Claims Intelligence #4: Navigating AI Adoption in Insurance with Joey Daryanani
Watch
Listen
Follow
Transcript
Understanding Innovation and AI
Sean Merat
Thank you, Joey, for joining us on this session. I really appreciate your time. As a way of intro, I mean, Joey brings a very rare and deeply practical perspective to today's conversation. He spent over four decades with CSAA, including more than 35 years of those just across the claims organization, holding numerous leadership roles helping build and scale high-performing operational teams.
In his current role, Joey oversees the support services for CSAA's operational team and helps lead the company's work around legal-system abuse, which is one of the most important and complex issues facing insurance today.
And then earlier in his career, he led Claims Central Services, including areas like subrogation, salvage, total loss, Medicare, claims training, basically everything. He's also helped launch and lead Specialized Services, a business that grew out of an employee-innovation challenge during a pandemic and which expanded into claims handling, training, and call-center support. So, Joey, great to have you and thank you so much for joining us.
Joey Daryanani
Sean, thanks for having me. Always a pleasure.
Sean Merat
So, I mean, you know, with over four decades, 40 years with CSAA, and much of that, you know, across the various functions and operations, I'm curious to know, when you hear AI innovation and insurance, what's feeling genuine to you? Like, what is fresh and new? And what feels like a continuation problem that insurers have been always trying to solve?
Joey Daryanani
Yeah, I think—great question. You know, if you don't mind, I want to start off by splitting those two words. So let's just take innovation to start. Innovation is nothing new to us; in recent history, that's what we've done. So I want to dissect innovation first. When innovation was the big buzz that everybody was talking about, everybody thought this had to be some transformational idea around innovation. As we look back, it turns out that while we had some transformational ideas as an organization and as an industry, we really celebrated the small incremental wins, right? So I think that's important to focus on when we talk innovation.
Now, as I think about the word innovation, I think of it as being the "what." What are we trying to do? What are we, what problems are we trying to solve for? Now you've sort of introduced the word AI. Right? And AI to me is the "how." It's not the "what." So I think we put that in perspective, is making life simpler by trying to figure out what tools we have. I think it's important to sort of stress that. To me, emerging technology is a tool. It is not the "what."
So, if to boil down the answer to the question you've provided is I don't think anything has changed as far as innovation in the insurance industry. It is ripe—we continue to make some good progress in that space. With the introduction of emerging technology and the "how," I think it'll even make it better for our people. I hope I answered that question.
Sean Merat
Yeah, I think that makes a ton of sense. I had a conversation with, you're familiar with Garry Kasparov, the chess player, and he, the way he would talk about this is that AI is a tsunami and it's going to, you know, it's, but it's us humans that basically have to direct it. So very similar, I think that analogy kind of feels similar to what you're saying, is we need to give it direction. It's a tool, and so we need to know what we're going to solve for and use it as a tool.
How do you come to deciding which problems would be best solved with such technologies?
Joey Daryanani
That's the interesting question. So when we had a conversation before this call, I think I made it very clear that I love my toys and I continue to love my toys, right? I'm the guy that when I'm talking to potential vendor partners, I don't think there's one in the recent memory I have that hasn't talked about their ability to roll out AI. It almost is, I don't even know what it means anymore. Right?
This is where I give credit to our CEO, Mike Zukerman, who has made it very clear that the platform is: do less things, but do them correct and do it better. And that's where the answer, I think, to your question is for people like us getting disciplined on what we're trying to solve for, actually based on business outcomes and return on investment, I think is super important, and not take in, not ingest more than you can chew, because you do that, most often I've seen that sort of fall off the rails and things not get implemented for a few things, a few reasons. One is price. Second is the organization only has so much appetite to execute.
Sean Merat
Right.
Joey Daryanani
And with IT resources the way they are internally, we have to be careful and really pick & choose and get very clear on what is the problem we're wanting to solve.
Practical Applications of AI in Claims Handling
Sean Merat
Yes. I think, I agree with you. I think, you know, we're, we have a front-door seat with many different carriers and that is a refreshing take because, it's, you know, a lot of carriers are doing AI for the sake of doing AI, whatever that even means. And so it gets very expensive very quick, and I think it's prudent to think about the efficacy and the applicability.
So, across like, if you think about the positions that you oversee or have overseen in the past, like, whether in, whether it's claims handling, shared-services training, the fraud detection, litigation, can you give us, like, maybe a few practical examples of some of these, like, how you would think about the problem statement, how it can be solved with some of these newer frontier—or just AI technologies—and how you would track basically the success?
Joey Daryanani
Yeah, you know, this might sound like oversimplifying, right? So let me just start with the use case around trying to summarize and evaluate demand packages that we get in the casualty world. I go back 40 years and spread that demand across my dining room table and have to look at every document, every word to make sure that I've got it right, write that file up, present it for authority—that whole process might literally take a day.
The first use case that we're focused on is summarizing that so that our employees are focused on the relationship with either the lawyer or the claimant. It's focused on understanding the type of case we have versus trying to summarize it themselves. And it is focused on ultimate resolution, significantly reducing cycle time. In everything that I've said there, there is an ROI that is quantifiable, that is actually hard dollars.
Let me give you a "soft dollar" example: first notice of loss. We've trained our employees in the past to focus on quality assurance. While you're [on] a first-notice-of-loss call, your notes have to be perfect. Your summarization has to be perfect. Your next best action has to be correct. All of those things and because of that, the intent was great, right? But the outcome focused more on how can this pass QA?
Sean Merat
Yes.
Joey Daryanani
What gets left behind more often than not is we have between 12 and 15 minutes to build a relationship with our member on the phone reporting a claim.
Sean Merat
Right.
Joey Daryanani
And typically, and I love this example, when you have a member reporting a glass claim, and we ask, because we're following a script, what the weather was, knowing it just doesn't matter, but we are so focused on getting it right, right? We lose the relationship portion, which now we have the ability to go after.
So summarizing, note-taking, translation, next-best action, all of those things in a first-notice-of-loss example really helps us build the relationship with a member and really start that investigation on the right foot. And the third is, you know, we're a 120+ year-old company. We have a lot of data. And from a fraud-detection perspective, right, we don't know what we have in the database, more so what's in the database of the entire insurance industry. Let's just focus on us first. So being able to access who we're dealing with, what the files in the past have looked like, what the settlement has been, what the negotiation was, is especially important for our people.
So I've given you three examples that, to us, make it very clear on why we picked the first three. Now, there are about, you know, there's a bunch that go behind that, but pause on the other stuff. Let's get good at these first three before we move on to item number four.
Enhancing Claimant Experience Through Technology
Sean Merat
I guess what I'm hearing is, your priority is, you're putting the claimant experience up front, whether it's added relationship, better care, better intervention. I mean, all of these things if you have the proper systems that can automate the mundane tasks, you're able to, all of a sudden, it's like removing 70% from your adjusters' plates, and then they can actually do what humans do best, which is relationship, empathy, and just listening.
Joey Daryanani
So Sean, I think you hit it right there, right? We are completely—let me back up. We are the business of people helping people solve real problems. Okay? So, yes, being able to do these three things frees our people up to build the relationship with claimants and members.
And let's just take legal-system reviews. How we got to where we are today, in my opinion, is if you just listen to the ads on TV, if you look at the billboards, it's all about trusting your insurance adjuster. That's the whole premise of people that take advantage of the system. And if we can do a better job building trust with a very little time at the beginning of that claim, you know, either while we're handling the property damage while the injury claim is pending, or taking care of them and proving the ads and the commercials wrong, to me that's your advantage, and it's all about trust.
Sean Merat
Yeah, I fully agree with that. And that's lost on, you know, we work with many of the tier-1 carriers and oftentimes we find ourselves being the, I guess, the seat in the house at the table on behalf of the claimants, and internally we actually in our company, even though we sell to the insurers and they're our partners, we like to think that we're building technologies for the end claimants, because if we can actually improve their experience, we've already, we're doing right by our partners. But that's a conversation that I find that often gets overlooked in a lot of boardrooms where they, you know, because the easy wins are just too accessible around. And you can win a few battles, but you will end up losing the war.
Joey Daryanani
Yeah, so let's get real at this point, right? What are the things that can easily be handled within the insurance space, right? How can we influence the bottom line the quickest? And the answer to that is expenses. We have control over that. It shows up really quickly, right? Whether it's an increase in expense or a reduction in expense and this is why the world, not just insurance, has focused on the expense side. That is not what we're here doing at CSAA. Over here, it's really: how do we improve the workflow by either summarizing documents, faster cycle times, fraud detection, reducing repetitive tasks, all on behalf of the member/claimant to mean building that trust. So that's how I make the connection. Instead of going where the mind goes really quickly, which is how do we reduce expenses? To me, this is not or should not be an expense.
Choosing the Right Technology Partners
Sean Merat
That makes a ton of sense.
Now, here's the difficult question. I'm actually really curious because I always sit on this side, never sit on your side. And we had the benefit of, you know, we've been in AI shop since before AI was cool, and we've racked up most of the tier-1 carriers. And back then, you know, AI had a whole different meaning. Now, every day, every conference we go to, there's, everyone and their mothers are having an AI company basically launched. From where you're sitting, there are so many different providers offering solutions to the very problems that you've listed, that you want to focus on, and, I guess, you know, they all look impressive, right? Of course, as a vendor or a startup, most of your job, most of your focus goes into creating really nice demos, right? Because that's what sells.
How do you, you know, distinguish between all these different providers to try to figure out which one is the right one for you? As you pointed out, your IT resources are constrained. What is the winning operational process here?
Joey Daryanani
Wow, loaded question, right? So, as a 100+ year company, one of my learnings are, you absolutely need to get disciplined around what's your strategy instead of bolting this on and bolting that on and waking up one day, all of a sudden and you've got a bunch of stuff that either isn't supported anymore because these companies are no longer around, or you've just not had a good enough strategy around what you're trying to do.
Let me explain that a little more. So there are, I'm making this number up, 200 things I want to do in that space. If I did all of them at once or I did all of them really quickly without a strategy around data and without a strategy overall around how and what the interface would be for our employees, I don't think it'll succeed at all, and we'll be right back where we were when we started the journey around replacing all our core systems. Right?
So the way I think of it is, in the example that I've given you of 200 different things, I run training—I can't go to our employees saying, "Hey, I'm going [to] teach you 200 different things."
Sean Merat
Right.
Joey Daryanani
There's just not that capacity in the human brain to ingest that. The way I think of it is, if we're really focused on it, is, first, let's decide strategically what is the interface of the employee. Is it the claim system? Is it the policy system? Is it the HR system? Is it, you know, whichever, is it [the] finance system? And if you—I've just thrown out five different systems—the way we're thinking of it is: one interface with those different systems sitting behind it, where you interact as an employee with that one, whether you want to log in volunteer hours, or you want, where in the policy does it say this about additional transportation, right? Or where is the next company hauling in? It's one interface. And then the way you train to it is the way Apple sends in their upgraded sort of software, 1.1, 1.2, 17.8, where you don't see anybody complaining about, "I got to download 17.8." Why? Because it works. So as we train, what I've emphasized with the training team is, let's use that approach. It isn't, I'm going [to] train you on 200 different things. I'm going to do 1.0, 1.1, until we get to wherever the sort of current technology is and whatever year that is moving forward.
But to be able to do that, you got to be very disciplined on the platform and the interface of the employee. I'm interacting with my iPhone. I'm not interacting with the apps.
Sean Merat
Yeah. That makes sense.
So just, I guess, firstly, on the, on trying to pinpoint the correct technology to use, I hear you on the fact that it should, you know, integrate with your system of record or whatever that is, I guess. So I'm assuming, you, are you thinking about this from like it's an in-house built platform or a core system, like a Guidewire, being your one interface? Like, how are you thinking about that one interface?
Joey Daryanani
Yeah. As many, well, there's a question around build versus buy, and that's a whole [other] session for us.
We're getting really good at trying to decide what's our core competency and what do we build versus what do we buy, right? So there are vendor partners out there that we have utilized that are going to be much better than us ever building it because that's their line of business, right? And we get to take advantage of the improvement in whatever output they have, right?
So, again, as I walk the floor at the conferences, I know which players have the reputation and I know which ones are new. It doesn't mean I don't learn from the new ones, but I have to also be very guarded around, "Hey, buyer beware" around historical ability to deliver and execute and all of that. The other thing you hear frequently is, "Hey, we'll get that done in four weeks." We all know that doesn't matter, right? So, strategy is important and discipline is important. I can't stress those two things enough.
Sean Merat
Yeah, I think, [that] makes sense.
So would it be fair to say that your preference is to work with reputable providers because it gives you, like, it's too noisy outside of those? Like, the newcomers, while they may have very attractive offerings, it's hard to separate fact from fiction and the most prudent path would be to go after what's being provided by the more reputable and wait until some of the newcomers maybe build a reputation before…because otherwise it may be, it's too much of a time-suck. Is that a correct assessment of what you're saying?
Joey Daryanani
I think that's a fair statement. I think where I'm more sort of leading towards is, I'll give you a live example. This week, I looked at a really good total-loss valuation for auto capability from a startup. And, while it was appealing, my first question is, "Have you engaged these large estimating organizations, and why aren't you partnering with them?" So that when we use their solution, it's more of a global solution that gets protected, right? To me, the answer to that is critical. Because if I was them, instead of peddling their stuff to each carrier, if they did it to the top four, they've got the entire market covered.
Sean Merat
Yes.
Joey Daryanani
And it's a win-win, I think, for the carrier with peace of mind, saying, "I'm not dealing with a startup that may or may not be there tomorrow, I'm dealing with the people that I'm already doing business with my estimating platform that has all these solutions built in."
Fostering Adoption of New Technologies
Sean Merat
Right. Yeah, I think [that] makes a ton of sense.
Now if once you've identified the right…so you already talked about identifying the end problem, what is it we're trying to solve, you've identified, kind of, the technology and how you want to scale this through that one platform that you mentioned. Now it comes to the people that are going to be used, and you spoke about them, you know, having an easier upskill experience by having to be on a singular platform.
But what is, how in your experience have you been able to get adoption through these folks? Because, as we know, I mean, depending on the line, but, you know, for claims adjusting, you know, adjusters are, you know, often they also have, like, many decades of experience. They don't like change a lot of times, you might feel some resistance with technology. Have you experienced that? I'm sure you have. And, I guess, how do you approach the adoption aspect of this?
Joey Daryanani
I think you'll agree that the simple principle around "what's in it for me" applies not just to leaders, not just to CEOs, but to the front lines as well. Using that as a premise, what I see frequently, and I would encourage us never to do this, is to lead by power, is we've already made up our mind, we're going down this path. Somebody's going to deliver a PowerPoint in front of two-hundred people and expect adoption tomorrow.
I love having conversations, especially with people that are actively resisting—love it. Because then I get to learn, I get to understand what the pain points are, I get to show and tell.
What I worry about are the passive resistors, where you don't know what's on their mind, what's quiet, and for leaders, I make it a point to hit up those quiet ones more than I do ones that provide me with the noise that I need. And, one, is being very frank. I talked about trust in the industry. Here's another place where trust is super important. I need our people to trust me when I sit down and say, "Hey, this is how I believe it'll benefit you and us and the member. What do you think?" And getting them involved at the onset versus, "Hey, I just heard this PowerPoint and it's the flavor of the week or the month or you name it." I think adoption suffers and that's where most of these expenses kind of collapse, is, you never get the full benefit because people are threatened.
Sean Merat
Right. Why…how would we go get over that fear for some of the employees, you think? About that, that this is an existential [crisis]?
Joey Daryanani
Yeah. The way I personally did it, and I don't know if this is the right way or not, is when I was faced with emerging tech, AI, it was important for me to learn it first. And I took the position that I'm not the smartest person in the room and I need to learn. Where I pushed my team was the trainers, I wanted further ahead, because they were going to be the ones that would pull the people along, right?
So I learned first, and when I did that, I encouraged all of us to play. Let's just play with it. Let's just throw something in there, make it fun and see what happens. And when we did that, what I found was people had fun, and that was okay because then it wasn't a ploy.
And then I evolved from having fun with it to, okay, the way I want our people thinking about it is, what is one thing that we should be doing today that we're not able to do because of "fill-in-the-blank excuse." So in your world, Sean, what is that? Sean tells me what that is and say, "All right, now let's focus on a business problem and let's have fun."
And I'll tell you, most people have been able to adopt to it where, while there…I'm sure there are people still worried about it, it's, they're more apt to try it and say, "You know, it's fine." You know, when we introduced, I'm going to date myself, email. How many people said, "It's not going to get to where it needs to go. People are going to steal it, blah, blah, blah." And if you look back at the impact email has had on the human race, you know, it didn't take the jobs, right? It may have in the post office, but that was replaced with packages today. So that doesn't count.
Sean Merat
Right.
Joey Daryanani
But people, today, without email and the next best thing after that, there's five other things, is, we've just come to normalize that in our lives and this is how I want to think of these changes.
Sean Merat
Yeah. I think that's very sensible. We, very similarly I think, when you think back not that long ago, big organizations had buildings full of accountants, which basically tallied numbers which has been replaced by Excel sheets. It hasn't replaced the people; it's replaced some maybe, but at the end of the day, the world is full of accountants and it elevates the work that they can do.
And I think your positioning is very sound initially, which said, this is just a tool. It's not here…tool don't necessarily replace people. It's there for people to do their jobs more effectively and do better and bigger jobs. I think the only existential crisis is for, potentially, for individuals who resist the change or improvement or—and that's, that's beyond AI. I think that's beyond anything else. If you don't want to grow with the times, like, if you didn't want to adopt email, in your example, it wouldn't logically, it wouldn't and it shouldn't turn out so great and so well.
Now I think as, and to your earlier point, I think this is an interesting metric that we have. So we onboard thousands and thousands of adjudicators, claims adjudicators per client, per line of business even. And what we found is that the most outspoken and critical end users end up being the highest adoption…biggest champions, highest adoption metrics. And this is across the board. Because I think exactly, as you said, you know exactly what you need to address with them. And they're doing this not because they don't like the tool, they actually want the tool to succeed or the initiative to succeed. So those are your best champions. The passive, how would you find, the passive…what did you call them? The passive disagreers?
Joey Daryanani
Resistors.
Sean Merat
The passive resistors. How do you identify that?
Joey Daryanani
Yeah, that's the hard part, right?
Sean Merat
They're camouflaged!
Joey Daryanani
If you think about what you just said, as I look back over my career, nothing's changed. You're going to always have the active ones and the passive ones. And my secret sauce to deal with the ones that I think are passive is having those one-on-one conversations, saying, "Hey, here's how I'm thinking of things. How are you thinking of it? And, our staff meeting on Tuesday, whatever, I really want you to give your opinion." Give them some time. Right? Because at that point, they're either going to tell everybody what they think on Tuesday, or they're going to tell me what they think right now. And frankly, it's okay. Because again, as leaders, we should thank those that give us honest feedback. Because more often than not, we deliver a product that absolutely is a failure on the front lines and it's tied into some bonus or some goal idea or whatever, and now you're dealing with something that doesn't. We underestimate the power of people. And, frankly, at CSAA, that's our secret sauce.
Addressing Legal-System Abuse with AI
Sean Merat
I think that's—makes a lot of sense to me.
The, now, switching gears a little bit on the legal-system abuse. I think, there's, that's a major pressure point right now. How do you think that we could utilize some of the newer technologies, AI technologies, to better understand the litigation risk, the claim complexity, but without crossing the line for over-automation and, you know, because we know that can go down a wrong rabbit hole, and unfair decisioning, things like that?
Joey Daryanani
So if we can agree that building trust is our way of getting to the claimant first, let's talk about what's available out there, right?
I talked about building trust in person. The other way to build trust is getting that case to the right claim handler as quick as possible. Right? So that's a workflow improvement that I think we need a significant improvement on just as an industry.
The other piece is sharing of information, right? So I'm not talking about antitrust here. I'm talking about, I talked earlier about leveraging the information, the data we have internally. But now we're also starting to see carriers get together and say, "How can we share information without sharing personal information?" I'm not interested in the personal information, but I do want carriers to sort of understand what's going on because, frankly, the other side has done it and they've done a very good job with it. I'm sure I'm on a list somewhere that talks about my experience and my way of negotiating and what I do and don't make. And I think we need to do a better job. But again, I'm very encouraged and I'm starting to see things move away.
Sean Merat
Do you see any pitfalls in some of the utilization that, if deployed incorrectly, foundationally, going to lead in things, like, you know, unfair decisioning or liabilities cost?
Joey Daryanani
Yeah, so this is where I go back to human-in-the-loop, right? For us, I don't foresee, and I could be wrong, I don't foresee a time where any decision that impacts our member is done without a human. I just don't see it. And, if I'm correct, and again, I would double down on that, if I'm correct, then I think it will handle that risk because there is that risk every single time where bad things could happen and as in any technology, as in any process, as, I mean, again, take technology out of it, right? It happens today. So getting the right person involved in the right case is especially important, especially when there's a decision that impacts your customer or policyholder in our case our member.
The Future of Insurance Operations with AI
Sean Merat
Got it.
And, I guess, my final question, if we were to look five years ahead, what part of insurance operation do you think will look meaningfully different because of AI? And what part do you hope remains fundamentally human? I guess the second part you already answered.
Joey Daryanani
I just answered that, right? There is plenty of things we can do to improve workflow. And I mentioned the three that we're going after. There's a whole world around summarization. There's a significant number of initiatives around faster cycle types. We talked earlier about fraud detection and connecting the dots and who is related to whom in the world, and reducing repetitive tasks.
Will jobs change? Of course. Have jobs changed over the 40 years that I've been here? Of course, right? I've had to reinvent myself a dozen times over the 40 years. That won't change. So while I don't know what tomorrow will bring from a technology perspective, I'm sure jobs will continue to evolve. But the human-in-the-loop, in my opinion, stays.
Sean Merat
Incredible. Thank you. That's all the questions. Sorry for bombarding you with all of these, but, all has been incredibly valuable. Really appreciate you taking the time to chat with me today.
Joey Daryanani
Of course.
Sean Merat
Thank you so much.