Shawn Hanson and Nick Gregory are attorneys in the San Francisco office of Akin Gump Strauss Hauer & Feld LLP. Their practices both focus on insurance, and for a sector not widely known as an innovation hotbed, there is a tremendous amount going on in what is called InsurTech. Since 2011, InsurTech companies have raised $5.67 billion across 464 deals. Established insurers, back-of-the-envelope startups, regulators and other industry players are all looking for innovative ways to handle risk. The interview has been edited for length and style.
New York City is emerging as the Silicon Valley of the rapidly growing InsurTech sector. Put this in context for us. Why is a city not known as a traditional haven for startups proving so attractive to InsurTech innovators, despite what is widely seen as the tough regulatory environment?
Nick Gregory: First, it’s important to remember that New York is a hotbed for startup activity. Venture capital investment in the United States is concentrated in a few big metro areas, and while it’s true that San Francisco and San Jose invest more venture capital into their startup ecosystems than anywhere else, the New York-Boston corridor is right behind with a significant amount of venture capital investment as well. Second, New York is at the center of the insurance world. Large, established players like AIG, MetLife and Marsh call New York home, and you’re in close proximity to Hartford and Newark, which are also insurance hotbeds. New York happens to be at this unique intersection of the established insurance world and the startup world. That’s why we’re seeing this kind of InsurTech activity in New York.
As for why people would be choosing a regulatory environment that might be perceived as tough, we’ve talked to people about that exact situation. While some entrepreneurs view that as a potential roadblock, others view it as a potential benefit – you get a framework for what those regulatory issues are going to be, and it makes expansion into other regulatory environments a little easier.
Shawn Hanson: There are two related points. First, one of the big challenges here for InsurTech companies is the patchwork nature of insurance regulation across 50 states. This is oversimplifying, but, because New York is, to a large degree, the center of the domestic insurance world, there’s a “if you can make it there, you can make it anywhere” approach to regulation. If you’re compliant in New York, there’s a reasonably good chance you’re compliant in most of the other states.
Second, another unique feature of an InsurTech startup is that because this industry is so regulated across the 50 states, many with fairly antiquated regulations, a lot of companies are bringing in seasoned insurance people capable of working with people from the startup world. New York presents a great talent pool for people with long, deep experience with insurance and insurance regulation, and how a business has to operate. There’s a trick to getting those very different cultures – startup and experienced insurance folks – working together. New York has a great talent pool for what these companies need at an earlier stage than you would expect for most startups.
A Bay Area company we work with, Embroker, is a good example. Even before their series A round, Embroker had a seasoned insurance professional as its COO. The CEO, while very tech-savvy, also had some experience with the brokerage business already, by virtue of a private equity background. Embroker also brought in a GC and director of compliance very early who similarly had long-term insurance experience. New York is a hotbed for similar talent.
Finally, one thing that’s different about the InsurTech business is that, more often than in most businesses, there are strategic investors. Not the smart, money venture capital people, but insurance companies coming in as early-stage investors. You probably saw that with XL Catlin, Munich and some of the other companies. A lot of those companies are very New York-driven. The strategic investors playing a major role in InsurTech right now are often located in New York.
Many observers say the regulation of insurance needs to evolve quickly to keep pace with InsurTech development. There are some positive signs, such as the creation of the Innovation and Technology Task Force by NAIC, which is designed to stay on top of new products and services from startup companies. Where does regulation need to go given long range societal trends, such as sharing information and leveraging data?
Gregory: Regulation has to adapt at pretty much every level. You can see how regulation is going to be an issue when you’re developing insurance products. If you’re developing new business models, like peer-to-peer insurance models, when you’re filing your new rates, or even when you’re adjusting claims, you can imagine that an entrepreneur developing code in the garage may be able to price risk more effectively than established insurers. But there are all sorts of regulations that entrepreneur’s team may not be aware of, such as the factors you are allowed to consider when pricing risk and the steps you need to take to get your rate formulas approved by different states. That can be a major issue for someone who is very tech-savvy but not familiar with the regulatory environment.
Hand-in-hand with that – and this is something we’ve talked to the NAIC about – is the idea of using artificial intelligence or AI. AI has incredible potential to change the way that we analyze risk, but as it gets better and better, it’s also harder for consumers and regulators to understand. A phrase I’ve been hearing is, “making intelligence intelligible.” How do we make sure that the intelligence we’re using is intelligible to humans – including regulators?
Here’s an example: one promising application using big data and AI is that your car may someday have a sensor that detects when you brake hard, or you accelerate too quickly, or you speed. Your insurance company may get that data in real time. It may give you a smooth driving score that you can look at and then try to improve. In fact, you might be able to compete with your friends and family about who has the best smooth driving score. By the same token, if you get a letter from your insurance company that says, “Your rates just went up,” you ask why and they say, “Because the algorithm said so.” You can imagine the customer confusion that will create.
Hanson: The answer to some of this – and the NAIC seems to be thinking about this – is called a “sandboxing” approach, which offers a way to work with regulators to permit private companies to experiment with new programs and concepts within defined confines. Within certain sandbox boundaries, a company can go to regulators and say, “This is the innovation I’m proposing. Within these boundaries, let me experiment and see how it goes.” Some regulators have already signed off on similar programs for FinTech. This sandbox approach is somewhat implicated by the NAIC’s Big Data Working Group, which is talking about an approach to reviewing complex models at the NAIC level. That obviously isn’t going to take away the 50 states’ ability to regulate, but it’s an approach that offers the potential for real innovation within the current regulatory environment.
At a more macro level, insurance regulation is going to have to recognize the modern social norms of sharing and leveraging data. That is going to be a challenge.
Gregory: Another interesting aspect of this is the extent to which this is going to be controlled by state-by-state regulation and the extent to which companies can contract with consumers to provide for some of these innovations. For example, regarding getting a lower auto rate, perhaps companies will be able to contract with their consumer beforehand and say, “This is the data we’re gathering and using, this is our software, and here’s a rough idea of the formula we’ll use to set your rates.” That way you can get out in front of that issue – the interplay between what’s going to be regulated from on high and what can be done on a contractual level with the consumer. It’s an interesting topic.
Hanson: Another facet that is gaining traction is what’s being called “RegTech” - or innovative solutions to help companies automate regulatory compliance. That means using technical innovations to enhance, regularize and speed up the compliance process. A really exciting aspect of RegTech is the possibility of regulators partnering with new RegTech projects to help this efficiency effort. One of our local Bay Area accelerators is having a RegTech conference soon. There’s a chance for the regulators themselves to become innovators, in a sense, by using some of the same approaches to their supervisory obligations. That’s only starting to bubble to the surface. It’ll be interesting to see how that plays out.
Gregory: One final point on this. There’s a lot of room for startups to be proactive in engaging with regulators – to start thinking about their business models, thinking about the kind of regulations that might be implicated by innovative new models, getting together with regulators or working with third parties to work with regulators to vet some of this stuff.
Given the complexity of the U.S. system, will insurance regulation evolve more quickly in jurisdictions such as Singapore, the UK and Dubai, where the relaxation of some traditional rules is leading to experimentation in sandboxes and innovation hubs?
Hanson: That is already happening. But that’s not the complete answer. For some of these innovative approaches, even with a significant amount of your operations offshore, and a significant amount of your operations basically done through AI or machine learning, or automated in some fashion, if you’re going to adjust an insurance claim, you still have to have an insurance adjuster who’s licensed to do that work. You can’t move it – at least it seems to me that there are some real impediments to doing that, both for underwriting and claims assessment.
In dealing with consumers, if your underwriting and marketing process is completely automated, you still have to have a licensed agent or a licensed broker transacting that business in the United States to some degree. Even if some of the innovation takes place in these other hubs, and it plainly has and will, if you want to do business in this market, you’re still going to have to deal with the 50-state regulatory structure. It’s likely to be a little bit of both, and that’s not even accounting for all the other consumer regulatory statues that aren’t insurance-specific.
Gregory: It’s interesting to consider what the regulatory sandbox idea would look like in the United States. Because of the McCarran-Ferguson Act, almost all regulation of the insurance industry is left to the states. If we do get a sandbox in, say, Iowa or Connecticut, how helpful is that if Texas and California and New York don’t adopt the same kind of regulatory framework? Does it make your business models more efficient, or is it one more headache? That’s just one potential obstacle that you face when you’re trying to get a regulatory sandbox up and running here.
One analyst suggests that, as important as InsurTech is to driving operational improvement for the insurance industry, the true sustainable advantage will come in breakthroughs in engagement innovation, including the development of customer-centric products, value-added services and new business models. Tell us about the new products and business models currently emerging, and where you see the most promise.
Gregory: Customer engagement is absolutely crucial, and there’s a lot of room for improvement. We actually see a lot of companies that are trying to improve customer engagement. One of the issues with insurance generally is it’s not perceived as a glamorous or fun product. It might be fun to order glasses online, try them at home, and send them back. It might be fun to get food sent to your door. But if you look at figures about the amount of time the average person spends thinking about insurance in a year, it’s unbelievably low. People don’t like parsing insurance policies, or comparing policies online. Trying to get customers more engaged in the process is an issue that people are trying to solve.
A lot of tools that are coming out are trying to change the game on that score by, for example, allowing customers to see, in real time, what they’re paying for, or describing in simpler language what a policy gives them. Credit card companies have been fairly successful in getting people to become credit card gurus. Engaging customers like that, to think more about how to get the best insurance deal or value and giving them the tools to make it a user-friendly experience, is going to be a major opportunity for innovation.
Hanson: There are insurance regulations that affect the marketing you can do. It will be interesting to see how, as companies try to make the user experience more interesting, that’s going to track with those regulations, which are fairly formulaic – even rigid – about how you market to consumers. There are hoops you have to jump through to sell an insurance product in most states. They will have to be accounted for.
That’s a business challenge for the creative people, and there are a lot of them out there. It’s also going to be a challenge for their in-house lawyers and compliance people. It’s possible regulators will see that these regulations don’t really match up, and maybe we’ll see some regulatory reform once the startups hit critical mass and have the ability to go as a group to some of the regulators and the NAIC and say, “Here’s what we’re thinking. Can we make some adjustments?” Hopefully that will be coming soon.
Gregory: Here’s one example people in the industry see all the time: simplifying policy language. Policy language is often tedious, and much of it appears – to a layperson – to be overwritten, confusing or irrelevant. So making policies more user-friendly for the consumer is a great idea that will make the products easier to use. On the flip side, industry veterans can tell you the reason every single one of those apparently unnecessary words or provisions are in the policy. Often, that’s a direct result of decades of case law and courts interpreting specific policy language, which means another add-on, which means another case, which means another add-on. On the one hand, you have fresh eyes that might be able to bring a new and unique and simplified perspective. On the other hand, you have eyes that have been looking at these policies for a while and know where the language came from. There’s a tension there.
There’s a lot of experimentation around blockchain technology, which some see as highly promising in the insurance sector. For example, Everledger, the digital vault of the future, has used blockchain to provide a ledger for diamond ownership and transaction history verification. What impact do you see blockchain having on the insurance industry? Are there any regulatory or other barriers to widespread adoption?
Hanson: There’s no doubt that blockchain is with us right now. Something that just caught my eye is that Ernst & Young and Microsoft, working with one of the big marine insurance companies, just implemented a blockchain-based marine insurance platform. At the proof of concept stage they were able to show real efficiencies. So it’s not coming; it’s here now.
I don’t see a lot of insurance-driven barriers to the use of blockchain. In fact, I think some of the more forward-thinking regulators see blockchain as a possibility for simpler, more efficient and much quicker supervision of the industry. In addition to protecting consumers, a high priority for most insurance regulators is regulatory solvency – that carriers doing business in their states are capable of responding to their contractual obligations. That generates an enormous amount of regulatory work, including collecting what is often basic financial information and quickly understanding it.
Blockchain potentially can simplify the process of solvency supervision and make it more reliable. Blockchain sets the data, and to some degree the operation, in stone (though the Stone Age may be the wrong metaphor). It makes the data more reliable and accessible. Once everyone figures out how to do that, which is going to be in the RegTech space, blockchain offers some real opportunities for regulatory enhancement, not barriers.
Gregory: What blockchain can give you is the potential to reduce the amount of litigation surrounding disputes over things like chain of custody. When you have a blockchain that everyone buys into and agrees is an objective record, the opportunity to reduce those kinds of disputes is immense.
Hanson: That’s going to apply in spades in the reinsurance world. I’m not sure how it’s going to work out in the classic, consumer-facing business of selling and adjudicating claims. I’m sure it has real opportunities there. But I think it will soon affect some of the more complicated and sophisticated programs and relationships among insurance companies, and relationships between insurance companies and their regulators, both domestically and in other jurisdictions. I think we’ll see real impact right away.
The recent hurricanes sparked a lot of discussion about the use of unmanned aerial vehicles, and advanced data analytics, to reduce losses. Is this just talk, or do you see real reduction of risk for the industry through the use of these and other technologies going forward?
Hanson: There’s no doubt that it’s the real thing. Here’s the hidden piece of the drone story, the part you don’t hear about as much in the popular press. Drones can significantly enhance underwriting. At a conference recently, I heard an interesting story about crop insurance. Rather than having someone inspect the crops as part of the underwriting process, they were able to use drone technology to recognize and analyze the visual data to materially tighten up the underwriting process in terms of expense and time.
Gregory: You can imagine that in other contexts as well. If you need to look at roof conditions, that’s the perfect application for a drone. It can collect large amounts of data that would take humans significantly more time to complete.
Hanson: Frankly, you can see it for simple auto claims. Rather than a consumer having to take a vehicle to an adjuster, the adjuster can do a visual inspection by using this kind of technology. It’s plainly the real thing.
Gregory: That does tie back into the regulatory points we were talking about earlier, and what regulators need to do to make sure they’re staying apace with the changes we’re seeing. Each state has its own regulations that govern claims handling, claims adjustments, whether or not a claims handling practice is reasonable. If AI gets good enough that it is reliably producing damage calculations or valuations, is that going to be seen as a “reasonable” adjustment process? Those are going to be the kind of challenges we will see moving forward.
The Zenefits matter in the state of Washington, and developments in Louisiana, have highlighted the significant differences in how state insurance commissioners view anti-rebate and inducement laws. Please explain the importance of the issue of unlawful rebates, what the state of play is today, and where you see things going.
Gregory: The purpose of these anti-rebate statutes is to prevent kickbacks from brokers to customers. The question now is if I’m a company and I’m giving free software to people, is that the kind of thing that’s supposed to fit into what I’m going to call the anti-kickback statute? If you read the Louisiana advisory letter on this, it’s pretty clear that that’s not the kind of thing that would fit into that more antiquated, anti-kickback statute.
Hanson: The rebating issue is the classic example of regulatory language that was developed in one kind of economy under one set of circumstances and that is worded so broadly that it’s actually now become somewhat counterproductive. It seems to me, honestly, that with the exception of Washington State, you see the Louisiana approach becoming the touchstone for regulators looking at these issues.
In many respects, this isn’t anti-consumer, it’s pro-consumer. Zenefits has had its hearing before an administrative law judge, and we’ll probably be seeing a ruling pretty quickly. I would imagine they did a good job of explaining to the administrative law judge that their style of doing business has many benefits to consumers, and isn’t in any way a detraction for consumers who need insurance. I’m hopeful that Zenefits will come out as a positive story. Overall, I think the Louisiana approach seems to be the touchstone these days. I’m sure there will be wrinkles here and there, especially as things get more complicated. But if you can explain effectively how what you’re doing helps the consumer, or helps the insured, regulators are willing to listen to that.
Gregory: One of the ways in which Louisiana is a great template through which to view this issue is because it sets forth in detail when a broker is providing something of value, and it’s offered on equal terms to the entire general public, that you can’t claim it’s an inducement to the contract because you’re giving it away irrespective of whether the person is entering into the contract. As Shawn said, the trend seems to be that that’s a more reasonable framework moving forward.
As in so many areas, the changes advanced by InsurTech innovation carry major privacy and data security implications. Tell us what your clients are seeing and what, if any, impact cyber rules enacted by the New York Department of Financial Services will have.
Hanson: As these companies raise more money, get through their series A and series B financings, and get bigger, they trigger all these rules. Everyone is very focused, in the case of bigger companies, on compliance with New York rules. I think as a practical matter, this is a good example of what we talked about earlier: New York, even though it can be considered a tougher regulatory environment, is the center. The general thought is that if you can meet the New York rules, you’re likely to be in compliance with whatever else is coming down the pike from any other state.
It sometimes gets lost, but a lot of these companies have HIPPA issues, too. There’s so much focus on New York cyber, but you have to be thinking about HIPPA at the same time, and trying to make those two regulatory frameworks work together.
Gregory: Let me echo Shawn’s point about smaller companies. They typically aren’t focused on things like cyber compliance, because they’re focused on raising funds, developing a product, marketing their product, getting a team together. Getting a lawyer is not always the first thing that comes to mind, let alone to start really focusing on compliance.
What this New York regulation does – the first state regulation of its kind – is to shorten the runway. Smaller companies need to be thinking about it a little earlier to make sure that when they launch their product, they’re covered and not at risk of some of the security breaches that we’ve been seeing.
Hanson: The message is out there. These days, when I talk to pre-series A companies, if I don’t bring it up, they bring it up. Those rules have penetrated to the point that even pre-series A companies are at least thinking about them.
A recent survey showed a lot of interest in InsurTech innovation, but also underlying concern about working with untested new players. In your practice, how do you see the traditional insurers and the emerging players working together?
Hanson: One thing that’s interesting about InsurTech is the number of what I call strategic investors, but let’s just call them what they are: insurance companies that are investing in InsurTech companies. Not only do they bring their money, but they bring their knowledge, their experience, their expertise in the industry. I think that that’s one way.
InsurTech has also benefited from the various accelerators. At least with those that we deal with, there’s a real opportunity for big strategic investors, both venture capitalists and insurance companies, to be around the brand new companies. They can talk to each other, learn each other’s language, see each other’s perspectives. It really helps. Plug and Play is one of the startup accelerators that provides that, and there are others both in the Bay Area and in New York.
There’s a third way, which I think they learned more from the FinTech experience than from the InsurTech experience. A number of the investors, when they size up these investments, one of the things they size up is the management team. And one thing they take into account is how much real, practical insurance expertise there is on the management team. Then you see series A companies, and even some pre-series A companies having real experienced insurance people, both in operations and compliance and elsewhere, on their executive teams. Then the regulators and the NAIC efforts go hand-in-glove with that.
Gregory: One concept that you run into a lot in the startup world is this idea of moving fast and breaking things. That’s how you innovate. What both investors and startups themselves understand is that the playing field is a little bit different in the insurance world, because you’re dealing with a highly regulated industry. That means striking the right balance between getting your ducks in a row and rushing forward. From what we’ve seen, people are putting together teams that address that concern, or at least trying to.
What about artificial intelligence? Does AI have a role to play?
Hanson: Everyone in this business is talking about machine learning and artificial intelligence. This probably reaches beyond the insurance business, but the regulatory environment for that, in the InsurTech space, is very exciting. It’s going to have its challenges, but that’s what in the next year plus is going to be the most interesting: how to make AI work with consumers, how to make AI work with telematics, how to make AI and machine learning work with regulators.
Gregory: At what point do we unleash AI? Do we let AI do its thing and calculate risk with little human involvement? That’s an interesting question, and that’s obviously not just for the insurance industry. You see it in all contexts. For example, there are software companies that use Big Data to make parole decisions.
Hanson: Nick always reminds me of this. It’s not machine learning versus the old underwriter/claims adjuster model, and how those two work. But it’s interesting and may be relevant. I’ll let Nick give the chess example.
Gregory: Everyone is aware that a computer designed to play chess has (for quite a while now) been able to beat the best grand masters in the world, reliably. That said, what chess grandmaster Garry Kasparov reminded us recently is that the best chess-playing computer assisted by a human grand master can apparently still beat the best chess-playing computer on its own. You wonder if there’s going to be a stage, and how long this stage will last, where we have the benefit of artificial intelligence, but it’s closely monitored and assisted by humans who understand it and can complement it And I think that all ties into the question of how we can understand, explain and assist AI in the best way possible.
Hanson: That goes back to the talent pool in New York that we talked about earlier. You almost need an insurance grand master to work with your machine learning approach. And finding those people and seeing how that plays out, and how long it plays out for, will be very, very interesting.
Shawn Hanson is a partner in Akin Gump Strauss Hauer & Feld LLP's litigation practice in San Francisco. He leads the firm’s InsurTech practice and also focuses on commercial, false claims and regulatory litigation involving health care and life and health insurance. He also has an active practice involving all aspects of ERISA, as well as private equity, partnership and similar corporate disputes. He can be reached at firstname.lastname@example.org.
Nick Gregory is a member of Akin Gump Strauss Hauer & Feld LLP's litigation practice in San Francisco. He is a member of the firm’s InsurTech practice. Mr. Gregory focuses on complex commercial litigation, with an emphasis in the areas of insurance, class action defense, and regulatory litigation. He can be reached at email@example.com.
Published October 25, 2017.