Broker Buddha CEO, Jason Keck, is joined by Caribou Honig, Co-Founder of InsureTech Connect and Partner at SemperVirens. On this week's episode, Jason is celebrating a major business milestone. Broker Buddha has converted 10,000 carrier applications into online smart forms. Jason is joined by Caribou Honig, Co-Founder of InsureTech Connect and Parter at SemperVirens. Caribou is an influential member of the insuretech industry and someone who Jason has admiration for. The two discuss the value in conferences, the importance of structured data online and targeting niche opportunities.
Learn more about Caribou Honig
Jason Keck: Hi, and welcome everyone to this very special episode of The Enlightened Agent. The podcast that brings you conversations with top insurance professionals and industry leaders. My name is Jason Keck. I'm joined today by Caribou Honig, Co-Founder of InsureTech Connect, which is responsible for the largest insuretech event every year, ITC Vegas, as well as a Partner at SemperVirens, a boutique VC fund focused on the future of work. Caribou, welcome to the show!
Caribou Honig: Jason, thanks so much for having me I'm glad to be here.
JK: Yeah, I'm super psyched you're here. I remember going to my first InsureTech Connect in 2017 and being just completely blown away by it, so I'm a little starstruck by having you on the show, but pretty excited about it. I think anybody in the insurance industry probably knows who you are, you're a celebrity I think at this point. I know there are a lot of people who don't know everything about you, so would you mind taking a few minutes to maybe introduce yourself and talk about what keeps you busy?
CH:Yeah I'm happy to, this should be fun! You know, by the way you might be starstruck, but I keep telling my sons that I'm kind of a big deal and insuretech, and they're like, 'sure Dad, whatever' if they only knew. My background is pretty simple, I started my career at Capital One a couple of decades ago, really during its hayday hyper-growth. It was great and super training. I left there in 2006, took a little time off, watched Netflix, played with my kids, and listened to the universe. From there I reconnected with a couple other former Capital One executives, we created a boutique venture capital firm, QED Investors. We did a lot of investing in areas like fintech and some adtech. Around 2015 we started to focus on insuretech, and for those of us who were in and around insuretech at the time, we all know there wasn't much going on, but it smelled like something interesting was about to happen. I was really excited for that, and I looked for good industry conference but couldn't find anything. Thus InsureTech Connect thankfully was able to collaborate with Jay Weintraub on that. I left QED in 2017, I had some issues to scratch outside their scope. Now I'm kind of doing what I do. I'm very fortunate, I have my hands in a few different pies. I continue to work on the conference side of the things. It's actually, and don't tell anyone, it's actually kind of addictive. Once you create a tech conference you just keep doing it over and over. I get really excited when I can help people. I like to say my north star these days, and this sounds probably a little corny, but I really like helping good people to succeed at doing interesting things. It turns out, there's almost no better way than to bring a couple thousand people together and have them like actually interact about the right stuff. I've done that now on another conference focused on future of work, collaborated with Jay and team on another one on prop-tech. We did one just recently on logistics-tech actually have one coming up in June on Medicare of all things and maybe some ag-tech food-tech.
JK: You're a conference creation junkie, you can't get enough.
CH: I totally am. The good news is when you partner with someone who does all the work and knows everything, then you don't actually have to do much. That's kind of my relationship with Jay. He knows everything, he does all the work, has a team and I just get, you know, more than my fair share of credits. So it works out pretty nicely, it's a good deal.
JK: Good for you, man. I tell you this year in particular, getting out of the office getting into a space where people were focused on opening their minds, as opposed to being more focused on delivery and execution was game changing, right? I mean, mentally for me, it was super exciting. The feedback we got at the conference was super positive, opened a lot of doors, you know, couldn't walk through all of them, but certainly walked away with a lot of energy. I think conferences are pretty special like that, right? You get people together out of the office, they have permission to think beyond their four walls and, you know, magical things happen.
CH:It's way more efficient for everyone to jump on one airplane ride, all converge in Vegas, have the 20 meetings they want to, and then fly back then for you to have to Jason, fly to 20 different places to have those 20 meetings. I will say, it does kind of suck organizing mass gatherings in the middle of a pandemic, that's not easy.
JK: Do you have any good stories from this year about that, that you can share? Notable, or extra chaos that came out of it?
CH: It's really, it's almost a blur the last two years and ITC was relatively unscathed. We lost one show from an in-person perspective. Future Work Show, we actually lost two years of it in person, just from where the timing and cycle was. It's actually coming up in just a couple of weeks in mid mid-March, as long as there's no new variants, I think we'll be in great shape. I think it's more just there was a mix of extreme eagerness for people, but also a mix of hesitation at the same time. There was this sort of uncertainty like, okay, I'm here now, we're all vaccinated, can I take off my mask to talk to you or not? People were still getting a feel of the protocol back in October. I'm an optimist, I think we're mostly behind the major risk factors now. That's one area where I've got my hands, as you mentioned, I'm a partner at this boutique VC fund now called SemperVirens. I'm really focused on future work kind of investments, everything from like workforce, software, healthcare tech, because the employer is such a gatekeeper for so much of healthcare in the U.S. and then the sliver of FinTech and InsureTech, which has this overlap here. I get to do that then, otherwise, you just basically make a nuisance of myself. Do the occasional podcast, try to put forth the occasional, a hair brain idea.
JK: Love it. Love it. I really appreciate you having on the show, very impressed with all the different things you have your hands in. It sounds really fun and I imagine it pulls you in different directions all the time, which can be challenging right? To be successful, you have to focus. So, find your lanes and try to go as hard and fast into those as you can. This is a pretty special podcast for me because as you know, today, we're announcing as a company that we've hit a major business milestone. For those who don't know, Broker Buddha built a technology that converts carrier applications, static PDFs, into interactive, online smart forms that agents can use to gather information from their clients and produce signed documents for quoting and binding coverage. It's like TurboTax for commercial insurance. Today, we are announcing that we have converted 10,000 carrier applications into online smart forms. This is a huge milestone, 10,000 is a huge number. This matters because in order for agents to place any type of coverage, that isn't a bog standard, small business commercial policy, they need carrier apps. I can tell you that there are a few things more frustrating than trying to fill out information on a PDF using kind of antiquated PDF annotation tools and trying to click in and fill data in into spaces where it doesn't belong. It's just seems very 1999.
CH: We rehearsed this, but I want to interview you a little bit here. You know, does the 10,000 matter in terms of your product getting any smarter? Is it one of those things where there's a feedback loop so that, you know, it was dumb for the first hundred, then it was a little smarter for the next thousand, and by 10,000 it's actually become like, okay it's learned from error correction of mistakes in the first hundred.
JK: What matters is that the connection between the forms and the linking of information is very powerful. In other words, if you answer questions on one form and the same question exists on another form, the information automatically maps over, and over time, we're starting to build a global data model of questions beyond just traditional ACCORD questions, and over time we'll be spinning up kind of standardized applications for different lines of business, which consolidate questions across the different carrier apps. Yeah, they might want to change those over time, but you know, those are minor changes that we can make tweaks on. I think the power of the size of the library is just the convenience for the agent to know that when they start using the platform or anytime that use a platform, they'll know that there's a very high probability that they can just go online, type in the name of the form, and it'll be ready for them and ready to go. As opposed to having to dig up the latest version, attach it, even multiple versions to an email, to send to their clients and force their client into the doctor's office, like experience of lying out multiple forms and answering the same questions, which we want to try to avoid.
CH: I love the evolution of how the back office or the front office goes from paper, like this is just audio, right? So going from paper to PDF's, right? And then PDFs to actually structured data. There is real value in going from paper to a bunch of pixels, but then the next level is PDFs are pretty close to being unstructured data. If you can then turn it into structured data, that's sort of the unlock in my mind.
JK: There's a lot of companies out there who are doing PDF extraction, right? You're seeing all these tools, trying to extract information from PDFs. We've gone one step beyond that and we're trying to get rid of PDFs. We're actually extracting the questions, so that as people start to answer those questions, you know, yes, we can put it back on the PDF because that's historically how people need it, and sometimes they need signatures on that and we've got the integrated signature to support that. What this really does is it opens the door for us to be able to deliver that data to their management systems, to their carriers, to their data warehouses in a way that allows them to learn from the information and leverage the information. If they want to do benchmarking and they can do benchmarking. If they want to just bridge the data into a carrier system to speed up quoting, they can do that as well. A lot of power and having structured data online. I wish that in 2022, we weren't talking about migrating onto web forms, but we are, and that's okay. It's a great opportunity for us. I think candidly, it creates a lot of opportunity for others. Now that we have structured data, and I guess that's kind of one of the questions I wanted to flip back at you, which is, as the insurance industry starts to get a better handle on data, right? As agents are able to get structured data, as carriers are able to get structured data, how does that change the experience for the buyers of insurance? What does that mean for them? How does it make their life better? What can you imagine there?
CH: This takes me back to my first 10 years back at Capital One. We were one of the original, like really, how do you use data in the business strategy? Ultimately it comes down to a few areas. One is risk and risk based pricing. When you're underwriting, you are using data, I hope. When you're underwriting, you're using data and the more data you can get your hands on, the more available it is with lower costs, lower friction, and the more sort of structured and sort of robustified it is then the better that the underwriting can be. So for one thing, like, I think for the insured, it's making sure that they're actually getting priced at a level that is appropriate and commensurate with the risk, right? You don't really want people to be forced to cross subsidize someone else's risk. Right. And if I'm less risky, price me appropriately.
JK: I think that's one piece of it. I think that there's sort of a pure relevance aspect as well. If I'm a small business applying for getting insurance, and you've go me bucketed in the wrong SIC code category, that creates an extra 45 minutes of process where I'm trying to clarify that with you, but the actual right data is somewhere out there, you just didn't have it about me? That's actually really a hassle. So the more that you can ensure relevance, the better. Then I think there is just an efficiency side, like an operational efficiency, even a marketing efficiency, which accrues largely to the benefit of the insurer or of the agent. If I can be getting in front of the right customer, with the right product, then that creates operational efficiencies, maybe marketing efficiencies for me as the agent. In the long run, like efficiencies in the cost side of things, for the carrier or the agent, will eventually migrate at least in part to the benefit of the insured. If you can create efficiency through data, that's a good thing in the long run for the insured.
CH: That's right. At a minimum, it will speed things up and hopefully it also gives a better rating. The carrier is able to rate the risk faster and more accurately and potentially give you better pricing.
JK: There's not as much risk in their pricing. They can, they can sharpen their pencils a little bit.
CH: They may give you higher pricing, but only because you warrant it, not because they're making a mistake of cross subsidizing, right?
JK: They're not adding an extra 20% on top because of an unknown that they don't have. They're actually, you know, maybe you maybe deserve the extra 20% or maybe you don't. The thing that we keep hearing is that the holy grail for agents is the ability to link policy data, with claims data, with exposure data. There are frankly no great systems that support that today. You know, the policies stored in their management system, the claims data is often with the carrier and the exposure data is on PDFs. What we're encouraging, all the agents to do is really think through what does that benchmarking look like in a perfect world? What are the different ways they want to segment their customer base? Is it down to individual s.i.c. code or is it, they want to group industries in a certain way? What are the revenue bands look like? How do they want to segment their customers? Then, we can start to pull in policy data and the claims data in a way that gives them the ability to say, this customer, relative to the industry is spending too much. Whereas, their limits are too low and really give good advice on that.
CH: I'm reminded often when the statisticians were building big models, right? Predicting risk response, whatever. 90% of the time and effort actually went into basically assembling the data into a way that then the statistical tools could use. It wasn't really like, oh, there's a ton of like, uh, how do I do this? This is called analysis or machine learning. It ended up actually being mostly about how do I get the right data, just put into a normalized form. We'll go on a tangent here if you don't mind. If you're really geeky, one of the funniest examples of like not understanding what data you've got and then how do I make sure I'm building the model on the right data is, I'm building risk predictions off of FICO score, right? FICO score is your credit score kind of thing. Generally the FICO score runs 500 up to 850, 500 as a range, and I have no idea, but 500 is terrible. You've probably gone bankrupt. You're not paying your bills, whatever. 850 is like pristine. There's eight people in the U.S. who have it. There's a lot of data science behind, like what every 30 points means in terms of increase in your overall repayment risk and things like that. Terrific. When you're building a model, trying to find some correlation between, well, you had this kind of score, and did you pay back your credit card or your mortgage loan? Or, are you high risk on auto insurance, because there's a correlation there too. Well, it turns out that the credit bureaus also have a special number, if you have no file or some sort of thin file. I don't recall all the details. Something that would typically be correlated with very poor credit, very poor credit performance, not necessarily because you've done something bad, but you haven't shown up, you don't have the history.
JK: Like my wife who came in from England and has no credit or had no credit.
CH: So she would probably have a credit score initially of 900, because it's not a contiguous scale. 900 is like the designated code for no file. All these clever stats models get all screwed up when they're finding this very clean, natural relationship between 500, to 600, to 700, to 820 to 850. Oh, so pristine! 900, terrible!
JK: That's complicated. Although I kind of like it sort of suggested positive. We're going to start with a very positive number and then kind of work your way down.
CH: You've got these reports that like the average FICO score of this population here that was underperforming was actually 750 because it averages a bunch of 500's and a bunch of 900's. So there's all this energy that went into why this anecdote right? It's mildly interesting, I think, hopefully three of your listeners think it's mildly interesting. So much of the energy is around actually getting the data right, and the statistical sophisticated, hard to say, statistical analysis was the easy part. Getting the data in the right form for the models to be built off of, that was actually the energy.
JK: We actually have built some pretty powerful translation tools recently.Kind of an if this, then that type of tool. If you answered the question, if you had some question 1 like this, then we're going to map that to a slightly different category over here kind of thing. That translation capability is super powerful and allows us to automate some of the things that you're talking about there and data capture. So good to know that what we're building is solving problems that have existed for a long time, just never been solved in insurance. So kudos to us on that one. Caribou, as you know the show is called 'The Enlightened Agent', because we like to share stories about amazing people who do amazing things. Enlightenment is defined as 'the state of having knowledge or understanding'. I wanted to ask you if you know, any stories of any enlightened agents, or agencies out there who you feel need to be recognized.
CH: I appreciate that. Let me shine a light actually on the, I guess they're technically not an agent, they're a broker, I have a subtle distinction. Hopefully one that people won't mind, it's actually a company, a benefits broker, that we've got partnership with at my VC fund SemperVirens. It's called Sequoia benefits and they're doing two things that I think are really interesting right now. One is just in terms of defining their target market. They are very focused on venture backed, high growth, quote Silicon Valley companies. I say quote because it could be anywhere in the U.S. but Silicon Valley type of companies. I think that sort of recognizing that you find a niche, you find a population that might have some relatively distinctive needs, and then you lean into that. I actually think that's a trend happening right now in insuretech broadly is sort of a focus on niche opportunities. Where you know in insurance, even a niche can be sizable. Then when you really kill it on that niche, then you've got permission to go into adjacencies. Instead of going after giant things, people are starting to say, okay, let me solve for this niche brilliantly, and then expand from there. Sequoia is doing great in terms of we've identified a niche, we're going to go after venture backed tech companies. Then I've seen from a product standpoint, what they're doing, I think is quite insightful. Sometimes, simplification is the hardest thing to do. Taking away choices in a way that actually creates a value for the customer. If done poorly is a disaster, but if done well, creates actually a lot of value. So I've seen them start to create bundles of ancillary benefits. They're not saying, okay, client, you can only choose this or that product. But, we know you got a lot of other things going on in the benefits and HR world. You know, there's this COVID thing too. Let us pick some best in class products that sort of go together and gel together and can be priced as one bundle for you.You can choose bundle A or, bundle B or, bundle C, we can give you some benchmarking about what other companies like you, which bundles they are selecting. I think that the simplification, taking away the cognitive load, the effort required because does an HR executive at some company really wants to have to go through, and decide on which fertility benefits, and which smoking cessation benefit, and which disability benefit? No, they just want to know that they've got a best in class collection that sort of gels together. That's something that I would shine a spotlight on that they're doing.
JK: Those are two very powerful approaches and reminds me of Founder Shield, which is the company we went through the incubator with. They did the exact same thing, focused on tech companies, startups initially, now later stage tech companies and they had bundles. They had a bundle of insurance,this is PNC, not benefits, bundles for the startup. I think they had a bootstrapper, the startup and the venture or something like that. It was basically your pre-seed, your seed companies, and then you're kind of series B and beyond companies. It just frankly made it easier for someone to say, I don't know what I need.
CH: I fit into this category. If you brand the bundles properly, then you help people make choices around things that they don't understand. They don't need to understand it, their job is not to understand it. It's actually your job as the agent or the broker to understand it. It's their decision, but it's your recommendation.
JK: That's right, that's exactly right. I like that. So I liked the guys at Sequoia, we know them well and clients of ours, so I'm happy to be working with them on the PNC side. I'm not doing much in the benefits space these days, maybe one day in the future.I'm glad to hear that they've taken that approach and I've no doubt, especially being out in the bay area that they're crushing it so good for them. Frankly, good for the companies that they work with. So a great story, thank you for sharing that. This Caribou, has been what I would call an 'enlightening conversation', if you don't mind me going one step further on the pun. I really enjoyed having you on the show. Before we wrap up, anything else you'd like to say or share with our listeners?
CH: What's on my mind lately is just the notion of how to help people, and not look for a quid pro quo. There's a lot of transactional folks out in the world, but gosh, I find that there are also a lot of people who are not just relationship oriented, but they are really looking to be helpful to other people. Knowing that it'll all come back to them in karma, even if it doesn't come back in something direct. So I'd say just I'm trying to do my part, I'm trying to train myself to be like that. I hope your listeners are leaning into that as well, because I think that makes things a better place here.
JK: Yeah, a hundred percent. I love that concept of giving without expectation. Sharing, and hope that if I look at you, and the success you've had, if that's what you've been practicing, then sounds like it's come back to you in lots of great ways. Not that that was the reason, but you know,maybe it's serendipitous, maybe it's karma who knows what it is. Congrats on all your success. thanks for coming on the show, I appreciate you being here.
CH: I'm glad to be here, thanks for listening.
JK: I'll see you in Vegas later this year, is that part of the plan?
CH: That's right, in September.
JK: It's September, alright. Thanks, Caribou.