Insurance Blog | Accenture

InsureTech Connect’s Caribou Honig explains how insurtech is upending the customer experience; how data is enabling new risks, businesses and operating models—and why insurers should aspire to be dentists, not toothbrushes.


  • Caribou suggests that carriers should be more like dentists than toothbrushes—that is, focusing on making the customer relationship as painless as possible, rather than trying to architect more opportunities for engagement. 
  • Channel partnerships that bundle insurance with other purchases—such as travel insurance with purchase of a plane ticket—can be a smart play to improve the customer experience. 
  • Digital-first insurance distribution may enable uninsured or underinsured populations to “break through” actuarial tables. For example, telematics may help a good driver in a high-risk demographic demonstrate how safely they drive.

Vanguards of insurance 

Welcome back to season two of the Accenture Insurance Influencers podcast. This season, we’ve spoken to Scott Walchek about how micropivots have contributed to Trov’s success, and Ruth Foxe Blader about why Anthemis invests in financial products for a digital future 

Trends in insurance, with Caribou Honig  

In our first interview with Caribou Honig, we looked at the current state of insurtech and what to expect at InsureTech Connect 2019. In this episode, we dig into today’s insurtech trendsseparate reality from hype and discuss the social implications of digital-first insurance products.

The following transcript has been edited for length and clarity. 

Welcome back to the Accenture Insurance Influencers podcast, where we look at how the industry’s vanguards aren’t just adapting to change, but creating a dynamic future for themselves. I’m your host, Eagranie Yuh, and today my guest is Caribou Honig.  

Welcome to the podcast, Caribou.  

Thank you, very glad to be here. 

Historically, insurers haven’t engaged that much with their customers beyond point of sale and claim, if something happens. And it seems like a lot of players are trying to change that. I understand that you have a different view of things, which you explain in terms of dentists and toothbrushes. Can you share this with our listeners? 

Yeah, sure. This is probably the best metaphor I’ve come up with in the last couple of years, maybe my whole career.  

I think there’s a little bit of bank envy for insurance companies. They see this level of engagement with the banking product that looks a lot like a toothbrush. You use it twice day. When I wake up, I check my credit card balance, and just before I go to sleep, I check my checking account to make sure I don’t have any overdrafts. Okay… good. I can sleep well tonight. It starts over the next morning. And that is a very high level of engagement, for reasons that I think are pretty intuitive.  

You see a lot of insurance companies––and start-ups as well––asking: Can we become like that toothbrush? After all, if the banks can, we can too, right? And it’s not clear to me that that’s the right aspiration. It may actually be that the right way for insurance, as a product and as an industry, is much more like the dentist visit.  

Dentist visits are very valuable, but very different from your daily use of the toothbrush. You go in once every six months, you sit in the chair. It may not be all that comfortable for an hour, but you get done what’s done, you hope there’s no cavities. And then you set your appointment for six months later.  

It’s relatively low engagement, slightly uncomfortable, but it gets the job done. And it may be that that’s actually much more the right model for insurance, that “Yeah, you go do a quick checkup, make sure you’re covered appropriately, hope you don’t have any cavities—I mean claims.” And set an appointment to check again in six months and then pay your bill. That’s fine. 

It’s not as satisfying as the notion of the high-engagement toothbrush, but maybe the real answer is that the industry should aspire to be a great dentist visit, rather than a toothbrush.  

Now, I will say the exception to thatif there’s going to be a toothbrush model for insurance, my instinct is it’s going to come out of the Internet of Things type of strategies. Think about the Apple Watch for monitoring your health for both life and health insurance. Or a water monitor embedded in the pipes in your house. Every night it sends you a little message: “your pipes are fine.” That actually seems relevant. 

It may actually be the smart toothbrush, ironically enough, that proves to be an example of insurance actually being more like a toothbrush. There’s at least one start-up I know that offers a smart toothbrush as part of their dental insurance package. That’s taking it literally, but I guess that would work.

I’ve heard others say something similar. For example, telematics is very popular right now and insurers love it because you get all this data, you can personalize the premiums, you can better understand your customer.  

But at the same time, Ive also heard a lot of customers say, “I dont want my insurer telling me how to drive. I dont want them to say, hey you shouldnt be braking so hard.” It feels like an imposition, and so I wonder how widely that applies to this toothbrush approach.  

Do people really want their insurer to be involved in their lives, or as you say, is there some balance between the toothbrush and the dentist? 

I think that’s right. That’s exactly right. 

So let’s follow this argument to the next level. If insurers should be looking more at the dentist experience, what is an insurerour analogy is getting a bit muddy nowneed to do to be a really gooddentist? 

What makes for a good dentist? One is that they’re thorough. Another is that they try to minimize the discomfort along the way. And when they do find a cavity, they try to help you navigate your options and take care of it quickly.  

You call up your dentist and say, “I’ve got a toothache, really bad toothache.” You don’t want to hear your dentist say, “Well, let me send someone over to take a look at it in a week and a half and they’ll take a few pictures of it and then we’ll get around another week or so to figure out how we want to deal with it.” (And I’m trying, of course, to create a slightly absurd toothache adjustment process there to really kill our metaphor.) You want the dentist to say, “Sorry that your tooth is hurting so bad. We’ll find a way to work you in today. Come on in.” 

Or—I’m going to go really off the rails heremaybe what the dentist should say is, “You’ve got a camera in your hand right now, try to take a couple of pictures of it for me. Text it over to me and I can give you, in the next five minutes, a quick perspective about whether this is something that needs to be seen today, or whether to get you something for the pain but it can be dealt with in three days.” That is the broader theme I like, around starting to outsource some of the activity in a self-service way to customers themselves. 

And you’re starting to see this, maybe not so much in actual dentistry, but you are starting to see this in parts of insurance, especially around the claims process or the data entry process. The Internet is great at enabling companies to source some of the work needed for the product or the application or the claim, to actually go to the consumer. But in a way which actually makes the consumer happier, because they can do it more on their terms, on their timeline. You know, I think that’s pretty interesting and I think that’s where the world is going here.

One thing that strikes me in the description of our insurer/dentist, is that there is this increased focus on the customer experienceI think a lot of these insurtechs really have the customer experience at the core of their offering, and see that as something that is less prominent with the traditional insurance experience. 

Why is customer experience so important and why is it such a central feature of the insurtech world? 

I think of two main areas for it. One is the claims process, and the other is the positive selection and value proposition aspect of the customer experience. 

First, the claims process. Where does customer churn come from? I think it mostly comes from the claim side. (I suppose the other place it also comes from is the, “Hey, you can save 15 percent in 15 minutes,” so there’s a pricing thing.) But I think a lot of customer churn ends up being the result of dissatisfaction with the claims process—the claims experience. So thinking about how to make that claims experience fairtransparent and easy for the consumer should reap real dividends in terms of lower churn. And there is certainly the example that we seem to keep coming back to: the 15-second claims experience.  

This is one of the things I love about parametric insurance modelsyou can structure it so there is no such thing as a claims adjuster or a dispute, or claims resolution. The threshold is triggered by, “Oh, you have a drought as measured by the weather satellites on this acre. Therefore, three days later, you’re going to see this money as the parametric crop insurance in your bank account.” There’s no claims experience other than receiving money in your bank account. I think that’s a pretty good claims experience. I really think that focusing on the claims customer experience—structurally, not just the handholding aspect of ithas power for customer retention. 

The other area is if you think of the customer experience wrapped into product design. Not just thinking about what color palette you choose for your mobile app, but the value proposition. What is the product you’re offering and how are you structuring the exchange of value? 

I think about Health IQ as an interesting example. They’re the folks who follow me around the Internet showing me the ads: “If you can run a six-minute mile,” which I can’t, by the way, “then we’re the life insurance company for you.” That’s a customer experience wrapped much more in the depth of the value proposition than the color palette. And that’s one which I think is clearly pointed at driving positive selection and avoiding adverse selection.  

I think rolling the customer experience into channel partnerships is another example, right? I’m seeing a lot of activity where start-upsin particular, are building out some novel product and then embedding it into the right point of someone else’s customer experience. 

Think about travel insurance; this has been around for a while. You’re buying your ticket on Expedia and that’s the moment when an insurance company will offer you, in partnership with the channel partner Expedia, the insurance product. That’s actually very smart thinking about the customer experience.

One argument that I’ve heard from the incumbent side is that bundling—such as the Expedia and travel insurance partnership—can be a really smart customer experience play. I think there’s a fear on the incumbent side that it renders the insurer invisible. You see insurance getting bundled with luxury car purchases, or car sharing and these kinds of things. I think there’s a fear that insurance starts to become invisible, thus exacerbating this customer experience problem. Does that make sense? 

It does, but I think it comes down to a question of, “where is your source of structural advantage? If your source of structural advantage as an insurer is your brand—brand awareness and the likethen that’s a big problem in this context. This shift is a threat to that kind of competitive advantage.  

On the other hand, if your competitive advantage is, for instance, we have the best technology stack and APIs for seamlessly integrating our product into channel partnerships, then a world that moves to that embedded insurance model actually favors that insurance provider, because by definition they have an advantage in that integration. I think it can cut both ways. 

A similar capability that cuts both ways is data. So I think that incumbents have the benefit of tons of historical data, of really good understanding of historical risk, and we’re coming up to a point where maybe that’s not sufficient, or there are different ways of doing it (for example, dynamic pricing based on Internet of Things data)Can you talk about how data is enabling this next generation of insurance technology?  

Again, I think in some ways it comes down to the claims side. Claims is where data is coming in most easily, most readily. And I think that it’s a simpler environment there, especially compared to underwriting 

You referenced telematics earlier—it’s really hard actually. It’s a long process to incorporate telematics data into automotive underwriting. Not that there isn’t value in it, but it’s a different regulatory regime you’ve got to manage. You may need to rebuild a whole bunch of models in order to incorporate it into underwriting, invite some regulatory scrutiny. It may not actually be culturally aligned with any sort of traditional actuarial science. There is, in practice, a distinction between actuarial science and data science, even if there shouldn’t be.  

But if you bring some data to look for fraud in the claims side, it’s just a lot easier. You can bet the learning cycle is much quicker to prove or disprove that some set of data is actually incremental. I think that the regulatory hurdles are a lot less on bringing new data or new data methodologies when you’re looking at the claims part of the process. 

Can we talk briefly about blockchain? It seems like it’s either the next Internet…or completely not the next Internet. What are you seeing? 

I think blockchain is a very interesting kind of technology. It’s interesting in part because ledgers are everywhere; ledgers are very commonplace, and so any new, really distinctive technology for maintaining a ledger is interesting. There are distributed ledgers of other things. People use Google Docs all the time. 

I’ve got a magnet on my refrigerator and it’s actually a shared ledger. It’s called the grocery list for the Honig family. If you want to add something, you add to it. And if you want to cross something out, you can cross it out. It’s not distributed. It’s centralizedon one refrigerator.  

There are some benefits of blockchain. But I don’t think it’s as magical as the image some of the advocates want to create. It’s simply got different tradeoffs than other types of technologies for maintaining a ledger.  

When I think about, “OK, when would a distributed ledger be the right answer, when would blockchain be the best?” I think about two criteria, which both must be true. If they’re both true then blockchain is really, really interesting 

  • You have multiple parties maintaining their own ledgers of some common information. 
  • Those partners periodically spend time reconciling their distinct ledgers of that information.  

So if you’ve got 10 different parties, all recording the same basic info on their own company ledgers and they have to come together once a quarter and spend weeks trying to reconcile them, then that’s a great place to look into having some sort of (probably private) distributed ledger, some sort of blockchain.  

On the other hand, if you’ve got 10 companies that are maintaining their own ledger, but they never feel like they have to come together and reconcile it, then I don’t know why some sort of distributed version of that ledger is better than a centralized approach like you’re doing today.  

So I just think it needs a real framework approach to understand when it’s better and when it’s not as good. What are the tradeoffs? What are the pros and cons, compared to maintaining a sticky note on the refrigerator or a Google Doc or an Oracle database?

Interesting. Another theme that comes up in InsureTech Connect is this reminder of the human side of insurance. I can’t remember your exact wording, but it was basically, “it’s a technology conference, but at the end of the day insurance has an effect on people.” And I’m curious why you feel compelled to mention that. 

I feel compelled to mention it because it is so easy at an industry conference—and particularly one that is focused on what’s coming and technology and innovationto lose sight of the big picture, to lose sight of the why that goes beyond just the selfish economic part of the why. 

Look, I spent nearly a decade at Capital One, in an industry that should exist. Credit cards are actually quite useful and increasing access to credit is quite useful. Capital One does that. But like a quarter pounder or the Whopper (whichever comes from McDonald’s), there can be too much of a good thing. So you get phrases like “drowning in debt” or “being up to my eyeballs in debt” or “being on a debt treadmill.” I guess over the years there, I came to get a little bit more personal maturity, thinking about the implications of those things.  

One of the things I like about the insurance industry and getting to play my little role in it, is that you don’t have people talking about, “I’m up to my eyeballs in insurance” or “I’m on some sort of insurance treadmill; I just can’t get off.” If anything, it’s much more, “How do we get more people to get the level of protection that’s in their own self-interest?”  

There is also, back on the claim side, making sure that claims are being adjudicated fairly and effectively for people on all sides. So you do get an adversarial aspect of that between the customer and the insurance company. 

But insurance is actually a really valuable thing and for all the nerdy excitement I get about APIs, and moving insurance to the cloud and blah blah blah, I also get a real personal satisfaction, and I think everyone involved should get a real personal satisfaction over the good that this industry does for people. 

When I was preparing for our conversation I asked a coworker of mine who has attended ITC (because I have not) and that was one of her observations: that as much as it was a tech conference, she was surprised—happily surprisedto see a lot of discussion of how these new products were helping get insurance to people who were un– or underinsured, and how there was quite a human drive to a lot of these companies. 

A lot of technology innovation, a lot of “classic” Silicon Valley is around product innovation and design. And when you’re doing good design work and good product development work, you actually need to have the human front and center of your mind. 

I like to say you want to be building for a focus group of one, and just make sure you’re serving that one ridiculously well. And then you hope that there’s other people like that one person you’re trying to serve––but really make sure you’re serving some needs of some person. The more you can actually name them, the better—not abstractly, but “Oh yeah, this person I know would get value from this product or from this design.”

There’s an interesting post on the Atlassian blog talking about ethical design and some of the unintended consequences of something that may start off as fairly benign, or as something that aims to do social good.  

So I wanted to talk a little bit about some of those implications, particularly because it’s insurance. Do we run a risk if all these companies are creating digital-first products? Does that, perhaps, favor a more privileged customer? And if you take that to an almost (hopefully) absurd consequence, do you end up with this bifurcation of higher-risk people who are not able to make this choice between a traditional product and a newer productand these preferred risks, who by virtue of their privilege, have been able to choose something that offers them a better experience and a better product?  

Is that a possibility and is there a responsibility of insurtech as a whole to make sure that doesn’t happen? 

Well that’s a small and easy question to answer… I may be slightly equipped in that I was a philosophy minor in college, we might rely on that a little bit. I try to give a little thought to these kinds of things.  

But look, I’m no expert in ethics. Not by a long shot. Tools are just that, they’re tools. And how they get used is the consequence of many, many different stakeholders and parties and entities in our society. Which is to say as a backdrop, the range of outcomes is very wide and the range of interested parties is very wide as well.  

Now when I try to take this to the question of social implications of insurtech, I find it useful to split that between the underwriting side and the marketing side. 

I actually find the underwriting side to be a little easier to get my head around. I think as a society we say that we’re not going to price, or make available, a product based on something that you can’t change about yourself. So what falls easily under that, for instance, is ethnicity. Gender is another one.  

Genetic factors are another example. I can’t really change my genetics. Now that we can read a human genome for so much cheaper than we could five years ago, you still can’t use it for employer-based health insurance.  

We also add, there’s some things which you could change about yourself, but which you should never have to. Think religion. 

As a general principle, those things which you can’t or shouldn’t have to change about yourself, we’re not going to underwrite on. But those things which are your choices, which are behavioral—like do you slam on the brakes hard, or do you smoke, or do you take your diabetes medicineswe are much more willing to price against. 

Now there’s exceptions to these things. In the US, as I understand it, we still allow, at the federal level, for genetics to be used for life insurance underwriting. And we make a distinction of, okay, well there are some places where it’s more pernicious, and other places where it’s more acceptable. 

Gender is actually a really interesting one. Developed economies have come to different conclusions about when and where gender can be used for underwriting of auto insurance. In Europe it is prohibited, point blank. I believe most states—not allin the US it is permitted. And that’s a really interesting set of social policy choices. On the one hand, we like the notion of gender neutrality. On the other hand, when you enforce gender neutrality for auto insurance, it means that in practice, a lot of women are going to be subsidizing the insurance premiums of men, on average. 

I like telematics because it allows underwriting based on something that’s behavioral. If you drive aggressively, in a way that creates more risk, you can start to make choices to drive less aggressively. And so I think about the person who may be actually a very safe driver, but has demographics or other data, like credit data, which puts them into a more at-risk population. But by letting them prove through telematics how safely they drive, they can break through the actuarial data tables. I think that’s actually very empowering for someone.  

The marketing side is a little bit more subtle, and in some ways is a little bit cleaner for me. The benchmark can’t be absolute uniformity of availability, because that’s kind of a false benchmark. I think of the benchmark as “What’s the availability of it today?” and “Is making insurance available through your smartphone going to improve that for a consumer?”  

I personally live in a comparatively economically poor area. It’s probably a banking desertI can’t say for sure that it’s an insurance desert, but I suspect so. And I also believe that penetration of smartphonesmaybe not the latest iPhone, but a perfectly capable smartphoneis pretty high in my neighborhood.  

So I think, and I hope, that the availability of digital-first insurance distribution and insurance products will go a long way to leveling the playing field. This is a complex, real question. 

I’m glad you’re more of an optimist than I am, Caribou. I think that’s probably why you were the VC and…well, there are a number of other reasons why I’m not the VC, but you’re definitely more optimistic than I am. So thank you for that.  

Well, let’s revisit in five years and see who’s the better report on that. We’ll see how it goes. 

Today we’ve been speaking with Caribou Honig, the chairman and co-founder of InsureTech Connect. Thank you very much for taking the time to speak to me, Caribou. 

It’s been an absolute pleasure.


  • The insurance claims process is a prime candidate for reinvention, whether through improvements to the claims experience, providing more self-service options for customers, or using new and novel forms of data. 
  • Incumbent carriers should examine their source of structural advantage, especially as embedded finance shifts the focus away from brand awareness and toward customer experience and technology capabilities.  
  • Many insurtech products have potential to provide insurance to populations that are currently uninsured or underinsured. 

For more guidance on insurtech: 

In our next episode, we’ll reconnect with Ruth Foxe Blader of Anthemis to talk about why scale is so integral for insurtechs right now. And after that, we’ll come back to Caribou Honig to talk about how incumbents and insurtechs can foster a more resilient and agile industry—together.  

In the interim, you can catch up with earlier episodes of the podcast. 

What to do next: 

Contact us if you’d like to be a guest on the Insurance Influencers podcast.

Submit a Comment

Your email address will not be published. Required fields are marked *