Welcome to Insider Insights where we dive into hot topics facing the financial services industry. Today, Kartik Sakthivel, chief information officer at LIMRA and LOMA, sits down with Brent Williams, founder and CEO, Benekiva, to explore how AI is transforming the insurance claims process.
Are you ready to talk about some AI and claims?
Absolutely, absolutely.
That's awesome. So, you know, obviously, AI is taking the industry by storm, right? Every single facet of the value chain is going to be disrupted, disintermediated, and quite candidly, accentuated by AI, right? It's everywhere in the value chain. One of the key areas, Brent, where we're noticing a lot of focus as far as AI is concerned is in the area of claims, which you know a thing or two about, right?
Yeah, absolutely.
So Brent, just for our audience, what do you think are the biggest challenges our firms, our industry, face when they're integrating AI into the claim systems?
Yeah, I think there's a couple of them. One is just where to start.
You look at the insurance operations and gosh, there's all kinds of different places where the use case for AI can be made.
And my answer about, you know, those use cases have been the same for years, tested on a few use cases. If it works great, do more use cases like that. If it doesn't move on.
And the second piece is readiness for AI, you look at the insurance industry and, you know, it can be argued that, you know, the P&C insurance is five years behind banking and life and annuity and health is five years behind P&C or more, from a technological perspective. And so you've got a lot of desperate systems, a lot of old systems, some new systems and connecting and intertwining all those can be very difficult for some carriers.
That's great. So, as we look upon, to your point, or as we've talked about before, P&C is about three to five years ahead of life and annuities, right, in terms of all things technological stuff.
From an AI perspective though, Brent, I will argue that our industry — L&A, worksite benefits — by virtue of the work that we've been doing with LIMRA and LOMA, some of the best practices, some of the tools, frameworks, scorecards, we're probably better equipped for long term sustained success across financial services. But going specifically to claims, right, if you think about P&C, in terms of claims processing, and with respect to AI today, right, what are some of the most promising use cases of AI in automated claims processing that you're seeing?
Yeah, document summarization is a huge one. And one of the earlier ones that we did pretty early, you know, we started our AI journey back in, gosh, 2019, 2020.
We built our first AI engine for death certificates. And, you know, it was a really cool tool, know, in our industry, there's this need for death certificates on death, on certain death claims. And, you know, we just started asking the question, why? Well, why do you go through the whole process of waiting for an individual to get an actual test certificate? And then that individual has to mail that into the claims department. The claims department then, and this is where it gets, where it blows your mind, it comes into the mail room, the mail room scans that death certificate into their document system — document storage systems — right, and then routes that physical death certificate up to the claims department. Now, only reason that is, is so the claims staff can open that death certificate and feel the embossing on this death certificate, and then they can validate that that is an actual death certificate.
Well, what we said was, can we challenge that? Can we challenge, can AI help us solve that issue? Because one of the things I've shared with the team is anytime you take a picture, anytime you take a scan, anytime you take a photocopy of anything, those things use light and light casts a shadow. And if we can cast a shadow on an embossing of a death certificate, we should be able to train an AI to look for the shadow. And if we identify the shadow and identify this is a real death certificate, now we can extract all the data from that death certificate and the underlying data, in our case, we extract the main data, you know, the PII data, plus the cause of death. But then there's also a section in most deaths certificates for underlying death conditions. And that can be extracted and sent over to the actuarial engines for better pricing of those products.
You know, we don't necessarily need that data from a from a death claim perspective, but the carrier needs that data in order to price their policies better.
So, by building an AI engine way back then to help solve that issue, we've transitioned a lot of carriers to not need the original death certificate anymore. So, it saves that whole process of mailing that certificate in. That started our journey back in, like I said, 2019, 2020, we built that engine. And now we use it all over.
So Brent, that's an interesting use case, right? So I can see the value proposition of that particular use case as you described across the value chain, right? It doesn't have to be limited to claims.
So here's something interesting. I heard you describe a new word, right? And ninety nine percent of the time I make up one hundred percent of my words, but this was a new one for me, right? It was humanomation, right? So can you explain humanomation? Can you explain to us the genesis of the word?
Yeah. So that was a brainstorming session with the team and we were trying to think about, you know, how we can differentiate our platform from other claims and servicing platforms on the market. And the reality is, you know, technology is an enablement for humans. And so we kept coming up with, we do automation, we do extraction, we do all these things with data, and we give the human more time, you know, our biggest thing is we give the human more time to have that face to face connection with the policyholder beneficiary. You think about it when those people are going through those processes.
You know, they don't necessarily want to deal with a robot on everything.
Yeah.
There needs to be some compassion. There needs to be some empathy. But when you look at the overall claim process, there's a lot of tasks that have to be done. And the reality is this technology can take a lot of these tasks off of the plate of the claim staff. So now the claim staff can add more value and do higher value work for the insurance carrier. So we put all these words up there and we're discussing it back and forth as a team, very similar to how we came up with the name of Benekiva. And we were just taking a bunch of words and we started marrying words together.
And one of the people on the team, I actually think was Soven on the team. He came out and he said, ‘Humaomation.’ And I was like, ‘Oh my gosh, that's it!’
And literally, once we identified that word, and then obviously worked with marketing to create a quasi-definition around that word, we immediately went out and trademarked it and that's our differentiator.
So humanomation, I think is thematically, Brent, extremely applicable to what we've been hearing from an AI perspective across the industry, right? The whole concept of human in the center, human in the loop.
I will say from a claims processing point of view, that need is even more acute, right? Because no customer is calling an insurance company's claim center to say, ‘Hey Brent, how's your day going?’ They're calling you at a pretty dire time in their life. That is where the value proposition of our industry kicks in.
I love what you said about AI being able to free up capacity for folks to be able to express a little bit more empathy, to have the time to express a lot more compassion, to shepherd our customers, right? The beneficiaries through the claims process. So, with that as a backdrop, Brent, what role do you believe humans should retain in the totality of the claims process?
Yeah, it just depends on what the use case is. You know, for example, back to that summarization, you know, an AI engine can summarize thousands of pages of medical documents, but human, it would take them weeks to do that, right? So why would we even have a human do that anymore? Why not have the AI do the summarization and then have the human just do the checks and balances of the summarization. And if parts that, you know, in the summarization that don't quite make sense, well, then you can actually, you know, go to the exact point in the medical documents — or whatever documents you're looking at — and then make a make a human decision on that.
But the reality is, if we can shrink that timeline, it takes to go through the overall claims process and summarizing, like I said, thousands of documents as is part of that process. Well, then it just makes sense that we can offer a better experience to policyholders and beneficiaries. And the reality is, you know, in my prior business as a financial advisor, I used to joke all the time, but I haven't had a single person buy a life insurance policy from me hoping to use it. Nobody does, right? But they expect that when somebody files a claim on that policy, that it'll just be handled, it'll just be taken care of.
Another great use case for technology too, is knowledge transfer. You look at the average tenure of claims and servicing staff at any carrier. And I would argue that it's a long time.
And, you know, there's some stats out there that have some studies that have been done that, you know, we've got more people in the insurance industry leaving the industry than coming into the industry.
Oh, that's a problem. You know, that's a problem that we're going have to deal with. And when you go into a carrier — and this is a true story — went into a carrier and they said, you know, that that lady over there, she's been with us forty one years and she can't go on vacation without taking a company laptop with her, because of the knowledge in her head. If something comes up, you know, to me, that's a business risk.
Yeah.
You take that knowledge and you need to, you need to put that somewhere. You need to, you need to be able to transfer that.
And the, and the other thing, the third thing is, the new people coming into the industry.
You know, you take a person right out of college or, you know, entering the workforce, and you put them in front of a green screen and tell them to arrow this function F that, you know, by the end of the day, they're looking for another job.
They're not going to work in technologies like that.
Yeah, no, that makes sense. And you know, what's interesting is you and I have talked about this multiple times, Brent. What was the adage, right? AI is not going be for your job, but your job might be at risk from someone who knows how to use AI, right? So, with that as a backdrop, in your definition, in what ways can the concept of humanomation enhance employee roles rather than replace them?
Yeah, it's getting them, taking those, I hate to say mundane, but some of those tasks that a human just shouldn't be doing, and we're all guilty of it, right? We all have tasks in our jobs that we probably are not the person suited to do those tasks. You look at the cost benefit analysis of an employee, you have a higher-level employee probably shouldn't be doing, you know, checks and balances on emails, right? You know, so that's a great use case for technology to come in and do those things. I look at it as an enablement.
I leverage AI every single day in my job and when AI really started picking up pace, very similar to what I did on the sales side, I said, I need to learn about this. So I rolled myself in an AI mastery class and I went through it. And the more I learned, the more I was excited and wanted to use the tools. I mean, it got to a point where there was one portion of the class where they're telling us all the different tools that are out there and what those specific tools do. And immediately, go out and I buy that one, I buy that one. And I said, okay. Well, I do a lot of Google searches. So, yeah. I probably need perplexity and, you know, so I just start buying all of them. But the interesting thing is now that I'm so ingrained in them and using them all the time that I find myself using the different tools for different tasks.
Yeah.
And you know, I think, but again, I just put myself in everyone else's shoes. It enables me to take something that would have taken me twenty minutes, right? Shrink that down to twenty seconds. That's huge from a capacity perspective. So, you know, when you tie it back to the earlier thing that I spoke of, we have more people leaving the insurance industry than coming into it.
Well, if you can have one person do the job of five, well, that's a partial solution to that problem. I'm not going to say it's the solution, right? But it's a partial solution.
The way I've been characterizing it, Brent, is if there are rote repeatable operational tasks that you perform as part of your job, think about how you can automate them and free up your own capacity, right? So, one of the things that is needed in our industry is for individuals to develop some level of domain knowledge, build up your domain knowledge. So imagine if you will, if you have a person on your claims team and you can free up ten percent to fifteen percent of capacity and invest in taking LOMA 281, LOMA 29, their FLMI journey, to be able to invest in educating themselves about the domain, and think about how indispensable that individual would be when they have a much broader level knowledge about the function that they're serving. I think it's a remarkable opportunity for us. So let me pivot to this question. So, I've said this to you before, Brent, but the true essence of who we are as a human being comes out when we dial an 800-number and you have an AI on the end, right? And look, AI bots have progressed significantly over the past IVR, right? The telephony has progressed significantly in the past five years.
We don't have to aggressively yell operator or punch out to zero, right? There is that balance that we need to strike between having that. We're a human-to-human industry, a person to person industry, a connection-to-connection industry, right? It's founded in relationships. So how do you strike that balance, Brent, between AI driven claims automation and humanomation of the industry? Where's that balance?
Yeah, it's different for different claims processes. You know, it could be argued for certain claim processes. You don't need a human involved at all. Right? Dental, vision, those types of claims, pet insurance. I mean, you take a picture of the invoice, and the system does the calculations you know, shows up in your bank, right? I don't need to talk to anybody. But for those more complex claims like long term care claims, disability claims, certainly death claims that involve, you know, empathy throughout the process. Those claims are the ones where we tell every carrier, you know, the benefit to Benekiva or our solution — Humanomation solution — isn't taking your current process and duplicating that in Benekiva. The benefit is reimagining your entire process.
And now think of it from a perspective of where and when does the human or should the human be involved? And when you look at your overall process from that perspective, you'll have some amazing a-ha moments and you'll have a-ha moments that yes, I don't need to be doing this mundane tasks. We've got tools that can do that. Right? But when you get to the point to where it's time to interject, you know, put the human in the loop humanomation in the into the process. Now, all of a sudden, you've got the extra time and capacity to be able to do that. And it really, you know, I read the, I don't know if you've ever read the book on Zappos, but I, you know, Tony Hsieh, he wrote this book on Zappos. And the thing that amazed me about that was they were passionate about their servicing staff being on the phone with their customers. There were servicing staff that would be on the phone for three hours.
Think about that. There are shoe companies selling shoes online, and they've got servicing staff that will be on the phone with you for three hours to discuss shoes and those things. That's a customer experience that you know, is unmatched. And then you look at the other side of that coin, right?
There are some people that want that handholding process, so to speak. But then there's the other side of the coin, the people that want the Amazon-type claims experience, right? I click a button, I take a picture, I'm done. Right?
So, the reality is both of those types of people and clients are out there. So, we have to be able to accommodate both.
And it's literally where meet them where they're at, whether it's the carrier side or the policyholder beneficiary side, because the carrier side, maybe they're not ready for AI. Great. We have a solution that doesn't include AI. Right? And just meet them where they're at.
I love that. Engage with how people want to be engaged, right? I think people react favorably with options. Now, Brent, as we round the corner of our chat today, you've been in the industry for a while, right? You've seen a lot, you've done a lot.
I know five years is an eternity in insurance, especially when it comes to the intersection of insurance and technology, specifically with AI. But if I were to ask you, I'm like, ‘Hey man, just pull out your crystal ball.’ What do you see the evolution of claims processing in our industry in the next five years? What would you tell me?
Well, I've said since day one, that founded Benekiva, that I believe there is going to be a day with rules, roles, data, verified data, where a policy holder or beneficiary won't even have to file a claim, where the claim will just be processed.
I believe we as an industry can get there. Now granted, there's going to be some things we're going have to, you know, deal with from a regulatory perspective, but the reality is, is we can collect data from multiple different sources and come to a high probability conclusion that that is a claim that needs to be paid and we should just pay it.
So, I may never see it in my lifetime, but my dream is you buy an insurance policy and when an accident or something happens, the policy just pays. You don't have to file anything. Just, you know, you'll get a message from the carrier saying, ‘Hey, Kartik, sorry that you were in that accident. We're going put a little money in your bank account, take care of this,’ whatever, or ‘Hey Kartik, I'm sorry for the loss of your loved one. The money will be in your bank account in forty-eight hours.’
At the end of the day, Brent, regardless of whether it's technology, any flavor of it, or AI specifically, that is the value proposition. That is the promise of our industry. We're going to be there for our customers in the time that they need it most. Right? At the end of the day, that's what it's all about.
I've said it for years, and even when I had my financial advisor firm, the role of an insurance company, the whole reason we have insurance companies — in my opinion is to pay a claim. Now granted, they have to, you know, pool everybody's money and go invest that money and make money in order to do that. I get that. But the reality is, people buy insurance policies, so that when and if they have a claim, it gets paid.
Wonderful. Awesome, awesome conversation, Brent. We’ve got to do this again, yes?
Absolutely. Absolutely.
Right, bud. Good talking to you, Brent. We will see you again soon.
Sounds great.
Thanks for listening to LIMRA's Insider Insights podcast series. To hear future podcasts, subscribe at LIMRA.com/podcast.
