I had the briefest of interactions with Warren Buffett — and he nailed it. 

When an authorized biography of him was published in 2008, I wrote a very favorable review for the Wall Street Journal. This was just weeks after my own book, "Billion Dollar Lessons," had come out, and I took the liberty of mailing Buffett a copy, along with the WSJ's rave review of B$L. 

Buffett had a cameo in my book (written with Chunka Mui) on lessons to be learned from business failures because he once invested in USAir while it was in the midst of doing a bunch of dumb things. I figured that connection, plus my having written a review he surely liked, might merit at least a glance. Who knows? Maybe he'd even read parts of the book and say something nice while Chunka and I were out hyping it.

Exactly one week after I mailed the book, I received a return letter from Buffett. He thanked me for the book, adding:

"Yes, that investment in USAir was the worst I ever made. I expect to make a worse one soon."

Buffett has said a lot of folksy, smart things to a lot of people over the decades, and I've been reading as much as I could for more than four of those decades, so I thought I'd mark his retirement announcement with some of his greatest hits, especially ones that apply to insurance. I'll start with my favorite: 

"It's when the tide goes out that you see who's been swimming naked." 

I've always liked this line because of its sense of accountability. For decades, I've watched companies try to blame troubles on anything they could lay their hands on — an earthquake in Japan, storms in Europe, sure, whatever, whether or not they did much business in the affected area. But the best companies just kept their heads down and worked their way through the problems, making sure they kept their swimsuits on even as the tide went out. 

Look at Geico. After Berkshire Hathaway acknowledged in 2021 that it had fallen behind on telematics, Geico worked and worked and caught up — as Matteo Carbone described for us last summer. Even as supply chain problems and bad driving habits left over from COVID caused many auto insurers to try to raise rates in a panic, GEICO had a combined ratio in the first quarter that started with a 7. (It's not just GEICO. Progressive, which pioneered the use of telematics to price risk, never had a blip and recently announced plans to hire 12,000 people.)

(If you're interested in learning more about Buffett's pioneering work in insurance, I recommend this piece by Adrian Jones.)

Here's another great one: 

“If you start fooling your shareholders, you will soon believe your own baloney and be fooling yourself, as well.”

My favorite study of all time is one by BCG that Chunka and I cited in B$L. It found that 80% of executives thought they had the best product in the market — and that 8% of their customers agreed. Surely influenced by that, in a cynical moment Chunka and I wrote in our book that "marketing is when you lie to your customers; market research is when you lie to yourself." 

Companies, including insurers, would be so much better off if they could take a brutal look at themselves.

(That quote comes from this article in the Washington Post, which includes a number of other worthy lines.)

"It takes 20 years to build a reputation and five minutes to ruin it. If you think about that, you’ll do things differently." 

'Nuf said. Insurers know this all too well.

"Someone is sitting in the shade today because someone planted a tree a long time ago."

While sitting in London's Hyde Park once, I marveled at the grace and beauty of a section enclosed by trees that had been espaliered — the leaves and branches formed what you could think of as a very broad, perfectly manicured box hedge extending from maybe 40 to 50 feet off the ground. I realized that those trees had to have been planted many decades before to grow to that height and be trained so well. So whoever planted those trees surely didn't expect to experience the serenity I was being allowed to appreciate. 

I dearly wish that more long-term thinking could exist in business, including insurance. Insurers do a better-than-average job of thinking about the long term, but we still get buffeted by tariffs and storms and so on and need to focus on that next quarter. I'd love to see more companies taking out a clean sheet of paper, designing the perfect version of themselves 10 or more years out and driving toward that vision.

My old friend and WSJ colleague Roger Lowenstein notes that as recently as this weekend, Buffett responded to a shareholder question by saying, “We don’t do anything based on its impact on quarterly and annual earnings. What counts is where we are five or 10 or 20 years from now.” 

How great would it be if the rest of us could adopt that attitude?

Well, Roger provides some numbers in an op-ed in the New York Times:

"Since [Buffett] took the helm of Berkshire — on May 10, 1965 — General Motors, then the largest American corporation, has greeted 11 new chief executives. Sears, Roebuck, the biggest retailer, has vanished from the scene. Eleven U.S. presidents have come and gone (two of them having survived impeachment and one forced to resign), and Coca-Cola changed its formula, but Mr. Buffett didn’t change his....

"Berkshire’s stock that day in May closed at $18 a share. When he delivered the news [of his impending retirement[, it was above $809,000 — almost 45,000 times as high. Over the same span, the Dow Jones industrial average is up just under 45 times."

Words to live by.

Cheers,

Paul

P.S. It's not clear that Buffett ever did make a worse investment than the one in USAir, at least by his telling. He once wrote of airlines: "A durable competitive advantage has proven elusive ever since the days of the Wright Brothers. Indeed, if a farsighted capitalist had been present at Kitty Hawk, he would have done his successors a huge favor by shooting Orville down.”

Paul Carroll

To start out, could you tell me how you think about chatbots?

Adam Fischer

When I started at Clearcover eight years ago, I was very anti-chatbot. They were these tree-based, logic-based chatbots that everyone hated using. Let's be honest. It didn’t matter how friendly you tried to make them. They just didn't work.

So we focused on our mobile app, and that has really paid off for us. Over 90% of our customers have downloaded and created an account on our mobile app. It's only for our customers, so people don’t do any shopping in the app. We just wanted to make it easy to interact with your policy after you buy it, whether that's getting your ID card, handling changes on your policy, or filing claims. Our app has always been at the epicenter of what we've done.

As generative AI took off, we started to say, okay, we finally have chats you can automate, and they are a good experience. We’ve really leaned into Gen AI for about a year, and it’s been a tremendous success both for customers and for Clearcover.

Paul Carroll

Tell me a bit more about what you’re doing with Gen AI. I've yelled at my share of chatbots over the years, and I’m still not finding them to be that great.

Adam Fischer

That’s a good question. Everything depends on the context. When it’s tight, and the chat is iterative, the AI is very accurate.

If you’re using a general tool like ChatGPT, the nexus of knowledge you’re interacting with is the entire internet. And the old saying is right: Garbage in, garbage out. You remember the famous story from five or six months ago, when somebody asked how to keep cheese from sliding off a pizza.

Paul Carroll

And the AI’s answer was: Use glue.

Adam Fischer

Right? And the technology was actually performing well. It had sourced different pieces of data and found a joke post on Reddit.

But when you have a very controlled use case like we do, there is a solid knowledge base with all the information the Gen AI tool needs to get the user to the correct answer fast and accurately. That's where Gen AI can really be powerful.

Paul Carroll

How do you quantify the benefits for Clearcover?

Adam Fischer

As a growing carrier without a large call center, we implemented Gen AI so we can answer a lot of our customers’ questions 24/7 now. A customer might have a question about billing. They might have a question about their policy. Now if it's 11 at night, one in the morning, maybe they were driving around and had a bit of an issue: Whatever the case may be, they can get that answer whenever they're interacting with us. We serve the customer in the moment when they want to be served.

As we're growing, our hiring curve doesn't have to be as steep. And the folks who are already here in our call center can focus on providing better service to customers when they do need to get in touch with a human.

Paul Carroll

My favorite geek joke is: Why did it only take God six days to create the universe?

Answer: Because God didn’t have an installed base.

I assume there are things Clearcover can do more easily than older carriers can with their legacy systems.

Adam Fischer

We're completely API-centric from the perspective of our custom-built policy administration system. So our chatbot can interact with it in powerful ways. If you have a question about your policy, the chatbot can pull up your specific policy and use that to provide your response.

As we see patterns, we try to integrate the chatbot more deeply. Because billing is a good portion of our chats, for instance, we've started integrating with our payment provider so customers can pay by link during those interactions.

Paul Carroll

How do you handle the coordination between your human agents and the chatbot?

Adam Fischer

For example, if you want to add a vehicle to your policy, we can gather a large portion of the information via the chatbot and then have a human complete the change and make sure everything is accurate.

If you have any questions, we're here, but we take a lot of time out of that conversation, which then lets our humans use their superpowers to provide really good service when they do get connected with the customer.

Paul Carroll

What are you finding people tend to do with the chatbot?

Adam Fischer

The questions are consistent with what they’ve always been. We get a lot of questions about billing and about adding vehicles and drivers or, say, about whether a customer needs to buy coverage from a rental car company.

We’re always keen on compliance and make sure we're not giving coverage recommendations or advice, but we can help you retrieve information you could find on our website.

Paul Carroll

I've often heard insurance companies say they want to interact with the customer more, because studies find a link between number of interactions and loyalty. But I think that’s often correlation but not causation. Right? Sometimes I want you to interact with me, sometimes I don't.

When do you reach out to customers as opposed to just waiting for them to come to you?

Adam Fischer

We focus on providing a good experience when we’re needed.

But we’ve been running a service pilot called Car Care for a little while, where we make it easy for customers to save money and book on common maintenance they need for their vehicle. Customers can browse shops and book online to save on oil changes, new tires, or whatever. We don’t bombard you with emails. You won’t get happy birthday emails from us or, “It’s July, so here are some barbeque tips.” But if you use a feature in our app that lets you report your mileage to us, we’ll let you know that you’re due for an oil change. We’re going to remind you of the benefits we offer, but we don’t assume insurance is at the top of your mind all the time.

Paul Carroll

I’m the perfect insurance customer. I’ve never filed a claim for anything other than healthcare. But people who do file claims often express frustration about how little they know about the progress of their claim. Shouldn’t it be easy these days to keep people posted, like Amazon does on its deliveries?

Adam Fischer

We have a claim center within our app that tells you, when you initiate a claim, what the process will look like and how long each step may take. The app also has a sort of Domino’s pizza tracker that will tell you where you are in the process.

We're also working on a product that lets people provide information to our AI on a claim, so they can do it on their own schedule and not have to wait for a rep to call or schedule an appointment. The Gen AI, which we internally affectionately call TerrenceBot, or Terry for short, will also answer detailed questions, like, Okay, what now?

Paul Carroll

If you and I reconvene in two years, where are we? Terry is out to the customer. What other things are going on?

Adam Fischer

The Gen AI is just going to keep getting better and better, so it’s going to be able to make the claims process more efficient.

And there's more than just the first party in a claim, right? There are third parties. There are passengers. There are other folks involved. We're planning on kind of spidering out our Gen AI to those interested parties, as well.

If you think about how a typical claim is processed, well, if the customer tells us there was a passenger, our rep has to call the passenger. Passenger's unavailable. You leave a voice mail. You play phone tag for a while. Finally, you get the statement. Now imagine a situation where we contact the passengers via bot and can collect that entire statement via the Gen AI and then have our human reps call to confirm some pieces of information and provide a nice finishing touch.

Reps won’t be spending all their time chasing x y z piece of data. Gen AI can be very good at doing that.

Paul Carroll

That sounds great. Thanks, Adam.

About Adam Fischer

Adam Fischer is chief product and innovation officer at Clearcover. He brings more than a decade of experience at industry leading consumer brands like Apartments.com and Redbox. In addition to being the first product leader at Clearcover, Fischer has worn many hats on and off during his tenure, including overseeing the rapid expansion of the technology department from 30 to 160 employees during a two-year stint as CTO as well as over 100% YoY growth for two consecutive years as head of growth. His latest role is creating an innovation team at Clearcover, which is currently responsible for Clearcover Advantage, which is aimed at delivering superior value for policyholders. He remains focused on embedded strategies, constantly striving to improve on the foundation he helped establish at Clearcover. He uses his passion for technology and building products to help bring industry-leading experiences that delight customers. Fischer holds a BS in information systems from DePaul University.

Selling insurance, whether automotive, home, life, or other types, has traditionally been a relationship-based experience. Many agents work with clients for years, and knowing their customers' evolving needs is key to upselling and building a book of business. As artificial intelligence (AI) seeps into every industry and consumers are more cost-conscious than relationship-driven, insurance companies and agents are taking a critical look at the technology to determine how it fits into the insurance business model.

On the surface, AI can lessen the administrative burden for agents who answer frequent and simple inquiries, while also helping to process claims, identify potential risks, and deliver personalized plans based on historical data. Up to 20% of claims filed are fraudulent, and AI is analyzing patterns to help insurers identify which cases are legitimate.

Yet, pushing too much of the relationship to self-service, automation, and bots can hurt brand loyalty.

While there are many potential benefits to AI, how it is implemented within the structure of the business is key to using it effectively and supporting a better customer experience.

Barriers Between Insurance Companies and Customers

Industry data has identified a generational divide between younger and more technologically savvy customers, who prefer digital solutions, and older generations, who prefer traditional phone-based services where they can speak directly to a human.

Across generations, some people are open to using self-service chatbots, automated SMS messaging, or AI agents, while others are less interested in this style of communication. In fact, up to 40% of people feel "unfavorable" toward chatbots due to past negative experiences or a lack of trust in the technology.

To protect the customer's experience when researching or purchasing insurance, filing a claim, or requesting support, insurance companies should consider an approach that integrates the human touch with elements of AI. This approach will ultimately optimize efficiencies and cost savings without sacrificing the quality of the member experience connection.

Taking a Human-First Approach to Filing Insurance Claims

There are several key downsides when human oversight is left out of the customer experience journey. Many interactions with insurance companies follow stressful experiences like a car accident or home damage caused by a natural disaster. While efficient, automation does not have the capacity to deal with these situations using empathy.

Customer needs are too complex for AI. Allowing a human agent to be the first touchpoint in the journey ensures that the customer is receiving personal and empathetic support, lessening their stress and anxiety and helping them through difficult scenarios.

For example, if someone experiences hail damage and is looking to file a claim, they may contact their insurance agent via the app, phone, or online. From here, a human agent can evaluate the customer's needs and communicate the best route for effectively filing a claim while helping to put the customer at ease. Then, it is up to the agent to decide if and when AI should be used.

In this case, AI could deliver basic information such as next steps in the claim process, common safety measures homeowners should take after a hailstorm, such as securing broken windows, and how to photograph damage. AI can also gather data from the customer, such as date and time of incident, address, and property details. AI can help schedule an inspection with an adjuster and automatically input all data into the customer record.

During this part of the transaction, it is important that the customer has the ability to reconnect with a human agent if they have questions or concerns to ensure they are not stuck in a frustrating loop with a chatbot or AI agent. Balancing this combination of human and AI interaction creates a sense of personalization, supports empathy, and frees agent time by offloading common or administrative tasks. It also supports brand loyalty because the customer feels supported by their insurance agent and always has a path back to a human.

Continued Optimizations for AI Advancements

Insurance companies should continue to monitor technology advancements and be open to adapting customer service models as AI evolves. There is not a one-size-fits-all approach when it comes to AI and automation. As roles and responsibilities of human agents continue to shift due to AI, it is important to document where humans and AI each add their own value to new and existing processes. Finding a strategy that effectively balances human support and AI will lead to increases in productivity and efficiency while still ensuring that customers are highly satisfied with their experience.

Technology plays a pivotal role in transforming the insurance industry, but it's not always an easy relationship. Many insurers struggle with outdated work methods as well as legacy systems.

While most insurance companies view technical debt as a major hurdle for innovation, it's easy for them to overlook the other two legs of the stool: process debt and organizational debt. These three legs work together to form a complex system, and neglecting any one of them can lead to stagnation.

Tackling technical debt

Technology is the backbone of the modern insurance industry, yet many companies still grapple with how to replace, integrate or phase out their older technologies.

Approximately 70% of IT budgets is consumed by legacy system maintenance, according to Forrester. Meanwhile, insurers struggle with complex integrations that are costly and hard to implement. As a result, they are constrained by outdated, siloed ways of operating.

For example, an outdated policy management system used by a life insurance company can result in long claim settlement times and difficulty complying with new privacy regulations.

The good news is that cloud, AI and other solutions can help insurers modernize their technology infrastructure and work more efficiently. For example, AI tools are available to help with migrating, consolidating and even converting a company's multiple legacy policy administration systems into a more modern, future-proofed solution.

Overcoming the weight of process debt

Insurers are under pressure to accelerate growth and innovation, streamline operations and provide faster, more reliable services to policyholders. However, they are often constrained by complex and highly manual, outdated processes and workflows. This leads to wasted time, money and productivity.

HFS Research estimates the insurance industry is burdened by $66 billion in process debt, which is the buildup of outdated, overly complex, or inefficient workflows and practices that made sense at one point but no longer align with goals or realities.

The key to overcoming process debt is to identify and address its root causes. This requires a thorough assessment of current workflows and practices, followed by targeted interventions to streamline and simplify processes. Then, businesses can recapture lost productivity, reduce waste, and ultimately achieve their goals more effectively.

For example, new technologies like smart workflow systems and persona-based portals help connect different front-, middle- and back-end tasks (like customer service, policy administration and billing/collections) so they can be completed automatically. This allows a company's external users (e.g., customers and producers) to do front-office work on a self-service basis, freeing internal staff to focus on more complex middle- and back-office tasks. As a result, insurance companies can offer more modern, efficient and personalized experiences for their customers.

Unburdening organizational debt

When insurance companies tackle technical and process debt, they often overlook the accumulation of inefficiencies, outdated practices, and structural impediments that hinder their ability to adapt and evolve with the times. In addition, organizational debt accumulates when the knowledge of these products, processes and procedures is not documented effectively and is only available from an aging workforce.

Think of organizational debt as the "interest" an organization pays for not addressing these problems. Overcoming it requires a fundamental shift in how teams collaborate, how culture manifests and influences decisions and overall team dynamics, and how customer needs are met at every turn.

For insurance companies, this means understanding the individual experiences customers crave — from preferred channels to accurate recommendations. It means using data intelligence to identify specific touchpoints that meet customer needs. It also means making sure that all the different departments in the organization (claims, finance, legal, underwriting, etc.) are aligned to the same goals.

From debt to innovation

Paying down each of these three types of debt isn't easy, but it is a worthwhile goal to pursue.

It requires a holistic approach that involves upgrading technology infrastructure, streamlining workflows, and aligning organizational culture with modern practices. New tools and solutions can help by automating manual processes, improving data visibility and reducing overall risk.

By shifting the focus from maintenance to innovation, organizations can explore new ways of working, create a culture that is flexible, adaptable and forward-thinking and be free to focus on what truly matters: delivering value to their customers.

The International Insurance Society's 2025 Global Priorities Survey found that two-thirds of the senior insurance executives surveyed listed artificial intelligence as a top priority for their technology and innovation agendas. That is up from 55% last year and represents a huge increase from 17% in 2021. 

But it also means that one-third of the executives DON'T think AI is a top priority for innovation. Hmmm.

The survey also found that "concerns over the speed of technological advancement have eased." Really? 

I just published an article that predicted that an AI available to insurers a year from now will be 10 times as powerful as today's, at 1/100th the cost. Whether that's precisely right, it's certainly directionally correct. So I, at least, am thoroughly discomfited by the speed of change and can't imagine why others aren't, too.

The IIS survey provides a great baseline every year for understanding the state of play in the insurance industry... and I have thoughts.

The survey — which I encourage you to preview here — includes a number of responses that suggest executives are alive to the possibilities of AI. For instance, among internal priorities, operational efficiency is a top issue for 50% of respondents, making it the highest priority for the second year in a row. I suspect that emphasis doesn't just reflect a need in a highly competitive environment but also shows an understanding of the huge number of relatively straightforward opportunities that generative AI presents for automating processes. 

I think the possibilities of AI also show up in the near doubling of respondents who said the aging workforce is a top priority (from 11% last year to 20% in 2025). Again, there is a huge need, given that hundreds of thousands of insurance company employees are expected to retire over the next few years. But AI also presents great opportunities, both to preserve the knowledge of those walking out the door and to provide data and tools to new recruits that can bring them up to speed much faster than in the past. 

The responses on cyber seem to incorporate some AI optimism, too. The percentage of those identifying cyber security as a top three priority in the political and legal category dropped to 57% in 2025 from 75% in 2024. While AI certainly makes hackers more effective, the good guys seem to be using advances in technology, including AI, to improve defenses at least as fast as the attacks are intensifying.

It's certainly encouraging to see a huge increase in the number of respondents saying they are focused on addressing technological advancements — 41% in 2025, up 16 percentage points from 2024.

But I worry that too many executives are still too complacent about all the change that AI will effect. Yes, we're almost 2 1/2 years into the generative AI era, and the sky hasn't fallen. But Amara's Law is undefeated. It says we overestimate the effects of a major technology change in the short run but underestimate its effects in the long run, and we're starting to move into the long run. 

I think insurers are getting a pretty good handle on the operational efficiencies available to them, but they should be acutely aware of the larger possibilities. Someone may figure out how to reinvent processes for claims or underwriting or to radically improve agents' and brokers' productivity. There's also a huge amount of effort going into producing AI agents that can operate as, essentially, employees, with considerable autonomy. Imagine a world where you can give every employee 10 or 20 or 30 AIs that work for them at essentially zero cost.

So I'm delighted to see the emphasis on AI and innovation in this year's IIS survey. I just want to be sure we don't get comfortable.

Cheers,

Paul 

In today's insurance market, policyholders have more choices than ever. Switching providers is quick, easy, and often encouraged by comparison tools and challenger brands. And while price has traditionally been the battleground, it's now only one part of a much bigger picture. According to new research from Nuvei, nearly half of policyholders who switch insurers do so for reasons unrelated to cost.

At the center of this shift lies the claims experience, and more specifically, the speed and flexibility of payouts to customers.

Why faster payouts are the future of customer loyalty in insurance

For many policyholders, filing a claim comes during moments of stress or financial uncertainty. They expect insurance to provide reassurance, yet many are met with delays, outdated processes, and inflexible options. Long wait times, lack of transparency, and rigid payout options erode confidence—often permanently. In fact:

• The average claim lifecycle exceeds 100 days, while most policyholders expect significantly faster resolutions.
• 48% of policyholders say they would pay more for a faster payout, proving that speed isn't just convenient—it's valued.
• Only 35% of claimants receive direct deposit payouts, despite 58% saying it's their preferred method.

These delays can seriously affect customer satisfaction. A slow payout undermines confidence in an insurer's ability to deliver when it matters most. With 40% of policyholders switching providers annually, that perception can be costly.

How flexible payouts can give insurers an edge in a highly competitive market

Flexibility is increasingly essential. Policyholders want to choose how they receive their funds, whether that's a real-time bank transfer, digital wallet, or scheduled installments for larger claims. Meanwhile, 18% were still paid by check, introducing additional wait times and banking steps.

When insurers fail to provide this flexibility, frustration builds. The result? 19% of claimants report struggling to access their payout, reinforcing the belief that claiming is more hassle than help.

Meanwhile, digital-first insurers are raising the bar. With streamlined onboarding, transparent communication, and instant payouts built into their platforms, they're capturing market share from traditional providers who haven't kept up.

To stay competitive, insurers must stop viewing payouts as a back-office function and start seeing them as a core part of customer experience and retention.

The bottom line?

Faster, more flexible payouts build trust. They increase satisfaction. Ultimately, they give insurers a lasting edge in a market where loyalty is harder than ever to earn.

Discover more insights, data, and strategies in Nuvei’s latest whitepaper: “Mind the Claims Gap – Why UK Policyholders Are Losing Faith in Insurance Products, and How Payments Can Fix It.”

You can't be everywhere at once, and your training and processes can only be so good. Adjusters, especially junior adjusters, can miss things in claims files that make a big impact on losses. These might be big misses, like with deadlines, or might be smaller misses that have a big impact, like whether contributory negligence was present. Both these kinds of errors can have a significant impact on your bottom line.

Claim leakage can be avoided with better tools, and AI is the ideal solution. AI can handle most aspects of file review with accuracy and consistency. However, you need a trained AI model to yield true value.

If you already use a product like OpenAI's GPT, you know that it can run into issues with deep or complex issues. Even many of the newest models that help with deed research still run into problems with lengthy and detailed output at speed. However, you should not compare your experience with the publicly available models on the internet against the quality that AI-specific insurtech companies can provide. AI is amazingly accurate when properly directed and trained.

Well-trained AI can handle virtually all aspects of a file review. After a file is closed, AI can also supplement the audit process to ensure your carrier's best practices were followed. AI has the ability to review tens of thousands of pages and compare any checklist against the claim's ultimate outcome and payout. AI can be incredibly proficient at this kind of outcome.

Using AI does not mean you (or your team) abdicate control over claim files, or the review of files. You should still validate the data. However, because AI has the ability to cite to specific pages relevant to its analysis, this process can be significantly sped up. This is especially true with mountains of medical files that are not relevant to the claim, or significant witness statements or communication logs, where only small bits of information are helpful or relevant.

This kind of review catches mistakes quickly and can be a terrific learning tool for your team. AI can speed up training time for new adjusters, who can immediately see areas of files that they may have not considered. Even when adjusters are manually trained, AI can be implemented to validate results and facilitate faster understanding of the key job functions and key performance indicators.

AI is not perfect, but neither are humans. The good thing about programming AI is that it follows your instructions every time the same way. Even if some of its output needs adjusting here and there, AI can be given a checklist of 10, 20, or even 100 things to review for every single file. It can effectively replicate your best and brightest over and over again.

Replicating your "best" is a key point as you consider software options. You want an AI that will replicate the best practices in your organization that are followed by the top 1-5% of all adjusters (or whichever team you are considering the use of AI with). Good AI software will start with your output in mind and work backward to determine what data, and ultimately what type of AI focusing, is necessary to produce that best output.

This process is different from a technology company that wants reams of data to "find" and tell you about your company's best practices. This generally produces average results (at best) and requires substantial internal training and focusing time to get you something very useful. The process takes a very long time, does not succeed, and costs substantial amounts of money. The reason why this can fail is in the inherent nature of how AI learns.

If you give any AI system 100,000 documents and you ask it to provide you with key concepts from all those documents, it will do a good job at summarizing them. It may even produce parts of a usable document as a template. This is because AI works by looking for correlations in the content of the documents. It is going to look for the things that most commonly appear. If you think about all the sections within the 100,000 documents as appearing on a bell curve, AI is going to go for the meatiest middle part of that bell curve. It will give you the results that closely match the middle because it is looking for correlation among the documents.

The issue with AI giving you the meatiest middle part of the bell curve is that the middle is the average. Nobody wants the average. Mitigating risk and reducing losses isn't about catching the average issues within a file - it's about catching the absolute largest number of issues no matter who is reviewing the claim. Average seems helpful in theory but is a failure in practice.

You do not want AI to produce average results, so you do not want it to evaluate the middle section of the bell curve. You want it to give you the very best, which means you want the results from the right-most area of that bell curve that represents the top of the top results. Conversely, you want AI to stay away from the very worst examples that reside at the left-most area of the bell curve - the place where the majority of leakage resides.

To get the best from AI, you must instruct it on your best practices. "Best practices" can mean either the best process/checklist you use, or the best example of a report that contains all the data you expect to see from your best people. Once you instruct AI on the best practices, then you can move backward into the reams of data to fill in the content. With the right application layer that directs AI, the results can be truly remarkable. This does not require creating a large language model just for your company's use, but rather harnessing smart applications built on top of the existing models.

Remarkable results can help reduce risk through better and more consistent file analysis, whether by an adjuster, outside counsel, or as part of a file audit. It can also reduce staff time by removing much of the labor-intense review of files that can take hours or days. Because AI doesn't get hungry, stressed, or tired, the time savings also means higher quality.

AI can offer greater benefits beyond time and file management. For example, AI can identify red flags in files, like excessive treatment, pre-existing conditions, or missing documents. It can provide an adjuster with a clearer understanding of property damage or bodily injury to better assess the claimant's demand. Using AI can even reduce the likelihood of a claimant getting counsel because an offer can be made within days versus weeks of the first notice of loss. The faster an offer is made, the less likely the claimant is to hire a lawyer.

Addressing demand letters is a new and powerful use of AI that smart carriers are implementing immediately. The plaintiffs' bar is already using AI to produce those demand letters, and the companies creating them brag about how much more money their AI-generated demands yield. One demand-generating company that recently raised funds on a billion-dollar valuation advertises that its users are 69% more likely to max out policy limits.

AI can effectively be used to counter these demands by recognizing holes in the file and presenting those to claimant's counsel. This includes identifying holes in coverage, such as endorsements or intentional conduct that might reduce or eliminate exposure. AI can do an initial review of liability by comparing police reports and witness statements to determine causation, and even flag contributory negligence and the lack of mitigation of damages.

As part of a file review, AI can also analyze damages and whether those appear excessive in light of the injury or economic information in the file. These kinds of robust demand responses point out all the ways a claim's value is not as high as the other side believes. This can yield higher leverage and lower payouts through appropriate risk analysis. This kind of analysis and response would be ideal for every file, but it takes a lot of time to do manually. AI offers the ideal solution, with the ability to produce a comprehensive response in less than five minutes.

AI can also reduce employee and customer churn. Using AI can lead to greater job satisfaction for adjusters and for customers. Your employees all of a sudden get to focus most of their time on the things that bring purpose and meaning to their jobs. They get to think more about strategy, talking to stakeholders, and analyzing files versus simply sifting through piles of documents that AI can do faster and more accurately anyway.

Customers are less likely to churn as claims are resolved faster and at fairer, more consistent valuations. When AI follows the same standards in every file, then variation in claim payouts stabilize, leading customers to appreciate the transparency and speed in which their claim is resolved.

The benefits of using AI are many, but it is not perfect. However, when AI reduces the time and energy it takes to review one file from 20 hours to two, that still equals a savings of 18 hours. And that is just one file. As you consider using AI for your organization, focus first on the best results that your best people produce. Then work backward. Also remember that perfect should not be the enemy of the good.

Early forms of artificial intelligence (AI) have played a role in shaping our technological landscape since the mid-20th century, from Grace Hopper's early programming breakthroughs during World War II to the codebreaking efforts involving the Enigma machine. Innovations like ELIZA—an early natural language processing program in the late 1960s designed to simulate human conversation—paved the way for today's AI-powered tools. Over the decades, AI has been quietly integrated into everyday life, from generating entertainment content and powering virtual assistant chatbots in banking apps, to recommending shows based on our streaming habits. That quiet presence changed dramatically in 2023, when generative AI tools, like OpenAI's ChatGPT, disrupted the market and brought AI to the mainstream.

Alongside these advances comes a troubling counterpart: deepfakes, which are capable of creating hyper-realistic videos, audio, and images that can be weaponized to impersonate executives, manipulate markets, and erode public trust.

This article explores the cybersecurity and reputational risks posed by AI—particularly deepfakes—and considers whether existing insurance products are equipped to handle them. How will the response to generative AI incidents differ from those traditional cyber-related incidents? As generative AI technologies continue to advance and become more sophisticated—and adopted on a wide scale—insurance providers face the challenge of determining how AI risk should be treated within the scope of existing insurance products or if they warrant their own distinct insurance product.

The Threat of Deepfakes to Businesses

Deepfake threats can take many forms. While the types of threats discussed in this article are demonstrative, they are just a small sample of the possibilities AI opens to cybercriminals. Like "traditional" cybersecurity security threats, AI threats evolve hand-in-hand with the underlying technology.

Blackmail & Extortion: Threat actors could use deepfake videos to manipulate or blackmail a company. By creating fake footage of executives or key employees in compromising situations, cybercriminals can pressure organizations to comply with demands or face reputational damage.

Social Engineering: Imagine a deepfake impersonating a C-suite executive, authorizing fraudulent wire transfers, or gaining access to sensitive information. This scenario is no longer hypothetical. A notable case saw a finance worker at a multinational company tricked into paying out $25 million to fraudsters who used deepfake technology to pose as the company's CFO. The ability of deepfakes to mimic the voices, likeness, and even the mannerisms of company leaders make them a powerful tool for cybercriminals.

Market Manipulation: Competitors or even nation-states could deploy deepfakes to damage a company's reputation, manipulate stock prices, or disrupt public trust. Fake announcements, altered earnings reports, or fabricated speeches from top executives could quickly erode investor confidence, causing significant financial losses. And once information is out, even if false, it is hard to contain. For example, on April 7, 2025, a misleading tweet on X regarding President Donald Trump's tariff policy caused turmoil in the U.S. stock market.

Reputational Damage & Liability Exposure: While reputational harm was once a major concern in the early days of cybersecurity, evolving public perception has made such risks feel more commonplace—though that may change as sophisticated AI-driven deepfake attacks push the boundaries of what's believable and trustworthy. Deepfake attacks can cause significant reputational harm —especially for high-profile leaders of publicly traded organizations. A CEO's image and trustworthiness are critical for stock performance and investor confidence. Deepfake technology has the potential to erode that trust almost instantly. Even if the content is later proven to be fake, the damage to a company's public image can linger, and the financial impact can be substantial.

Beyond public image, these incidents may lead to allegations that company directors and officers violated fiduciary duties, such as inadequate financial reporting, or failure to implement prudent AI policies or safeguards. Professional liability exposure may arise if AI adversely affects the rendering or performance of professional services.

The creation of fake content—such as a deepfake video of an executive making damaging statements—could also lead to immediate loss of consumer trust, stock price volatility, and lasting damage to the brand. This kind of damage is not only hard to quantify but also harder to recover from in a traditional sense, as rebuilding reputation takes much longer than addressing technical fixes or financial losses.

How Should AI Risk Be Covered by Insurers?

AI-driven incidents present unique challenges that may not be fully addressed or appreciated by traditional insurance policies.

From a policy language perspective, defining what constitutes an "AI incident" could be difficult. While deepfakes are a clear example, AI is also being used in various other ways, such as in decision-making processes, automation, and data analysis. Will all AI-driven incidents fall under this coverage, or will they need to be explicitly defined?

Furthermore, the complexity of claims associated with AI incidents, such as fraud or misinformation, may require new expertise and claims handling processes. For example, it could be difficult to identify liability in a deepfake scenario—will the board of a publicly traded company be found at fault for failure to implement adequate AI safeguards if a deepfake impersonates a CEO and causes stock price drops thus negatively impacting investors?

These challenges have created a debate over whether AI-driven incidents are sufficiently covered under existing insurance products or whether an AI-specific insurance product should be created to address these risks.

There are two schools of thought on how to approach coverage:

1. Traditional Coverage Perspective: Some argue that AI risk does not inherently change the covered risk, but rather changes the magnitude of the risk. For instance, traditional cyber insurance generally covers the financial losses incurred by an insured arising out of a cybersecurity incident; be it business interruption, crisis management costs, reputational harm, or damages arising out of third-party liability claims or regulatory investigations. If a threat actor group uses AI to infiltrate an insured's system, and then deploys a ransomware attack, the use of AI does not change the covered risk (loss due to a network intrusion), but rather makes it easier for the network intrusion to take place. The same can be said about other lines of insurance whose insureds interact with AI. Therefore, AI risk should not be covered under a standalone insurance product, as it is sufficiently covered under existing products. Notwithstanding, carriers should actively consider AI risk in the underwriting process and amend pricing and modeling operations accordingly.

2. Standalone AI Coverage Perspective: Given the unique nature of AI-driven incidents, some argue that this risk should warrant its own stand-alone product. Traditional insurance products were not designed with AI in mind. This could lead to gaps in coverage for losses involving AI. There is also a rising trend of specific AI exclusions in existing products. Without a dedicated product, businesses may find themselves unprotected from AI risks.

While this is far from a settled matter, it will be interesting to see how the industry reacts and adapts to AI risk in the near future.

Final Reflections

The rise of AI-driven risks poses a significant challenge for businesses and insurers alike. Whether AI-driven risks are adequately covered under existing insurance products or whether they should have their own distinct coverage category is a nuanced debate that requires careful consideration of the risks involved.

On one hand, AI-specific coverage could offer more tailored protection for financial, reputational, and operational risks. On the other hand, integrating AI-related incidents into traditional coverages might offer businesses more streamlined protection.

Ultimately, insurers must stay ahead of the curve by adapting their policies, training claims teams, and rethinking risk modeling. Businesses, too, must reevaluate their coverage and internal controls to ensure they are not caught off guard by AI-driven incidents.

Employee Retirement Income Security Act (ERISA) lawsuits have experienced a fever pitch, with 136 new cases coming to light in 2024, a shocking 183% increase from the previous year. This trend seems likely to continue into 2025, as new ERISA-related lawsuits filed against Southwest Airlines and Charter Communications were brought forward. While these legal actions are becoming prominent, the ERISA legislation experienced a milestone, recently celebrating its 50th anniversary, further indicating the endurance of this law and its strong framework in protecting employee benefits and emphasizing the need for clear guidelines of fiduciary duties for those managing retirement plans.

2025 and beyond will be high stakes for employers and companies that are maintaining retirement plans for employees, otherwise known as plan sponsors. Risk mitigation and contingency planning for protection of individuals and companies are essential.

A new consideration for plan sponsors?

ERISA, major federal legislation that took effect Jan. 1, 1975, governs employee benefit plans of almost all types, and holds fiduciaries – as broadly defined – personally liable for plan administration and management. It has evolved over the years to ensure it is up to date with market changes and retirement planning requisites to better support employees. Inherently, it is a long and complex legislation with numerous nuances that contribute to plan sponsors' challenges in maintaining compliance. Recent class action litigation indicates that there are standout fiduciary areas where plan sponsors are struggling – including 401(k) plan forfeitures in defined benefit plans, pension risk transfers (PRTs) and health plans, as well as who bears administrative costs for these plans.

Most current lawsuits challenge how forfeiture in 401(k) plans is handled, and the outcome could have significant repercussions for the sponsor community. Typically, forfeiture funds are those contributions associated with employees leaving their jobs before fully vested in their employer's contributions to their 401(k) plans. Plan sponsors can use these forfeited funds to offset their contributions. However, many lawsuits argue that ERISA requires these funds to be used solely for plan expenses or redistributed to plan participants. While the verdict is still out on the court ruling, one thing is for sure – should the plaintiffs come out victorious, there could be a massive shift in forfeiture policies.

In the same vein, PRT-related cases are under the spotlight. PRTs have long been leveraged as a favorable strategy by employers to eliminate their pension obligations and associated risks. Plan sponsors have typically conducted transfers to an insurance company through an annuity purchase or a lump-sum buyout. Yet recent court cases indicate that the tides may be turning as plaintiffs have filed a handful of cases alleging that these annuities are too risky and thus fail to meet ERISA fiduciary requirements. This is another area that plan sponsors would be wise to watch, as the outcome could result in higher standards for PRTs.

Health plan litigation is another area of concern for plan sponsors. In 2024, class actions against health plans were all over the spectrum, from actions on health-based wellness programs to how plans choose to provide pharmacy benefits to their employees, particularly their choice of pharmacy benefit managers. Plan sponsors should be keen on keeping up with the effects as they can broadly affect their programs and require significant adjustments.

The impact of an ERISA-based lawsuit

Legal issues are never on the agenda for businesses, as they bring forward an onslaught of consequences, but ERISA-related lawsuits can play a particularly malignant role in an organization's continued growth and success.

Firstly, the short- and long-term financial strain can be debilitating. ERISA lawsuits incur mountains of legal costs – from attorney fees, settlements, and more – potentially reaching well into the seven-figure range. Additionally, under ERISA, plan sponsors may face personal liability for confirmed fiduciary breaches, potentially leading to civil penalties, removal of fiduciary status, or criminal prosecution.

Companies should also be wary of the reputational damage an ERISA lawsuit can cause. Stakeholder trust can be eroded following a lawsuit, making it challenging to hold onto and attract new investors. Similarly, talent attraction and retention are heavily affected. Employee benefits and retirement planning support are now expected by employees. If marked by an ERISA-related lawsuit, top talent may look for other organizations that meet their long-term financial wellness needs. By losing top talent, businesses will struggle to maintain and grow their business performance.

These examples of potential impacts underscore the importance of companies and plan sponsors effectively managing ERISA compliance and fiduciary responsibilities. The best way to mitigate these issues and their impact is to avoid falling victim to alleged breaches and staying alert about legal rulings. However, given the complexities and nuances of ERISA, it can be challenging to keep pace. Realistically, plan sponsors and businesses must be prepared to address potential issues from all angles.

The need for fiduciary liability insurance

Plan sponsors may be aware that bonds are required by ERISA; however, their protection is limited to fraud and dishonesty. For comprehensive risk management and to better navigate the growing trend of litigation, fiduciary liability insurance should be at the top of the list for fiduciaries and their organizations. Typically sold in increments of $1 million, this insurance offers valuable protection against allegations of improper judgment related to employee benefit plans, including, most importantly, covering legal defense and even settlements.

While it is understandable that concerns about costs exist, neither the mandatory ERISA bonding nor the optional fiduciary liability insurance should be seen as expensive. Considering the backdrop of regulatory fines and penalties from the Department of Labor for non-compliance and the increasing cost associated with defending against litigation, the cost of insurance is quite reasonable.

The protection from this coverage extends to the sponsoring organization, officers and directors, and plan fiduciaries. As ERISA holds individuals with discretionary authority over retirement plans personally liable for decisions that harm employee beneficiaries, fiduciary liability insurance provides essential protection.

Additionally, firms should enhance their compliance processes by leaning on technology-driven solutions to stay current with new ERISA provisions and automate wherever possible. Leveraging tools and platforms that can support tracking vesting schedules and contributions reduces human error and oversight, often the drivers of fiduciary breaches. Furthermore, digital-first solutions can support generating audit-ready reports if needed to demonstrate ERISA fiduciary duties are being met.

Navigating the future of fiduciary risk

The long-term success of ERISA demonstrates the effectiveness of its framework in protecting American workers' retirement plans. As new retirement trends emerge, market volatility increases, and regulations evolve, there will be a continued emphasis on risk mitigation and compliance. It is crucial for plan sponsors to stay updated and not fall behind in these areas. They must remain vigilant in managing funds in accordance with ERISA, especially as legal scrutiny intensifies.

However, due to the complexity of ERISA, it is not uncommon for gaps to arise. A significant aspect of risk mitigation involves preparing for the worst-case scenario, particularly in facing potential allegations of fiduciary breach. Robust defense plans should include solid fiduciary liability insurance, monitoring evolving regulatory frameworks, and updating/automating compliance practices. Only then can organizations and plan sponsors have the peace of mind to run excellent plans for the sole interest of participants and beneficiaries, which in turn benefits themselves and the organization.