AI’s Turning Point: Technology That Talks Back 

Hi. I’m Paul Carroll. I’m the editor in chief at Insurance Thought Leadership, and we are here today to talk about AI, just a little subject that we’ll, you know, dispense with very briefly and and move along. As part of our future risk series, I’m delighted to be joined by Jaime Henry, who is the vice president of product management at Origami Risk. She and I have talked on this topic before, and she has a lot to say. Also with us is John Sviokla, who I’ve had the pleasure of knowing for going on thirty years now. Longtime professor at Harvard Business School, did a bunch of other things, including with me, and is now a cofounder of an outfit called GAI that is focused on helping organizations understand AI better. Most of them will tell you more about themselves and the and the companies they’re working with as we go along. So it seems to me we’re in kind of a funny spot with AI. And I I tell people I’ve been watching the same movie for almost forty years because that’s how long I’ve been covering technology since I started on the computer beat at the Wall Street Journal back in nineteen eighty six. So I’ve seen some of these things come and go, and this isn’t quite typical. Some people say we’re sort of in the trough of disillusionment partly because of a study that I think was a little off done by MIT on ROI. But I’m not convinced that we really are. Certainly, some expectations have come and gone, but I think a lot of progress is being made. So with that overly long preamble, Jamie, why don’t I I start by asking you where you and your customers are in the in the AI journey. Yeah. It’s been really interesting over the last two and a half years to watch how things have evolved. Last year, what I was consistently seeing is that interest was very high, but adoption was very low. And where clients were really focused, and and this was both within, clients we support on the pure insurance side, so insurance carriers, MGAs, and so forth, and even our clients that we support on the risk management side, which, you know, might be a Fortune one hundred client, they’re really focused on just making sure that they were getting their their policies and procedures in place around the expectations of of, you know, that initial iteration of generative AI use. Now as we get into, particularly the last couple of months in the summer and going into the fall, interest remains very high. But what I’m seeing is that the education has increased, and the amount of questions, the complexity of the questions we’re now receiving as clients are starting to understand how their vendors, are building AI and making sure that it’s happening in a secure, way so that they know that their data and their users are protected has evolved, and it’s been really exciting. And so while I agree that I think we were in the trough of disillusionment, I believe that we’re also on our way out of it. That being said, I think that we clients, and and even as general population of consumers, we’re still struggling to figure out exactly how to integrate AI into our everyday life. And that change management curve of, early adopters versus, you know, all the way to the laggards, I think we’re still in the early adopter phase. And what we need to do, is help our our colleagues, help our clients start to move through that change in a shrink curve. So, John, how about if you put this in some historical perspective for us? That’s if I go back forty years, but the last time you and I were talking about this, you’re going back to the thirteen hundreds in Turkey. So maybe put this in the broad sweep of history for us. Yeah. Sure. Well, first of all, it’s it’s lovely to be here. And, you know, my I’m very fortunate in my professional career. You know, our our business is to basically help people understand what does all this mean for them their own careers as individuals, their businesses, and also, their communities. And so, so we do a lot of research, talk to a lot of folks, do a lot of publishing and so forth, and a lot of education. We’ve seen that big uptick in that lately. So my background, I get, you know, if you look at the long history of automation all the way back, you know, as you say to there’s some, you know, reports of early robotics, you know, animatronic kind of creatures back in, you know, thirteen hundred and I think even before around six hundred, just mechanical devices that try to look alive. But if you go back, let’s say, to the mid eighteen hundreds, so late eighteen hundreds, and you say, what’s been the history of automation? It’s been to take things out of take power and effort and so forth out of animals and out of people and stuff it into the machine. And this is very different. November twenty twenty two at scale with ChatGPT three five is really the first time the machines talk back. You know, really and and that has profound implications for a number of reasons, primary among which is that the operating system of society and the operating system of organizations is conversations and stories. And now for the first time in history, we have a fluent in every language, fluent conversation partner that’s got a silicon base. Right? So we have a silicon hive mind that we’re now talking to, and that is, that’s fundamental. The second thing is we have to remember that much of the industrial revolution was a knowledge revolution. Everybody knows about Henry Ford and the moving, you know, production line that he got from a combination of, you know, the the millers and the butchers, you know, the the slaughterhouses. What only some people know is he paired up with a guy by the name of Frederick Taylor, and Frederick Taylor was the person who invented scientific management. So, you know, the old, you know, time and motion studies where somebody stands through the stopwatch and watches. And why was that profound? It was profound because up until that point, through the long arc of history, knowledge of how to do stuff sat in social groups, usually in the Middle Ages for sure and forward, the guilds. So you had the Silversmiths Guild and the Hoopers Guild, people who made barrels, and the blacksmiths and so forth. And they passed it, on in that guild, and you came in as an apprentice, worked your way up, did your masterpiece, and became a master. Okay? Taylor ripped that apart. That’s why Karl Marx hated Taylor because he was taking knowledge out of labor and putting it into the organization, into capital, into management. And we’ve been living on that. You know, you look at this coat and this shirt and my glasses and everything would be so much more expensive and so much lower quality if it weren’t for Taylorism. And so for the first time, we now can do not only Taylorism on symbol work, but these things can explain themselves, and they learn on a machine speed, billions of, you know, of, instructions per second, not just on a human speed. So if you look at the practice of management as having at its core, one of the big pillars is knowledge management. This is radically different. So we yeah. Somebody observed early on that AI has been around for a long time. This is the first time it’s had a voice. It also has eyes and ears, in in a way that it never did before. And I and I’ve been struck, just to add a little bit to what you’re saying, you know, in high school and so forth. People talk about the industrial revolution as sort of this, you know, James Watt radically improving his team engine and suddenly, wow. But in fact, were all kinds of things that had to happen. Because in those days, the the mills were run by wheels, so they were all by rivers. Right. Well, once you had the steam engine, they didn’t have to be by river. But then people had to figure out how to not only have rotary motion this way, but motion up and down. So then you could do things on floors and factory. But somebody had to invent the LLC so that you could actually innovate these things and have them fail and not crush the society and so forth. So there are all kinds of things that have to happen after the technology breakthrough to implement these things. Absolutely. So, Jamie, then my question to you is how are you doing that, and how are you helping your, customers do that? Yeah. And I think it actually goes back to the MIT study that you alluded to and and the the ROI and some of the failure rates that I’ve even heard kind of prior to that that people have been reporting. And so, as we saw, we saw a lot of businesses rushing out AI solutions really as a marketing play. Right? And to say that we’re we’re at the forefront of this, but they missed the mark. They didn’t necessarily serve the users where they needed to be. They didn’t, solve the problems that really needed to be solved, the problems that were really painful, nor did they hit the level of accuracy that helped support our our users and us as just people in society to say, like, this is really useful for me. I want to adopt it. This is worth it for me to change my habits to go adopt. And the other thing I would add to that, which is a conversation I’ve been having with people a lot lately and and talking about the industrial revolution, really, is also the fear that’s at play. The fear of this, you know, replacing jobs and how is it gonna change the future, and what does it look like for future generations, and what can we look to the past to teach us? And so what I’ve been advise advising both, people on my team as well as clients of this is inevitable. You would be in the best position to make sure that you’re using it, making sure that you’re educating it, making or educating yourself, excuse me, and identifying the use cases in that it is value add with you staying in the loop. So human in the loop is critical. I think we’ve all heard that. I, deeply believe it. And I talk about it as enrichment, as additive, and as, allowing us to make sure that we’re saving our critical decision making bank, if you will. We only have so much power to make so many decisions in a any given day. So how can we use AI to get us started? How can we add a use AI to enhance the work that we’re already doing, and allow it to empower us versus scare us? John, I know you’re a big believer in hands on. I mean, would you talk a little bit about, how that works with the executive education you do, and then more broadly, how you’re trying to get people to think about reorganizing and doing some of these sort of, you know, factory level, LLC level innovations to actually integrate AI. Yeah. Absolutely. The, it’s really important to understand that this is a bifurcated market, that we have a small number of people who are going really, really, really fast, and the majority of people not doing much. Right? They’re they’re running into, issues. And so whenever you hear an average, you know, from the MIT study or whatever, it’s not useful because what you wanna know is not, you know, not what the average is doing, but what are the best doing. Right? And, technology is an amplifier. So, you know, getting a bigger amplifier is not gonna make you a better guitar player. You know, it’s, you’re gonna annoy more neighbors. Right? And so so looking at averages is really a fool’s errand in this, and and I think all executives need to understand that. In terms of, hands on, it’s absolutely necessary because nobody’s ever experienced the technology. Anybody working today who hasn’t used this stuff has never experienced a technology like this. And, my former colleague, Jim Cash, who’s a professor at Harvard Business School when I was teaching there, used to used to say talk about visual literacy, that you can’t have a management team that that, doesn’t have enough hands on experience with the technology so they can actually recognize an opportunity. Not that they can do it, but they can recognize it. And this is, and so in our work, the the firms are doing the best. It’s a combination of top down and bottom up. So you have a top down group that gets hands on experience at least a half a day or more, preferably a half a day or more a quarter, but at least a half a day of work, working real problems because a lot of people don’t understand how to use these devices. What they do is the equivalent of walking up to one of those giant pipe organs with the three, you know, keyboards and the forty stops and the twenty seven pedals. And they go up and they play chopsticks on the middle board and say, yeah. Not that much different. Who’d pay all this for the organ? You know? I can do it with an electric piano. It’s like, no. No. No. You have to, like, play this thing a little bit before you understand what it can do. And, so that’s our job is to get them to at least, you know, hit a couple of keyboards and some foot pedals and a few stops, right, in terms of doing very practical stuff. The second thing is that I absolutely believe in the organizational maturity model, and it’s a pretty standard for those of us who’ve hung around, you know, consulting in in academia, everybody’s got a maturity model. Right? Okay. So here’s ours. It’s basically, do you have experimenters, people doing individual experimentation? Do you have pilots, that are doing or even production that are islands of automation, kinda level two? Level three is what we call orchestrators, people who basically have created, you know, a platform and then build on that platform. So Goldman Sachs, Blue Cross Blue Shield of Michigan, you know, so they’ve really committed that, you know, they they understand how to move expertise around the organization and so forth. And finally, what we call intelligence leverages. Those are people who are using AI to build AI. Now we’ve only seen those folks right now in technology, but we believe there will be intelligence leverages in most major industries. And the the key part for executives to understand about intelligence leverages is that they are gonna skew the profit pool in their industry toward them. So my prediction is, for example, I used to work at PwC. I no longer have a financial relationship with them or anything like that, but I used to work at PwC. And so the big four, I believe, have a profit spread that goes about if if a hundred percent is all the profits available for the big four, and they’re in a separate market from anybody else because there are certain things only they can do, like, you know, big company public, accounting, for example, that those people right now, the spread, I believe, from top to bottom would be roughly thirty five percent, twenty five, twenty five, twenty, you know, sorry, thirty percent, twenty five, twenty five, twenty percent, k, across those. So they spread the profits pretty much, you know, not that differentially. My prediction is within ten years, maybe even within five, the firm that really transforms itself is gonna grab fifty percent of the profits, the better talent, and be in a whole different learning curve than the other three. And that’s gonna happen in industry after industry. Now I know insurance is very, parts of it are very concentrated. A lot of it’s very diversified because of the regulatory, structures. But, anybody listening to this should really think about that because, you know, right now, you might, you know, be eating deer and wild boar. If you’re not careful, you’re gonna be eating chipmunks and cockroaches. I I love that, John. Yeah. I think you know, I’ve been thinking about that deeply. Right? So on two fronts. So how do I, from an operational perspective within my own business, make sure that we’re utilizing AI to, frankly, stay ahead. Right? Yes. But then importantly, how do we take this technology and put it into our products so that our customers, our clients know that they are can take advantage of what we’re building to then push ahead their own businesses. So many of our insurance clients are running their entire businesses off of Origami. And they’re and like everyone, they’re looking for those efficiencies. They’re looking for that productivity to help them move any levers that they can in order to improve their revenue. Right? But just a basic need of of of any business. And so to kinda tie back, I think, to some of the things that you were calling out is as we’re producing new AI features, one of the things we’re trying to do is is partner it very closely with another technology staple of of excellence in user experience. So we wanna make sure that we are providing these capabilities that aren’t overwhelming, that aren’t confusing to the end user because we need to ensure that we’re putting it in front of people and limiting the decision they have to make about using it. We wanna create functionality that is just simple, it’s intuitive, and it is producing that acceleration that AI is offering us. So we’re one of the things we just released, for example, is, our AI risk and control explorer. So it actually helps enterprise risk managers identify potential risks facing their organization through just a simple prompt, a description of their organization or a situation they’re in, and uses AI to generate, both potential risks based on that their organization, their industry, and that specific situation that they might be digging into, and then this the controls. This is such a great example of the benefits of AI because what enterprise risk managers were doing otherwise was just thinking. Sure. What might the risk be? And getting started is so hard, or it’s very easy to miss something. It’s so much easier to start, you know, from a piece of paper that already has some words on it. Absolutely. And and also having a conversation, there you know, there’s pretty good, you know, academic work that says some people learn by talking, some people learn by listening, and some people learn by reading. Well, the cool thing about, you know, AI is, okay. Tell me how you wanna learn. Right? You wanna talk to me? You wanna and and we’re just not used to having technologies like that that that really up the learning curve. The the other thing I’d I’d just like to pause and and emphasize something you said before, Jamie, about a lot of people don’t understand that this is different economically and different in terms of inevitability. So we go back to the dot com here. Right? And Paul and I know it well. You know, you had a lot of overinvestment. You also had some funky stuff going on because the investment banks were on both sides of the trade. Right? They’re they were doing they were doing the issuance, and they were doing the the promotion. And so there’s a lot of, and none of those guys really suffered from their crazy promotion, which is, I think, bad from an enforcement standpoint. But, anyway, you know, so you had pump you had on the inside of Newbetter who are pumping it up. You had people who are buying stuff way in it ahead ahead of, revenue. And I would say the Internet was relatively small compared to what we’re talking about here. Was about transactions and information availability early on. Okay. And still people made zillions of dollars. Right? We’ve never seen companies like you know you know, OpenAI has billions. Anthropic has billions. Cursor went to a hundred million bucks, I think, in three months. This is real revenue coming in. Okay? So they have that. The second thing is that these companies that were little companies when, you know, when the Internet started, they are bigger than countries. Okay. John Sherman, the guy who wrote the antitrust act to come to Sherman’s brother, in there, he was worried about companies becoming as strong as governments. The folks in Chicago School of, you know, economics, you know, kinda pushed that to the side, which has really done us a disservice. But he Sherman said, look. Why? We just fought a king. Why should we bow down to a company that’s gonna give us something essential? We shouldn’t bow down to them like a king. Well, guess what? Google, Meta, they’re kings. They’re telling countries to go stuff it. I mean, if France comes after them, stuff it. Brazil comes after them, stuff it. These the the combined market capitalization of the of the six companies I’m leave Tesla off because I think they’re a meme stock. But the six companies of Meta, Google, Apple, Microsoft, and, NVIDIA, and I’m forgetting one. Anyway, those six have a market capitalization over fourteen trillion dollars. You have to remember that the combined GDP of Japan and Germany is about eight point six in twenty twenty three. Google creates, just, no. Microsoft creates just under two billion dollars of EBITDA a week. So when the Europeans put down, you know, oh, a billion and a half dollars, it’s like, okay. Four days. Thank you. You know? And the you know, when they talk about, oh, they’re spending billions. Thirty percent of Microsoft’s cost base thirty, forty percent of Microsoft’s cost base is software creation. We know this stuff makes people more productive in software. So they can spend billions. They even if they make no money in OpenAI, it’s gonna be an enhancement to them. So you have that. The second thing is anybody who’s been paying the least bit of attention to the Ukraine and the Israel conflicts, Gaza conflicts, can see that the entire doctrine, provisioning, equipment, training, intelligence, preparation, aftermath is completely getting recreated by AI. So you have the most profitable and powerful companies that have ever existed on the face of the planet. And by the way, in twenty twenty three alone, the five of them generated five hundred and twenty nine billion dollars in EBITDA. Okay? And the Russian government, from my intelligence friends, collects about three hundred billion. Okay? So these things are more powerful than governments. They’re more profitable than any company we’ve ever seen. They influence the entire political discussion that happens. They’re at the leading edge of intelligence and, spycraft. This is this this is nothing like the Internet. I mean, that’s the difference between, you know, riding six foot waves, you know, on La Jolla shores in, you know, Southern California versus going to Nazarene and riding a hundred foot wave. These things are coming in. Boom. I also think we ought to separate the stock market from the things that businesses are doing with AI. I mean, in in the history of bubbles, people are throwing a lot of money, and not all of that investment can probably be justified because it’s you know, you talk about the concentration you see going with the big four potentially. That kind of thing is going to happen at some level with the AI companies. But the the pot of gold potentially for the winners is so big that they can justify what from a macro standpoint, looking at all of them is irrational. But whatever happened, that’s goodness for those of us who are trying to use AI because we have all these super smart, super powerful companies spending all this money to advance the state of the art. So to me, that’s I mean, at at some point, I think there will be at least a, know, some sort of retraction in the market, if not a crash among the a s AI stocks. But even when that happens, I don’t think it’s gonna derail what’s going on with the those of us who are benefiting from AI. Yeah. Absolutely. Look. First, let me make two comments. First, in terms of stock, it drives me baloney because, you know, drives me nuts because I used to teach some decision analytics. In any way, if it has a huge payoff and a small amount of of, probability that you will win, it is rational to overinvest because everybody’s buying a lottery ticket on a rational basis. So I have a one in fifty chance of winning a thousand times more with my money. Guess what? Most people are gonna lose. So the hype cycle is totally wrongheaded. It’s not about existing businesses doing marginal investment. It’s about people buying lottery tickets for a massive payoff, and you’re always gonna overspend on that. So it’s just now our insurance friend should understand that. Right? It’s not. That’s how investing works. Anyway, it’s just so the stupidity in that market should’ve just been crazy. Right. So So one one little detail there, don’t what you’re saying about the hands on stuff. It reminds me of one of my favorite IBM stories. They sort of discovered the email back in the late eighties, early nineties, and the CEO did the right thing and mandated that, his senior team start using email. And then he started getting these things that were done exactly as in our office memos. They were still dictated to the secretary. The secretary still typed them on the computer. They just, you know, were sent around that way. So that literally with people playing chopsticks in the middle of the, the pipe organ. Yeah. So, I I think it’s Can I go to second piece about Yeah? Yeah. The the the so, look, on the ground, you know, we spend, twenty three out of twenty four hours a day worrying about the leading edge of the practical. What are we seeing? We’re seeing first of all, these a really simple way to think about these things is these are power tools for symbol work and unstructured data. So if you hired somebody to put a couple of, outlets in your house and, like and you came home from work and the electrician’s there with a hand drill going like this, you’d be like, what the heck? I’m paying you a hundred and fifty bucks an hour, whatever I’m paying you. And and power drills. Right? No roofer doesn’t have a nail, you know, a nail gun. You look at any you look at any, you know, any construction site, it’s full of power tools. Okay? If you’re walking down a cube farm and people are sitting there working on computers and, you know, doing actuarial work or whatever, you know, marketing, any symbol work, they’re sitting there with hand drills if they’re not using AI. And you can see it for companies that are think that they can control this by telling their people not to use it. The surveys are saying thirty to forty percent of people are buying AI themselves and bringing it to work or using it home and bringing the output to work. This is the most important career accelerant of the ambitious people in your organization than they’ve had forever. And you think the kids coming out of school aren’t gonna have this, you’re nuts. And, so it’s just that horse has left the barn. So the question is, which one are you gonna use while you figure out how to make it more productive? Yeah. It’s Yeah. I mean, it’s so beautifully said. Yeah. I’m I’m seeing that within my own organization and recognizing that, you know, as a product person and driving a software product forward, I have to give those tools to our clients as well. We need to make sure that we’re putting this in. And all of this ties back around to where we started, and, John, what you called out. The importance of leadership guiding their teams in the inevitability of this, recognizing it’s happening whether you’re, you’ve set aside the time or you’ve set a strategic goal for this year or not. It’s a it’s a strategic goal that might be passing you by, and you need to to focus. And so it’s been really important for us to set the stage for our internal colleagues as well as for our clients of saying, adapt it. We’re giving you we’re giving you different tools. We’re giving you different operational processes to support your success in this. Absolutely. So what are some examples of companies that are doing this really well that people ought to look at? I mean, maybe, Jamie, I’ll I’ll start with you, John. I know you, like Jerry and Blue Cross Blue Shield and probably have a couple others. But, Jamie, what what are, among your customers, without names necessarily, examples you see of people doing this really well? I would say the first thing starts with that, the expectation that’s set in the beginning and those those leaders who do understand that it’s happening and it needs to be adopted. I’ll ask the question of some of every client that I meet. Where are you on your journey? And oftentimes, it’s not yet been a focus. And so, those clients who are doing it well, it’s quite simply a focus. Even if they don’t have the perfect answers yet, even if they don’t have all of the processes and the explicit tools they’re expecting at the individual contributor level, of how they’re gonna use it. Because not all the tools are out there. Right? Not all the tools are out there yet for every job. The tools that are readily available are a little bit more generic in nature, whether it’s Copilot, where it’s ChatGPT, and you just encourage people to think of their own use cases and take advantage of those tools to kind of target what might be their own use cases, say, just generation of of marketing content as an example, or one of the tools that we first rolled out was generation of emails embedded within our application, making sure that you just gain that tiny micro acceleration just through an effort of helping you generate an email, helping you make an email more brief, helping you change the tonal direction of your email to perhaps be more empathetic, which in a claims organization is critical and and, as you’re trying to move a claim towards closure. So that’s what I would say to the companies that I’m seeing that are doing this really well are are ones that have simply accepted the reality. Don, how about you? Yeah. Let me let me give you a couple of examples. So one is a company called One Digital, and, they do, employee benefits, retirement and wealth consulting, HR consulting, well-being solutions. Okay? They’re owned by private equity, and they, they launched a number of, months ago, actually almost a year ago now, eight different, digital employees. So you have someone who can help you with explaining benefits to a small business, someone who can help you learn how to close a sale, someone who can help you figure out your own benefits and capabilities. And now they’re up to a couple of dozen, and, they are deploying them. And their whole thing is not about saving money, although it does save them money. Their whole thing is to make the very best producers better. And anybody in insurance that’s got a Salesforce understands. You know, the top five percent probably drive a hundred and five percent of the, you know, economic value. Right? And so imagine taking those people and making them superior. Right? They’re excellent. You’re gonna make them superior. And I think they’re I usually don’t like war analogies, but the analogy here is is, special forces versus regular army. Right? And the way you take special forces, you take them up a level as you cross train them, you give them technology, you give them decision authority, you make sure they have the same values, and they are vastly more productive than regular army. Okay? That’s what’s gonna happen here. And so One Digital does that, and they have found that the people have curiosity and drive. So openness you know, if you were to take the the, what’s that? The the ocean model, the psychological model of openness, conscientiousness, Anyway, a couple other things. Right? Neuroticism and and that basically, it’s a profile of, you know, are you open to new things, or do you complete it? Are you neurotic about it and, you know, controlling? The people who are more on the openness and curious scale and actually are conscientious, get stuff done, those are the people who are taking this stuff and running. So when they implement, they look for those who are open and curious, and that’s moving the top end of the distribution. So my question to every insurance company is, in your critical areas, underwriting, sales, so forth, claims, not the whole organization, but if I could take the top ten percent of my organization and move them out of full standard deviation, I can use words like that here in insurance. Right? Move them out of full standard deviation. How okay. What kind of performance would I have? Very different management problem than trying to get the great gray mass of the middle to move. Right? This is about making the excellent superior. So that’s one. Second one, Blue Cross Blue Shield of Michigan. So Bill Fandrich and his crew, you know, he’s he’s, one of those rare IT executives that’s great on the operations side as well as IT. And so, and his title reflects that. He sits above IT and operations. Anyway, Bill, has been working for, many years about creating a, you know, good data platform and good, you know, governance. And, and, you know, in that business, if you get it wrong, you know, it’s not like Wall Street where they just kinda give you a slap in the wrist. You know, you actually end up in jail, like, in a orange jumpsuit, right, if if you’re not careful. The penalties are pretty severe. And so, in a five year period, he took them from the highest cost among the Blues. I think there were thirty three Blues now still. The highest cost per employee on IT, down to the lowest cost per employee on IT. Okay? And they’re using and they’re at what I would call the orchestrator level, and they’re starting to get into the intelligence leverage level, which is they’re using AI to route, you know, things like benefits requests, you know, doing approvals. There’s always a human loop on the approvals, but, you know, a lot of times the approvals and the and the benefits eligibility, you know, stretches out because of a combination of having to find out if the individual is the individual, what can I share with them, then looking at the policy? And, you know, they’ve got, over, you know, thirty million lives. I mean, you know I’m sorry, over five million lives and over thirty thousand producers. So really complicated, right, as to are you still covered and so forth. And and inside, they’re using AI to not only contextualize things, but to be able to lead the data there. So they don’t have to create a giant data lake, pull out the data that they need, do the inference that they need, and use it as, in a lot of cases, API calls, right, application program interface calls to over four hundred systems internally so they can rewire the processes to give a good customer experience. So they’re really doing a heck of a job. Now on the underwriting side, if you look at their financial performance, not very good. That’s a product design and some issues around politics and regulatory things that they can and can’t do given who they are and where they compete, but on the operating side. So I would say both of those, you know, unlocking individual contributors and really creating a new platform of growth. John, what I like about the theme that you just called out there was it it it those two success stories inherently kind of attack the fear that individual contributors have. Yes. Right? They fear being replaced. And what you just described was you’re you have the wrong lenses on. Yeah. This is a door opener. This is an opportunity for for growth, you know, not only for an individual’s career, but their entire organization. Yeah. And what I’m seeing is some leading companies, upstarts, they’re starting to expect when they do interviews for, in this case, for software programmers, they wanna see your resume, but they also wanna see what robots have you built. So kids coming out of school right now I I just did a talk over at Northern Kentucky University where they do a lot of, training for the for, you know, Kroger and since and they’re right there in Kentucky right over the river from Cincinnati and, you know, for so Kroger, P and G, and so forth. They said, look. You have to be graduating kids with robots. Okay. Show me the three or four or five, you know, the chat GPTs you’ve done or the Gemini gems or whatever or or, you know, the Python code you’ve written. Because it’s every knowledge worker, every symbol worker is gonna be like a chef. Now I’m not in the restaurant business, and you would not wanna eat my cooking. But my understanding is you never hire a chef without their own knives. A chef coming in without their own knives would be like, what’s everybody’s gonna have their own knives. Every knowledge worker’s gonna be like a chef. Show me your knives. Okay? And that’s gonna be part of the interview process. It’s gonna be a really I mean, just imagine as an employer because you look at, you know, early stage is coming down, right, one to three years, and just about and lots of different professions, but certainly software. So how do you become how do you look like somebody who’s got three years experience when you’re coming out of school? Robots. Right? Open up your knife case. Say, look. This is what I got. And you will I guarantee, same resume, this person has robots, this person doesn’t, this person’s getting hired. Yeah. You’re inspiring me even as a parent. What do I need to do to set up my kids? Absolutely. Yeah. Well, to to me, this changes the whole dynamic. I I’ve always been a little skeptical of change management programs because they tend not to work. As as Chuck and I Chuck and one of my frequent coauthor and friend mutual friend of Johnson and I wrote one point in a book with me. Everybody loves change. It’s a change part. But if if you don’t try to bring everybody along or insist everybody come along, you just go to those super producers and make them more efficient. Everybody can see what’s going on, and they have to get on board. If the hiring process changes, so you show up with your knives, your robots, whatever, then that changes the whole incentive structure as well. So you sort of lead from the front and then let everybody figure out what they’re gonna do to to follow behind you or or not. Yeah. It’s it’s it’s really funny that I think people I think people don’t understand what the dynamics are really gonna be. Everybody’s worried that, oh, you know, people are gonna be afraid they’re not gonna adopt it. That’s not gonna be the problem. The problem is your best people are gonna adopt it like crazy. And how are you gonna be able to pay them what they’re worth when they’re fifty times more productive than someone else without ripping apart the fabric of the organization or them spinning out and taking some of your best customers and best people with them? That’s the dynamic. It’s not it’s not, you know, how do I get every, you know, every Tom, Dick, and Harry on this thing? It’s like, forget that. This is about pulling the front end out. And and, I mean, it’s not good news for our society because we’re not making the investments in labor liquidity that we did at the end of World War two, the GI bill and lots of other things that, know, a lot of you know, college was cheap and all that other stuff. And, you know, health care wasn’t as as expensive on an absolute and relative basis as it is today. So at a societal level, I think we have to commit to and literacy and the creation of common good and common properties, data models, ecosystems that are open for our citizenry. Yeah. The the same way it’s kinda like a GI Bill for AI, if you will. That’s but that’s not a a problem. If you’re an individual company, the issue is that, you have to worry about, are your best people gonna leave you? And can I talk about one particular interesting soft underbelly of of insurance, which is really scary when you wrap your mind around it? And, actually, this was, put in my head by a a dear friend of mine. I I don’t wanna mention her name because I haven’t asked her permission from a brand name organization. And I think it’s a brilliant insight that she had. She said, what happens if AI can copy or duplicate our hundred and fifty year history of data? So all this entrenched stuff in certain of our lines, right, our long history really gives us a competitive advantage. And, of course, all the algorithms and pricing and everything around that. Imagine a synthetic data can recreate that like that. That’s a really interesting question. You know? And you look at for those who aren’t into it, like, a lot of the progress of things like AlphaFold in biology and AlphaGo and Go, right, were reinforcement learning that was done on synthetic data. And you look at there are providers out there like Seeker. It’s a a customer of ours. And, you know, Seeker has got the ability to as you’re cleaning up your data, say, oh, I don’t have sufficient data. Let me generate data. You’ve seen stuff in marketing where people are doing They’re they’re doing market tests on synthetic data. It’s like, oh my goodness. If I can do that in underwriting, that’s really scary. I mean, if I’m an incumbent, because it’s like, well, wait a second. All this all this I mean, data lakes, data houses, data boats, data water skiing, you know, all that stuff they do on the data lake, it’s copied. And that could get really scary. Yeah. No. I think we’re just at the beginning of this, and we’re, you know, getting it sort of implemented at low levels, and then aspirational companies are doing it at some higher levels and so forth. But, yeah, at some point, we get to that sort of level that that you’re talking about. So I I could do this all day. I find this fascinating and and love talking to you guys. But I think we’re sort of getting toward the end of our time. Ordinarily, I’d ask for practical first steps, but you guys have already sort of gotten into that with the, you know, try it and then going through some of these steps of Dawn’s maturity model and sort of journey that Jamie described. But just any final thought. Jamie, why don’t I start with you? Just, you know, we we we basically have ripped open the universe and said all things are possible. So you’ve got thirty seconds. Summarize it. I would summarize it by by doing that call to action again. Go use it. Don’t fight it. Embrace it. Play whether it’s within you know, there’s a number of applications. So whether you’re an individual contributor or you’re a leader, request on that the AI capabilities within your software stack are turned off and then prioritize the time you need to start taking advantage of it. One thing that’s been fascinating for us is that we are going down an effort where we are trying to democratize AI, where we’re trying to make sure that the business workflows that our clients have, that we can put in these points of AI, whether, you know, generation, summarization, things like that, at the point that the users need it without them having to wait on us to hard code it. And one of the things that we have found in doing that when we present this kind of system that we’re gonna that we have, that it immediately inspires every single person we’re talking to of saying, well, here’s my exact use case. Could you do that? And so people are very quickly inspired once they’re given the framework. And so I I love what John said about kind of those top performers. I think it aligns very well with that. Like, go seek it. Open up the potential. Be curious. K. John, bring us home. Yes. For anybody interested in AI and generative AI leading an insurance company, I’ll be inspired by the words of the greatest samurai ever to live, who is Miyamoto Musashi. And he had this quote, is fantastic, which is we must have a close view of distant things and a distant view of close things. And that is absolutely perfect because a distant view of close things means you need to get your hands on it and understand it and have a feel for it. And this guy had seventy two fights and obviously never lost because you lose, you lose. Right? Anyway, the and then, a close view of distant things because the fundamental economics of learning knowledge actuarial client service, the whole routine are changing. They’re changing as much as when mister Swift decided to slaughter cows in Chicago and ship them out dressed, you know, in refrigerated cars, and that radically changed the economics of of butchers and slaughtering. And you saw massive centralization. Okay? Same thing’s happening here. Symbol work is gonna fundamentally change education, knowledge transfer, customer service. And if you’re not thinking about what this means for the reinvention of management and please, please, please ignore the Gartner hype cycle. Ignore the averages. This is not a story of averages. It’s what will the elite in the segments you compete in gonna be able to do? Because I am telling you, ten people are gonna be able to do the work of a thousand or ten thousand. Alright? Just like in special forces. I had the chance of, to talk to, boomer Stuffelbeam, admiral Stuffelbeam on the eastern Mediterranean fleet, western Mediterranean fleet. No. I’m sorry. Eastern Mediterranean fleet in the second in the first Gulf War. And he had he was supposed to go through Turkey, but then the Turks chickened out, and he had to land in Northern Iraq in the Kurdish region. He put fifteen hundred special forces soldiers on the ground against ninety thousand Iraqis. Those are the numbers we’re talking about. So as a leader, you have to have that close view of distant things while you’re busy doing the distant view of close things and getting your hands dirty and understanding the top down, bottom up educational process. Yeah. Top those two. So I’m just gonna, say thank you guys for for talking again. I I found this fascinating. I hope our our viewers do. I will add one call to action for our viewers, which is, please share. I think this is a great stuff, and I think if you have any folks in your organization who are the slightest bit recalcitrant about AI and and know and understand the possibilities, I think this would be a great eye opener. So thank you, guys. We’ll put up a slash screen at the end so that people know how to get in touch with you and, of course, those of us at the insurance thought leadership. So thanks, John and Jamie. Thank you.