Skip to content
 
 

[00:00:00.00] [MUSIC PLAYING]

[00:00:15.52] REID HOLZWORTH: Want to go ahead, Nick?

[00:00:17.05] NICK LAMPARELLI: Introduce myself, or--

[00:00:18.16] REID HOLZWORTH: Yeah, do some intros. And then we'll get into it.

[00:00:20.57] NICK LAMPARELLI: Sure. Nick Lamparelli. I am the managing director of Insurance Nerds, which is something you guys might see online. It's a media and content organization. And we're focused on career development, but also marketing for B2B companies in the InsurTech space.

[00:00:39.80] CHIP BACCIOCCO: Hi, I'm Chip Bacciocco. I'm the CEO-- that's better. I'm the CEO of trustedchoice.com. I've been in the insurance business my entire career. Trustedchoice.com is mostly focused on helping independent insurance agents to-- it used to be survive, but now it's thrive in the digital age.

[00:01:00.75] REID HOLZWORTH: I'm Reid Holzworth. I'm the CEO of IVANS, and it's going to be fun. So, Nick, what did you learn? What did you take away from today? Let's start with you.

[00:01:08.05] NICK LAMPARELLI: Well, I'm particularly interested in the telematics play. I think my thesis has been, we have all of these sensors. And it's taken a good 10 years for telematics to finally get some traction. And I'm curious as to what you guys think about the next step, which is property sensors.

[00:01:31.20] So smart devices. So we got ring doorbells that can catch thieves. You can talk to them as they're breaking in your door. But we also have water detection devices. We have smart smoke detectors; motion detectors for HOAs, where they can see if someone's actually entering the pool area. And so I'm wondering, is it going to take another 10 years for us to be able to incorporate that into the insurance ecosystem?

[00:01:58.61] REID HOLZWORTH: I mean, personally, my homeowners, my care-- I'm not going to name names-- almost mandated the water detection on my house, literally. And I didn't end up doing it. But I can see it. It's coming. I mean, they were pushing hard for that with me.

[00:02:13.50] NICK LAMPARELLI: What kind of incentives?

[00:02:14.89] REID HOLZWORTH: It's premium, but it wasn't enough to warrant the cost and doing it and dealing with it. And I kind of liked the idea of it, frankly. But I'm kind of a non-standard risk.

[00:02:25.02] [LAUGHTER]

[00:02:26.12] CHIP BACCIOCCO: Yeah, I've been to Reid's house multiple times. He's a non-standard risk. Yeah, he has a full-sized motocross course on the property.

[00:02:36.47] REID HOLZWORTH: I do. This is true.

[00:02:37.73] CHIP BACCIOCCO: Yes. Yes. And how many race cars?

[00:02:41.50] REID HOLZWORTH: Not enough.

[00:02:42.79] [LAUGHTER]

[00:02:45.40] CHIP BACCIOCCO: Anyway, I thought the speakers today were great. I am struck by how far and how quickly we are already talking about adopting AI. I mean, let's just-- AI was the big topic today. It's the big topic in the industry.

[00:03:02.09] I looked it up this morning. And ChatGPT 3 or 3.5 or whatever, the one that was publicly launched, that was only 17 months ago. And generative AI is already moving into everything. And we almost can't build all the ideas we have and applications for it fast enough. But it's really encouraging to see the speakers that were here today and how they're thinking about how to use AI and where we're going as an industry.

[00:03:32.10] The last time-- and I'm old, I guess. This is going to be embarrassing. I remember Netscape Navigator. But it was absolutely transformative, as we all thought, wait a minute, all these green screens could be replaced with web applications. And people say, well, what does mean? Anyway, this now, this time we're in feels to me as like that time felt, like that an absolute open vista was opened up in ways that we couldn't-- as quickly as we wanted to, we couldn't start to think of all the ideas, all the ways we would use it.

[00:04:11.44] So I'm known for saying we live in the golden age of insurance innovation. And now what do I have-- say, the platinum age? I don't even know. It's going to be exponentially bigger. And I just thought the talks we heard today, the comments were insightful and great. And we're going to sing about it later. That's how good it was.

[00:04:32.59] NICK LAMPARELLI: Is there anyone not using AI on a consistent basis? So everyone's using AI. Is there anyone that's confused by all of the different options that are available by AI? And that's been my experience is the AI is moving so quickly, we can't quite keep up with what's happening.

[00:04:55.91] CHIP BACCIOCCO: Yeah, I agree.

[00:04:57.05] NICK LAMPARELLI: And we don't understand the underlying technology. So it's just like this magic box. But now I'm overwhelmed with so many magic boxes. Do I use Claude or ChatGPT 4 or Gemini? Which one do I use? Which one's better? I don't know. It depends on the circumstance. And how is the insurance industry supposed to adopt that when-- I don't know. Have you seen the results from some of the AIs?

[00:05:23.94] CHIP BACCIOCCO: Well, Leo shared some that were entertaining earlier. Yeah, I agree with you, Nick, 100%. It's daunting. I might have to pull my phone out. But this morning, I was like, there's nothing this stuff can't do. And then you start thinking about all the things you can do, all the things in our-- I'm a little biased for agents.

[00:05:45.58] I think, by the way, you guys could have done a whole day just on insurance company applications. Then you could have done another whole day just on insurance agency applications. You could have done a whole day on claims.

[00:05:58.69] It's unlimited. It's daunting.

[00:06:01.16] NICK LAMPARELLI: I would be very interested in finding out because I think there are regulatory issues with AI. I think a lot of insurers are going to be hands off for a little while. What are the low-- what's the low-hanging fruit? What are the easy applications, where we can quickly bring AI in?

[00:06:21.76] I heard crickets. Is that because I'm boring? What? Is it my-- what?

[00:06:27.92] REID HOLZWORTH: I think you're doing pretty good.

[00:06:29.80] NICK LAMPARELLI: But I'm mostly interested-- like, if we're going to adopt this, where are we going to adopt it? We're not going to adopt it right at the underwriting stage for the most part, because that's probably a pretty risky thing to do. But I have seen it in-- we were talking backstage. I have seen it in fraud detection. I think that's an easy one--

[00:06:50.55] REID HOLZWORTH: That's a great one.

[00:06:51.16] NICK LAMPARELLI: --where you don't have to worry about too much of the regulatory element of it. But I don't know. I don't know if you guys have any ideas where else we could see some early adoption.

[00:07:00.64] REID HOLZWORTH: Yeah, I mean, Leandro was talking about it. I mean, on the underwriting side, just repetitive tasks, research, things like that-- I really loved how he said, hey, treat AI like an intern. I thought that was really good. That was a key takeaway because it is. It's like, do the work, depend on it, but you got to check it. Right?

[00:07:20.53] CHIP BACCIOCCO: Yeah, he said check their work is exactly what he said.

[00:07:23.64] REID HOLZWORTH: --at least. But--

[00:07:24.93] CHIP BACCIOCCO: I agree with that. I do. I'll disagree just a little. Because what's the panel without-- the first thing that occurred to me on the underwriting side is-- and I totally agree, the regulatory side is a nightmare. But it's so easy. It's so tempting.

[00:07:41.49] Because Leo was also talking about the chess game. So that's one of the early AI moments I had was I watched a YouTube video-- and I encourage everybody to look it up-- of some researchers that taught some bots, virtual bots, not real running around, how to play hide and seek.

[00:07:59.77] So if you Google on YouTube-- don't Google it. Go to YouTube and go AI hide and seek. And the researchers were amazed. They thought maybe we'll be lucky in these little bots will teach themselves how-- there were two teams, little two guys try to hide. Two guys try to find the other ones.

[00:08:18.55] They thought, well, maybe it'll work. No, it didn't only work a little bit like he showed in that gaming example he had. They invented a new way to play hide and seek. They cheated. They found unbelievable insights.

[00:08:31.06] And so the first thing I thought about when I think of hide and seek a little bit-- maybe my gaming obsession-- is that insurance is a little bit like hide and seek. The losses are trying to hide from the underwriters. And the underwriters are trying to find the risks and uncover them, price them. And I guess insurance agents do the same thing.

[00:08:50.74] So an easy way-- my silly way of thinking, if I'm an insurance company, I have all the accounts I wrote pick any line of business you want-- property, auto, whatever. I know all the accounts I wrote two years ago. I know what premiums I charge. And now I know what losses I incurred.

[00:09:09.99] So a little bit like a lot of these games. You just run that through. And you have a little neural network that-- and you let it game itself. And the game is, OK, underwriting algorithm, write me a book of business that has a 30% loss ratio instead of 70%. And it'll do it.

[00:09:30.09] Now you're right, Nick, you might get in trouble. Because someone's going to be mad about why their price changed and a whole lot of stuff, but--

[00:09:38.46] NICK LAMPARELLI: Especially if you're in California.

[00:09:39.49] CHIP BACCIOCCO: Right, yeah, exactly. But it potentially shines a light on a lot of stuff that might be healthy for us to look at, anyway. So like I said, everywhere I look, I see potential. And maybe reckless potential, but potential.

[00:09:57.39] NICK LAMPARELLI: I'm curious in terms of who's out here. Are there start-ups out here? How many start-ups are out there? This is the first time?

[00:10:10.25] So I was taken from Jennifer Linton's presentation about the roller coaster, because that is absolutely true. And we had a heart-to-heart conversation last night about start-ups. And I think start-ups in the insurance ecosystem are especially difficult.

[00:10:29.10] And what I try to remind start-ups founders is, nobody wants to buy insurance. It is truly one of the more different industries. And nobody actually is lining up to buy our product. This is the House of Blues. I'm sure when some act books here, there's probably a line outside at the ticket booth. People are lining up for that. People line up for Teslas. They line up for iPhones. No one's lining up to buy insurance.

[00:10:56.11] And so I think a lot of start-up founders are ill prepared for the sales cycle that's required to actually sell into the insurance ecosystem. It is months or years. I tell every starting MGA that you have to bookend 2 and 2. It's going to cost you $2 million. And it's going to take about two years before you sell your first policy. Are you prepared for that? They usually say yes, but I have to keep talking to them and pull it out.

[00:11:26.89] But you guys have bumped into start-up founders and probably been start-up founders yourself. Right. I found the roller coaster to be completely appropriate. And you have to prepare for those troughs-- the anxiety, the depression. It absolutely happens to all of us.

[00:11:47.07] CHIP BACCIOCCO: You're the expert on this one.

[00:11:49.22] REID HOLZWORTH: I think it's really hard. I mean, it's really, really deep, what we do, first and foremost. It's not like you can build a simple application, sell it, and you're crushing it. You really can't. You have to know the space. You got to know the people.

[00:12:03.86] It's a big relationship business. Sales cycles are very, very long, very long, especially if you're selling into carriers. Agencies too. But it takes a while to build a reputation. And that's key in this industry and build the relationships and do it. It takes a lot of money, a lot of time, a lot of effort. But yeah, it is very, very tough.

[00:12:24.35] CHIP BACCIOCCO: And then there's all the-- even after you sell it, then you've got the change management and the culture adaptation--

[00:12:29.08] REID HOLZWORTH: 100%

[00:12:29.47] CHIP BACCIOCCO: --of the users.

[00:12:30.29] REID HOLZWORTH: Adoption.

[00:12:30.87] CHIP BACCIOCCO: So the adoption curve, you might have the leadership of a business. It's like, oh, this is a great idea. Well, that's step one.

[00:12:40.98] REID HOLZWORTH: Yeah, totally, 100%.

[00:12:42.93] CHIP BACCIOCCO: And then you got-- and you know this, again, better than anyhow. Agency management systems like, no, no, no, here's how we enter the data about this account, here's how we process an endorsement. And you're like, yeah, I know that's how you did it, but there's another way. And--

[00:12:57.21] REID HOLZWORTH: Every carrier, every agency is different.

[00:12:58.86] CHIP BACCIOCCO: They're all. They absolutely are. They absolutely are. Should we ask for questions.

[00:13:05.24] REID HOLZWORTH: I think it's time.

[00:13:06.41] CHIP BACCIOCCO: Is it time?

[00:13:07.35] REID HOLZWORTH: I think so.

[00:13:08.09] AUDIENCE: I'm just curious. The start-up, when you have an idea and you're starting to move forward, how much time is put into the validation? And you're validating at many different steps along the way.

[00:13:22.14] REID HOLZWORTH: You have to test it. I mean, you have to-- if you're bringing a product to market within this industry, you got to get the customers first. And it's the chicken and the egg. And maybe they're not customers to begin with, but you have to be real-time testing the products in a real environment.

[00:13:36.87] To come out with something before agents have truly tested it and giving you feedback, you got a long road ahead of you. And I think that a lot of people come into this industry not coming from the industry that say, I know how to fix it better, not having that industry experience nor the connections. And it doesn't really end well. It hasn't ended well historically.

[00:13:58.14] It is a network. All of us-- I think that this industry is very open to new technology and new things. I don't think it's hard to get in front of brokers or carriers in that way, especially on platforms like this. You can meet people. And you can show. And you can start to talk. And then you can start to incent them to try to use it and test it out and give you feedback 100%.

[00:14:19.15] And then you're going to get beat up, kicked down. You're going to cry, all the things. And then you got to turn around and make it happen. So that's 100%.

[00:14:28.73] CHIP BACCIOCCO: Yeah, I would echo that. I've been very fortunate to know some great mentors. And Pat Gallagher used to say when I was much younger and naive-- and he would say, everything around here-- he meant Gallagher-- starts with a sale. And what he was trying to chide me against, because I was in the ideas' part of the business was, you guys in your ideas, go sell it first, show me a check, show me an order, show me a contract, and then we'll build something.

[00:15:03.59] And if I had to guess, and I'm just speculating, many companies, many start-ups start off with a bespoke couple of contracts like Jen was talking about. And I think that's because you end up in a conversation with a potential customer. And the customer says, well, here's what I want.

[00:15:20.33] REID HOLZWORTH: Yeah, that's a good point.

[00:15:21.53] CHIP BACCIOCCO: And so you end up saying, all right, well, for this one, we're just going to build what the hell they want. And you have this imaginary perfect product, this platform. But in the short run, you need some customers. You're going to learn why you do it. And it might not be exactly the product or services that you imagined. But you're heading in the right direction. You're getting paid. You're building some reputation.

[00:15:48.90] I think you have to-- you can't just build in a back room someplace, which we all seen people do that. They build this-- they get coders. And they start building and building. They don't actually have any customers yet. And then they never get any. And that's the danger.

[00:16:03.92] REID HOLZWORTH: We built TechCanary for our agency originally. And we thought we knew everything there is to know about insurance. And as soon as we brought that product to market and got away from the agency space, we got our ass kicked because we did not know how everybody did it. And so it was a few years after that before the product came to fruition, to your point.

[00:16:26.78] NICK LAMPARELLI: I have to-- I'd like to break up the world, bifurcate it into ideas that are technologically based and ideas that are insurance based because they're slightly different. If you're bringing something new to insurance like the MGA that I co-founded seven years ago, was to basically work on the flood side of stuff, it was incredibly difficult to validate. It had never been done before, especially not the way we were doing it.

[00:17:00.59] So it was-- we had to work on the trust angle. We had to work on the angle of like, we're taking this conservative approach. Here's our plan B. Here's how we're going to make sure the losses don't get too big on this side. We're going to-- we're not going to write in Florida to start. Here are all the things we're going to do. Please, here's our tin cup. Please, put a dollar bill in it so we can get started. Whereas I think if you're building technology, you do get a tighter feedback loop.

[00:17:29.72] So you should be able to figure out an MVP, a Minimum Viable Product, something you can get into the hands of someone. Get feedback, figure out how to monetize that over time, and continuously validate it. You got your butt kicked sooner than we would have gotten our butt kicked because we had to build the whole thing just to go live.

[00:17:53.45] And then we had to build a portfolio before we knew this was going to work or this is not going to work. It ended up working, and it ended up working really well. But the validation around it, that's the biggest problem with insurance-based start-ups is it's probably literally takes years to figure out, this is a good idea or this is not a good idea or it's not going to work. So it's the split between technology and insurance.

[00:18:21.46] CHIP BACCIOCCO: Down here first, then we'll go to Chris.

[00:18:25.38] AUDIENCE: Chris has the mic, so we're starting with Chris.

[00:18:27.66] CHIP BACCIOCCO: Yes, Raghav, yes.

[00:18:29.29] REID HOLZWORTH: You are very lazy.

[00:18:29.52] CHIP BACCIOCCO: Yes, sir. Yes, sir.

[00:18:31.09] AUDIENCE: I won't ask five, I promise. It's kind of a comment and reactions. I think this AI, I'm fully immersed in it like a lot of people here. I think it creates promise in a couple of ways that addresses this question about how new entrants into the space can play.

[00:18:47.17] Because the blessing and the curse of independence is there's 40,000 independent agencies in the industry, a couple thousand carriers. How many technology providers? Do the math on how many combinations of running an independent agency there are. And that's the power of it.

[00:19:02.28] So no solution has been really able to hit real scale there. And I think AI, the customization of that will allow more organizations to find their way in to allow the unique nature of how independent insurance agencies operate to be impacted by it. But the other side of this thing is the question around it, and I saw this on the carrier side, was understanding the economics of this industry is really critical at this point in time and understanding what happens to $1 a premium.

[00:19:31.91] Because we're not manufacturing widgets. We're bearing risk. And every single person and every single client is different. And when a carrier has to take $0.60 of $1 a premium and put it on a shelf to pay claims, that $0.40 that left has to do a lot.

[00:19:47.99] And understanding what happens with that math is really important. And the companies who have come in and made a difference and penetrated the industry and understand that how the economics work and where their product can move the numerator and denominator have really the ones that have had success. And I think AI can allow us to find more ways to allow creativity in that space.

[00:20:12.19] NICK LAMPARELLI: Good comment.

[00:20:13.13] CHIP BACCIOCCO: Yeah, agree.

[00:20:20.89] AUDIENCE: So I'll continue on that. So you're saying it's difficult to latch yourself and to start. But AI is like a wave. The way I work as a software engineer, 60% of my code is now written by GitHub Copilot. I use ChatGPT all the time. My workflow has changed completely. It's impossible that insurance will not change completely in a few years.

[00:20:45.91] The question is, will this make it easier for start-ups to start other niches where it's easier? Or does it have to be almost like a PayPal style approach, where you do something completely different and almost go laterally around regulation to do something truly innovative and create your own market? How do you see this playing out with generative AI?

[00:21:13.50] REID HOLZWORTH: For me. I don't know. I think it's too early to tell, personally.

[00:21:16.90] CHIP BACCIOCCO: Yeah, I think it's too early. When you were talking about start-ups, did you mean technology companies or insurance businesses?

[00:21:25.50] AUDIENCE: Probably technology companies more than insurance businesses providing tooling to insurance businesses.

[00:21:32.10] CHIP BACCIOCCO: Yeah, that's--

[00:21:33.43] NICK LAMPARELLI: I think AI is going to help technology companies, as you said, more-- you said more than 50% of your code is being written by AI? So right there, it's already making a difference. So yes, I think there are start-up opportunities.

[00:21:48.92] I think for insurance-based businesses, you have to pick and choose the use cases. I think most carriers are going to be extremely careful where they implement the AI because they just do not want any of the negative publicity that could occur or the wrath of the regulator on top of them.

[00:22:10.46] And the AI has been so incredible. But when I think of things like climate-related extreme weather, I think the AI still has a long way to go. I still think we're going to be in a highly actuarial model of building insurance programs, risk-based programs, where that $0.40 of margin has to still do a lot of lifting.

[00:22:40.58] I've seen firms that are talking about their AIs just continuously dropping their loss ratio. And I salute them. Like, if you can do that, great. But eventually, the regulator's going to tell you to give some of that loss ratio back to the policyholder.

[00:22:57.48] And I feel this is like Alice in Wonderland, the Red Queen. We're going to be doing a lot of sprinting on this AI just to keep up with everybody else. That's probably my biggest fear about the AI is that we're going to have to have this massive investment, these resources. We're going to change how we do all of these operations. And the loss ratio is going to be exactly the same on the other end because all of the benefits just will go to policyholders. What's the correct balance?

[00:23:30.77] And I think we've seen it in the past couple of decades with the digitization of carriers. Have carriers really improved efficiencies with all of this digitization? I think some have. I think a lot of them have worse loss ratios than they had when we first started digitizing. So they're sprinting and sprinting and sprinting. And they're not making any progress. I'm hoping technology or an AI, we finally get some sort of scalability in insurance, we make more progress, but I have not seen it yet.

[00:24:00.69] CHIP BACCIOCCO: Well, I have to agree with you on that. So there's a lot of cynics-- I shouldn't say a lot. There are some cynics in our industry that have-- and you know what? They're not even-- let's call them pragmatists.

[00:24:14.90] And they have seen us as an industry spend crazy money on technology for the last 20 years. And there are lots of C-level executives that look around and observe what you just said, Nick, which is a combined ratio is about the same as it was. Profitability for my agency isn't too different. We still have about this many people here.

[00:24:40.05] So in other words, that's a really good point. We're all very excited about the potential, but not everyone's going to progress at the same speed. And there's going to be a lot of hiking down the wrong trail. There's going to be a lot of expensive projects that then get reversed, which have happened with every technology initiative in our business for as long as I've been here.

[00:25:07.03] And maybe that's just the nature of things. You've got to explore. And you got to walk down paths that you thought was the right idea. And then you get there. And well, maybe that was wrong. We took a wrong turn at Albuquerque, and we got to go back and try that again.

[00:25:24.46] So yeah, Chris and I have had a lot of conversations about insurance company-- well, in my own board. So my board-- just as a total sidetrack, my board, it has eight insurance companies on it. So I'm in an interesting position as a company that mostly sells to insurance agents.

[00:25:44.06] And one of their biggest, it's not even a private concern, is when the heck are we going to see an improvement on some part of my combined ratio for all this investment that we make? When's that coming? Because I've been waiting. And of course, then I say, next year, don't worry, it's coming.

[00:26:05.96] [LAUGHTER]

[00:26:07.88] NICK LAMPARELLI: But that could go into the nature of the business. So insurance has been around for hundreds of years. And this could be-- I think of the telematics and I think of smart devices and IoT. So our society is fully marinating in this technological revolution insurance hasn't quite got there. And I almost envision there has to be a change to the business model for that to happen.

[00:26:42.47] We're making these huge investments in technology. When is it going to affect the loss ratio? Well, maybe there's a whole bunch of other things in the business model that have to change, including distribution, how you market, what the policy looks like. The policies are virtually identical to what they were decades ago. And the customer is still as confused.

[00:27:04.47] So there's goes back to the very beginning of what we talked about. This is a product that no one really wants to buy. How do you sell something to someone that doesn't really want to buy it, but is in a lot of cases being forced to buy it? How do you make that-- how do you create some kind of emotional connection with that?

[00:27:26.91] That's what every major product brand is trying to do is create an emotional connection. Ours is supposed to be at claim time. But single digits of policyholders have claims. So that doesn't scale well. What are the other opportunities?

[00:27:43.47] And I see telematics and I see smart devices as-- imagine a smart device. I have a Ting electrical smart device. So it plugs into an outlet. And it's measuring the voltage in the currency. And I get a report every single day. And if the power goes out, I get a text message. And all I keep thinking about is if my insurance company had a logo on that Ting device, I would be thinking my insurance company is protecting me.

[00:28:11.99] And it's a way to generate an emotional response for a product. Most policyholders are thinking, what am I getting for this money that I'm paying? When am I getting my new roof? When's that hailstorm coming? When do I get my new roof?

[00:28:28.23] CHIP BACCIOCCO: --the drone. Yes, I have-- the drone is on the way.

[00:28:30.60] NICK LAMPARELLI: I paid all of this money, and my claim happened, and you guys weren't there for me. That's what everybody is worried about. And I look at these new opportunities with technology to just touch a policyholder every day and just say, you are safe. There was no water detected. No one's breaking into your home. There were no fires. Your electrical system is fine and dandy. Now you can sleep at night. You get this feeling that I'm getting my money's worth, my insurance policy is doing something for me, and I don't have to worry about it.

[00:29:02.49] REID HOLZWORTH: That's great point.

[00:29:03.25] CHIP BACCIOCCO: Yeah, well, I was just going to say, the trick is, and I think Jen Linton mentioned it, you can love them too much at the wrong time. At some point, they say, shut up, we've heard enough, thank you, I bought this so I don't have to think about my house burning down, don't tell me every day my house didn't burn down.

[00:29:19.20] But if we could change a little bit, I wanted to go-- I wrote a note earlier when Matt was up here, and he was talking about all that data, and I suddenly got very worried about who all knew my driving habits and so forth. So I'm literally ask-- this is my question. If Matt's still here, is that subpoenable? If I--

[00:29:40.31] REID HOLZWORTH: Good question, yeah.

[00:29:41.51] CHIP BACCIOCCO: Is that data out there that somebody could use it? There he is. Matt, is that-- can I be pulled into court and you're going to share my data about where I drove and how fast? No, he said no. There we go.

[00:29:54.79] SPEAKER: We do have a question back here. We actually have two questions back here. So we're going to start with Mark and go to Jason.

[00:30:00.26] AUDIENCE: Hey, guys. Mark Stender from Linqura AI. I want to go back to the use case you talked about earlier, Chip, because I think you were spot on for where AI is going. So AI that learns and acts is really where this is going.

[00:30:17.27] So your analogy of turning a dial, so to speak, to get the outcome you want is generating AI to get a very specific outcome. So spot on with that. I was wondering if you had an analogous or signature use case on the distribution side.

[00:30:34.78] CHIP BACCIOCCO: I do. [LAUGHS] Thanks for the question. So the distribution side has a whole different set of challenges. We're talking about insurance agents. And I love agents. I'm a licensed agent my whole career. So I am one. So I can criticize us.

[00:30:56.30] We don't know what our cost of acquisition is. We don't know which clients we serve well. We don't know how to stop wasting time on prospects that we really wouldn't serve well. There's going to be so much that AI think brings to the insurance agency operation.

[00:31:14.97] It's going to be very fragmented, to Chris's point before. Because you've got 40,000, 35,000 different agencies. And they're all going to adopt this stuff at different rates. But with the help of companies like Applied, there will be some sort of tool kits.

[00:31:29.73] I'll give you one example. So I would-- we always talk about lead scoring a lot. Certainly in my business. And there's several ideas for how could AI clearly help you see through your own data. You think you could score your own leads or somebody else does with some pretty simple logical algorithms like, hey, that guy that has a yacht, that might be a nice prospect for you. That's a high scoring lead.

[00:31:53.80] That's all true. But if you really applied the data, you might find out that chasing people that own yachts is not really going to be successful for your agency because either the markets you have, the products you have, the people you have, or how people with yachts behave.

[00:32:09.46] So I think that will be one of the earliest uses is actually analyzing what are you actually good at, where do your best prospects come from, how do you avoid wasting time just quoting and marketing. We all think we know which markets have the right products at the best prices and will be competitive at quoting certain things.

[00:32:34.88] I'll bet anything if you let AI game itself and try to do better than you do, it will reveal things to you that you-- just like that Go player that was like, Why the heck did he make that move? But then won the game. So we have to be open to the fact that AI is going to uncover kind of like Michelangelo used to uncover statues-- look at a rock. And he'd say, no, there's Apollos in there. I'm just going to set him free. But you have to be open to finding Apollo the way Apollo is inside of that rock.

[00:33:07.55] And I think we're going to be amazed at the suggestions that our own data on the agency side, everything, sales process-- we think our best producers are good producers for some specific reason. We might be very surprised if we could look at the data of why does this producer outperform that producer and then adopt those obvious results into the rest of our sales process for the other producers. And it's going to be exciting to find out.

[00:33:40.34] NICK LAMPARELLI: I think that's a marvelous use of AI in distribution is to-- listen. You have expert producers, producers who are elite. They're in the top 1%. And they've been doing it for so long, they probably don't even know why or how they're doing it. It's just instinctual at some particular point.

[00:34:05.54] They can look at body language. They can look-- they're just very comfortable when someone has an objection. They don't even know how they do it or where that came from.

[00:34:15.62] And it's critical to the younger producers to figure out how to get what I call tacit knowledge. It's institutional knowledge that they have that exists not just on the production side, but on an experienced underwriter or an experienced claims professional. They're almost doing it instinctually. And I think AI has a really-- it's the only way to shorten the arc of training for some of the younger professionals coming in and make them elite.

[00:34:48.43] The middle tier person that's fairly new in their career, they don't want to wait 10 years for them to become some head of some department. They want to move quicker. And the ability for AI to look at experienced skilled staff; figure out what is making them experience and skilled like the subconscious stuff, the words that are not said, the words that are said; and then translate that to a younger audience or get it into some sort of simulation so that the younger professionals can practice how to overcome an objection, practice to know when reading a submission whether this has got a low probability of making it through, I should just decline it up front-- even for a broker, look at a submission and say, I hate to turn down money, but this is going to be a hard placement. It's going to be really inexpensive account to hold on to.

[00:35:47.96] We never had those tools to be able to do it before. And the folks that develop that innate expertise had such a massive advantage over everybody else. AI could level the playing field.

[00:36:00.77] CHIP BACCIOCCO: It totally could. I'll just give one more example, if I can. So one of the things that we do for our clients is we record their phone calls, their inbound phone calls from prospects. We record thousands every single month. We're, obviously, want to let AI with appropriate removal of personal information listen to them and so forth.

[00:36:23.36] But I've done it anecdotally for years. I've listened to probably more agency phone calls maybe than anybody in our industry. I don't know. I made an observation a while ago, and this is the free tip that I have given out for years.

[00:36:35.69] The number one thing I think a producer/CSR, can do to dramatically improve their chance of winning a client is when they're in the first conversation, it's very normal for the client to say, I'm looking for, I need XYZ insurance. And the first thing when I hear these magic words on the phone, I see a doubling of ultimate success. And these are the words, I can help. It's that simple. I can help.

[00:37:04.77] Now I really don't want that to be my anecdotal story anymore. I want data to back that up. I want a neural network deep learning tool to have scoured thousands and thousands of phone calls and to confirm for me that my suspicion is correct, that those are magical words to use to a human being.

[00:37:24.91] By the way, the reason I think they work is because people are stressed out when they're buying this product that they hate having to buy. And they don't want to shop all day. And they just want someone to reassure them that they're done with the hard part, that it will go well from here. I just need to give you some information. You're going to give me some quotes. I'm going to pay more than I want, and I'll be on with my life.

[00:37:45.00] But those little things and having observed that-- but I'm looking forward to there's so much AI by being objective, by combing through just endless piles of data that we don't have time to and we wouldn't know what we were looking at if we did. And it's going to bring magical efficiencies. And then we'll have the next problem to deal with, and we'll deal with that. But yeah, that's my story.

[00:38:11.48] - We have one final question.

[00:38:13.40] AUDIENCE: Chip, what I want to say is you have no idea how really how right you are. As an agency owner, the use case for AI today is in productivity. And what I like to say is that's doing more with the same inefficiency, which is doing less with the same. And I think that that is where we try to look at it.

[00:38:31.36] You said earlier that the CEO looks at it and says, well, my combined loss ratio didn't change much. I'm pretty sure they're saying that. But if they are, they're missing the point here of what we're doing.

[00:38:42.04] Just because the weather channel.com can predict the weather or just because when you're typing on your cell phone, it predicts the next word, that doesn't mean that we can predict the actions of humans. And I think maybe in the future it can. I'm not saying it never can.

[00:38:56.59] But I appreciate what you said, Chip, there because it was good information you gave. And it's about the efficiency and the productivity, which you're more operational. And when it even comes to the data, especially with producers, when we're trying to go through producers and we're trying to mine through that data, that's not predictability.

[00:39:13.40] That's just, how can I be more efficient in the people that I'm trying to call to locate my client? And if there's anything that any companies could provide, it would be being able to-- letting agents harness those tools so that we can use those and we can help you sell your product more. So I appreciate you were right on there, Chip. Thank you.

[00:39:30.64] CHIP BACCIOCCO: Thank you, Jason. And I'll shut up in a minute. But I want to say, here's the predictive part. The predictive part is that you look at the data to predict what a new sales person or somebody you don't know, what could they do to improve all of their results. And then you let the people do it. And then you get the productivity from everybody.

[00:39:53.80] Like Nick said, there's things they're doing, the successful ones are doing, they don't even know they're doing it. They're sending a text message, but they don't know. But they tend to send them in the morning. And for some reason, that works better than sending text messages in the afternoon. I don't know. Whatever it is, there's little behavioral elements that we can pick up that will show us a pattern that predict a higher rate of success, a higher rate of productivity.

[00:40:17.72] NICK LAMPARELLI: I say we can predict a lot of stuff about humans. I like the Jeff Bezos' way of like, hey, we should be investing in focusing on the things we know humans are going to still want and need 10, 15, 20 years from now. So for Bezos and Amazon, it was 10 years from now, 50 years from now, 500 years from now, humans are still going to want a discount. They're going to want a deal. They're going to want lower prices. What scenario would exist other than luxury where people would want higher prices?

[00:40:49.90] I think in insurance, Jason, to your point, in some of the things that were said here, what AI can't do, AI cannot make me sleep better at night. A human can. And you're going to be OK. Everything's going to be OK. The hard part is done.

[00:41:08.95] And that goes for broker to prospect. I think it also goes from underwriter to broker. We're going to be OK here. It's to give that reassurance claims professional. There's still 10 years from now, 50 years from now. I don't think lemonade insurance is correct that a bot is going to tell me that I'm OK. And I'm going to just say, OK, I'm OK, I'm going to be able to sleep at night. There is a human element. Something has to be sold.

[00:41:44.11] It's AI plus human is the right combination. It's itself, it's just too nebulous. Folks will try to make it replace humans. But I think in our particular industry, it still requires humans to say, you're going to be OK. We're going to make things-- we're going to make things all right for you.

[00:42:06.16] CHIP BACCIOCCO: Yeah, we sell a social contract. It has social elements. You want the reassurance of Mary or whoever, your insurance agent to look you in the eye, even if it's over Zoom, and say, you're covered, and I'm accountable to make sure that that happens. And no clicking of any buttons is going to give you the same feeling as looking Mary in the eye and her telling you it's going to be OK.

[00:42:30.08] So I agree a million percent. It's a social contract. It has social implications.

[00:42:36.12] REID HOLZWORTH: Closing thoughts.

[00:42:38.90] CHIP BACCIOCCO: Haven't I talked enough?

[00:42:40.60] REID HOLZWORTH: Quick.

[00:42:42.47] CHIP BACCIOCCO: It's awesome to be here. We work in this amazing industry. And I enjoy every minute of it. So thanks for sharing your day with us.

[00:42:51.01] NICK LAMPARELLI: Technology is marvelous. Humans are, too.

[00:42:55.87] REID HOLZWORTH: That's awesome. Thank you, everybody.

[00:42:58.51] [MUSIC PLAYING]