIO involved early on, every step of the way, make sure that your Chief Security Officer is feeling like they are part of the business. They are part of the goals of the organization, and just not the person that is responsible for keeping everyone safe on the business side, I think it's critically important to do exactly what Andrew and Casey have both said, which is essentially make the business responsible for ensuring compliance and security, and ethics are evolving. It is going to continue to evolve. And I think we need to start thinking about moving away from, "Did we get this answer right?" to, "Is this answer safe?" And I think over time, that will happen. So, it's a really important part of what we do. It's an important part of a highly regulated organization and industry. And I think it's only going to get more and more challenging to meet those expectations of regulators.

Barbara Call 44:29

Okay, thank you. Great perspective. And so here's my last question with a look to the future. What's your prediction for the next big change in AI, and how can it and business leaders prepare. Andrew?

Andrew Robinson 44:43

That's a big question, because it's obviously the potential canvas for that question is just gigantic. I mean almost even indescribable. So I'll try to narrow it maybe to something more close to home, which is the world we operate in commercial insurance. I think that in the course of the next 12-18 months, the question of, "Are you using AI?" is central to your decision making in all regards. What you're doing in terms of risk, selection, pricing, adjudicating claims, making your back office as efficient as it possibly could be, I think that we'll probably see a pretty high tide across the industry in that regard. I think the differentiator is going to be a bit like, the money ball game, which is the cycle in which organizations are able to really create a virtuous learning loop, constantly adding information that can provide finer and finer insights around risk selection, pricing, how to optimize claim outcomes and so forth. Today, we're already in the process of getting the best practices across every one of our businesses applied into our decision making, because the models allow us to do that. But if I think in the next 18 months, we will probably multiply the amount of data that we're using towards that end by an order of magnitude. And of course, in the cycle of precision and so forth, gets better and better and better. And I think that you must translate that to one prediction. It's not that everybody's not going to be at some level of professional in using GPTs, large language models in their business. It's going to really be about the next turn, which is, how is that you're playing Moneyball, and being the very best around having the information that gets you super precise selection, pricing and I think that that's going to determine the winners. And I also think with a lot of talk around staff reductions and so forth, which I've heard amongst some of our competitors and peers, we have a tremendous track record of growth so far this year. We're growing as a company through three quarters, the first three quarters, 20% and I think that part of this is we can actually be a place that is becoming more efficient. We have one of the lowest expense ratios amongst the special or specialty peers, but still be the place where talent wants to come because we're growing our workforce, because we're growing given the insights that we can generate through this more virtuous cycle. And I think that that's going to start to become clear over the next 18, 24, 36 months.

Barbara Call 47:53

Yeah, I agree. It's a it's a huge question. Casey, we would love to hear your thoughts.

Casey Kempton 47:59

I'll just pick on one of the dimensions that you mentioned in there, Andrew. I think about risk selection, as well as all of the data elements that we use today to understand exposure and price exposure in any line and where we have had to rely on proxy data versus actual observations. And you think about how the market has moved without AI around, say, payroll workers compensation rating with actual payroll data, which has a lot of benefits to the customer, as well as the insurance company. But telematics and auto is an example where we're trying to have more precise exposure information to rate with more accuracy, specific to driver behavior. There are third party data sources out there today that I think offer us the opportunity as an industry to be more precise, to rely less on proxy data, and when the models are able to teach themselves about the potential insights across bigger and bigger data sets, I think we will start to get much more insight into drivers of loss in a way that I think is going to have a real advantage to customers. And I think we're dabbling in that and a lot of different places, whether it's property around roof condition or sediment condition as relates to non-weather water losses in homeowners. Clearly, in auto, we're getting a lot further on that dimension with telematics. I just think there's so much more potential in how that comes together and then how the models learn and make recommendations that we come to to rely upon with that human-still-in-the-loop. So, I'm putting my bets on on actual exposure data, allowing us to move away from proxy data.

Barbara Call 49:55

All right. Thank you Casey. And Keith? Just to wrap us up.

Keith Schlosser 49:59

So, obviously Skyward and Nationwide are advanced in this area, but I would just say, across the board, my prediction is that we're going to quickly move from more of a chat to an act when it comes to AI. Currently, we're in an era of generative AI, where models are creating text and images and answering questions and so on and so forth. But I think in the very near future, even now, it's not going to just talk to us, it's going to do very, very complex tasks. It's going to plan, it's going to reason, it's going to execute complex workflows. But most importantly, it's going to do it across different software systems, without constraint and human hand holding. It'll be important at that point to put guardrails around it and set the operating parameters in which this technology can do its work. But it's going to happen very, very quickly, and I think it's going to be a game changer. And, from the CIO perspective, and certainly CEOs and the most senior executives, the things that I think that they need to focus on is getting platforms. Andrew said it best. There are platforms out there that are huge, monolithic systems that can do everything for everyone. And they slow things down. There are platforms out there that are very specific AI platforms that can solve a particular need. I think the goal is, is to get a framework and a platform that can handle anything that an organization needs to do by leveraging the data. So that brings me to probably the most important point that's been hit on: data hygiene. Agents are only as good as the information that they can access. I think if an organization has not spent the time to understand their data, to clean their data, to structure their data, they're making mistakes. To prepare for the next wave, I think it's all about the data and making sure that you have a really well thought out guardrail system. And how can these agentic AI platforms operate in the context of your business. What limitations are you willing to put up with, or what limitations will you provide these agentic AI platforms? So in my opinion, we're moving from again, chat to act.