The InsureTechGeek 1: The InsureTech Geek Podcast with James Benham
On this week’s episode of the InsureTech Geek Podcast, episode number 1 of our new podcast series talking All Things Insurance Tech.
The InsureTech Geek Podcast powered by JBKnowledge is all about technology that is transforming and disrupting the insurance world. We will be interviewing guests and doing deep dives with our own research and development team in technology that we see changing the industry. We are taking you on a journey through insurance tech, so enjoy the ride and geek out!
Well, hello everybody. I am your InsureTech Geek. James Benham, here with JBKnowledge, here to talk to you about all things insurance tech. Now, I have been a podcaster for the last 4 years. This company, JBKnowledge, we have been in business for about 18 years. We work in two industries, construction, and insurance. I have been podcasting over in construction for 4 years, and I have been super eager to come over into the InsureTech space and talk about something I have worked in even longer than construction. That is insurance technology. It is an exciting industry. A lot is going on and it is hard to figure out how it impacts you.
I work with some of the bigger companies in the insurance space and some of the smaller companies in the insurance space, and it can be challenging for them to sort out all the tech that’s coming and sort out buzzwords from action items and to figure out what they should be looking at experimenting with and working on. So, the goal of this podcast is to get you up to speed. Now maybe you are already up to speed, maybe you are one of those InsureTech Geeks out there that is got this all figured out. Our goal is to entertain you immensely and connect you with some great new people that we are going to interview on this show. But hopefully, there is a bunch of you out there who has something to learn, and we are going to talk about some fascinating technology. We are going to cover stuff like predictive analytics, chatbots, and artificial intelligence, robotics, RPA, telematics, and wearables, portals, EDI, one of my favorite topics, blockchain. We are going to cover a bunch of different technology. Now, we are going to mix this up between discussions and interviews with people out in the insurance space, people in the insure tech space, and just as interesting, our research and development team.
If you do not know who we are, JBKnowledge is a 205 person, roughly insurance, a technology company headquartered out of Bryan College Station, Texas. And if you do not know where that is, it is home of the world’s finest Institute of higher education. Texas A & M, and you will hear me talk about that, probably to a level that will irritate people who are in the Southeastern Conference. Nonetheless, that is where we are headquartered. We have been working in the insurance business for about a decade and a half, and, started in the property inspection business, for underwriting inspections and then moved into the general property and casualty business. Then got over and into detailed into work comp and then jumped up and do the pharmacy and pharmacy benefits management space. And we have gotten to explore and build technology around a whole bunch of different sectors in the insurance space for the last decade and a half, and through it have seen an incredible amount of automation go into place.
And not just automation, but true innovation in the way that we are insuring things, people, companies. It has been transformative for the industry in the last 15 years and what is going on now. In the last 3 or 4 years, we have seen just an explosion in funding for InsureTech startups. We have seen an explosion in companies that are trying to bootstrap their way into this space. We have seen companies that are in traditional insurance companies starting their venture funds or dramatically expanding their venture funds in investing, starting their own company, so, there is just an incredible amount of innovation out there.
Let us just jump right into some of the topics that I think are of interest, and I think that you will find interest in as we go through this. And this will be a weekly show, as we go through interviewing people and talking about these different segments. Predictive analytics is arguably one of the hotter buzz phrases in the insurance space, and it can be a confusing topic to talk about. I meet with a lot of insurance executives and they will toss this phrase around, and I have found that very few of them could properly define it. So this is a field of practice that when executed properly, and we are going to interview some people here very shortly that I believe are setting it on fire in predictive analytics, cannot tell you what will happen in the future, but we will help you extract information from all these data sets we have been gathering to look at patterns and try to assign a probability, a better probability to future outcomes.
There is a phrase that we use pretty often, here at JBKnowledge. Give me facts, not feelings. Let us look at the facts. Let us look at the data behind it because gut instinct can lie to you. We built an interesting system, several years ago that benchmarked workers‘ compensation claims, and we were helping TPA’s benchmark their claim performance against a broader data universe. And that way they could figure out how good they were at lowering the cost of risk? Keeping claim costs down, getting workers back to work quicker. Of course, the bigger objectives, you want to make sure people get back and get healthy quickly and that they can do so in a cost-effective manner. And so, we were trying to figure out, the ultimate question almost everybody wants to know in InsureTech and that’s causality. What causes expensive claims in that particular case? What causes expensive work comp claims? And it was interesting as we dove into the data, how our assumptions about what caused claim cost to go up, were blown apart by the data once we started looking at correlation, and I am not going to ruin the surprise.
We are going to talk about this later in a different episode, but once you start looking at all the data points you can possibly collect, and then you start looking at the correlation between those data points. And then, of course, correlation does not equal causality. So, you got to figure out, is there just a correlation here? For example, you look at a data set and you would see that there is a bunch of people who eat red apples and it turns out a lot of those people also like to scratch their head. Eating red apples does not necessarily mean they scratch their head. This might be a correlation that might not be causality. As we dove into these data sets, we found a lot of correlation in some causality that challenged our assumptions about what drove work comp claims up, and that is where we are going to dive with predictive analytics.
Let us try and look at what’s truly predictive analytics and what is a system that someone is just taking what they have in their head, and systematizing it into a piece of software, where it’s human assumptions, not data-driven machine arrived at assumptions. So that is what we are going to look at with predictive analytics. There are some neat things that you can do around the space, and it is an area of a lot of innovation right now. We have got a great interview on that one coming up. Cause we are trying to move away from linear, traditional models and you know, move into a completely new way of looking at data and segmenting our data.
Chatbots and AI. This is another area that there is a lot of discussions around in some people in the insurance space are actively putting this into practice. I have seen chatbots being effectively used. I have also been to many insurance and insurance technology conferences where I have been told that someone has a chatbot, and then what I am presented with, is a text interface where you pick menu options. And that kind of reminds me just a little bit of personal background for those of you who have never met me. I am 40, I was born in 1979. A great year to be born. I turned, 11 and 12, and then got my first computer and started writing code. And I started writing software in a GW basic. That was the very first programming language I ever wrote software in. And then I wrote stuff in Turo, Pascal, and Fortran, and C, C++, and Assembly. And when we first started writing code a long time ago, we had text interfaces where it was a linear text interface and it would present you with options and you would type the number that you wanted.
So, when I see chatbots that are functionally what I used to build when I was 12 out of GW basic, it is not a chatbot. A true chatbot is the ones that have strong natural language processing capabilities where you can have a conversation with it, almost as if you are talking with a human, and then it can respond and perform multiple actions. So, this is what I look for in my consumer tech too. I am a big Google home fan. I have got all these different systems. Of course, I have Alexa as well. I have got Google home; I have got all these different systems. I have put controls in my house. And one of the things I particularly like about Google home is that I can speak to it rather naturally and then combine commands. I do not have to separate commands into separate line items. I can issue multiple orders in one sentence and it will carry them out.
And so that is what I am looking for when I look at chatbots and we are going to talk about chatbots. We have built some interesting chatbots, both independently through text messaging and a lot of the text message chatbots that I see on the market are interesting, but they are not exactly innovative. It is like, here are your options, pick an option, move onto the next option. It is almost faster going to a dedicated mobile app or a website at that point. But we are seeing some folks in the insurance space and including some sophisticated carriers that are implementing chatbots, to answer questions about policies and to file claims, and give them a natural language interface.
There is some cool tech around voice chatbots. I have been following Google duplex pretty closely, and if you have not checked it out, it was almost a year ago, it has been a while ago by Google as a natural language voice interface to do whatever you want. They are driving both to use their assistant. And it was interesting. Before I came down here to record this, I was reading a new story that restaurants that are getting phone calls from Google duplex are starting to pick up on it and ignore them or hang up. They find it creepy. Their response was not that it was bad. Their response and why they hung up on these chatbots was that it was too good. It creeped them out that the robot was as good at speaking as a human would be. And this has been interesting.
There has been some interesting human backlash, so we are going to talk about that when we get the chatbots. The human side of this, how will people feel about talking to or chatting with robots, and there was some backlash the day after the Google duplex announcement when it first came out from people saying that it was too realistic because Google included little nuance like aha, which convinced the listener. And so, they have been identifying it as Google. The only reason they know it is Google is cause itself identifies and then they get angry because it is too human–like. To will be honest, I would not have pegged that as being a big response from a lot of people. But there you go. So, we are going to talk about the human side. We are also going look at the technology behind chatbots, both voice and texts, and where it can be used and where it is unnecessary.
This is something that I say a lot, both in the company and outside that you can efficiently suck, and a chatbot could go to legitimately help you efficiently suck. You could be the worst interaction for your customer or your employees. Artificial intelligence and chat, you almost have to talk about those in a similar breath. True general AI does not necessarily exist yet. There are predictions that it will not exist for another 30 years or so. Who knows? It depends on who you talk to and how they box in general AI. If you are an old school sci-fi like me, and you saw 2001: A Space Odyssey, that was pretty close to general AI. You have a general use voice interface to talk and it can think and reason and make decisions. Of course, it went poorly in a little part of that movie. And there are other movies about that. It means science fiction likes to explore both the utopian and dystopian side of potential technology. AI though, when we speak about it, and this is an overused phrase right now in InsureTech, I am seeing a lot of people talk about AI and only a few people using components or specific forms of AI, like NLP, natural language processing, image recognition. They are using specific components.
So, we are going to be interviewing and talking to some people that are using this in real life. And another great Sci-Fi reference is Soylent green. Soylent green was made with people. A lot of AI is made of people. They are people that purport to be AI systems. I remember I signed up for, oh, I do not remember what it was. It was an AI assistant. I do not remember the name of it right now cause it has been a couple of years. But I found out after using this AI assistant, and sometimes it would be kind of laggy like it would take an hour for it to respond. To me when it was setting appointments and doing things for me. I would largely use this AI assistant for scheduling. And I found out that a significant portion of the interaction was being handled by human beings in the Philippines or somewhere who were chatting through this robotic interface. And so, it was AI’ish and they were trying to train it and get better, but it just was not game time yet.
But you are seeing some specific form of AI that can perform tasks that normally require human intelligence, like visual perception, decision making. Language translations have gotten way better. Last year, Microsoft announced they had a peered a human translator for English Chinese, and so they had been able to produce a language translation bot that was able to rival language translation. Now, this is a big deal. First off, let us just talk about language translation. There are entire companies, and I run into him in insurance conferences, whose entire business model is translating both voice and text for insurance claims. So that is a big win. Also, think about just the untapped, unstructured data and insurance that is out there. Video cam footage, webcam footage, security cam footage, photos, videos, free text reports. There are just terabytes and terabytes, and of course, ultimately petabytes and exabytes of data that is never used, that is never accessed, that is never properly put into a system where you can search through it and utilize it. So that is what we are going to talk about when we get to AI. And it is exciting.
And then you kind of jump into robotics and a lot of times we think of robotics, we will think of Boston Dynamics and the creepy dog robot. If you do not know him, from the other podcast I do. Jim is our technical producer here, and he hates the robot dog because it is got like a snakehead on it and it creeps them out every time. And so, he always cringes every time I talk about the robot dog with a snakehead. But that’s Boston dynamics. That is an actual physical machine robot that can make its own decisions, navigate 3–D map an area. You can think for insurance claims it is going to be brilliant to be able to send a robot into a potentially dangerous building that might be unstable, and you are trying to do a condition assessment for a claim. You are going to have the option to use robots, in particular, autonomous robots that can decide to walk around and scan the building and come back out in the report to you.
But a lot of times the most common robot we are seeing is software RPA. That is robotic process automation. It is a form of clerical process automation. It is when you are using software robots to perform tasks, and so you will train it, you will tell it what to do and repeat it over and over and over again. It extracts data out of systems. This is good for extracting data out of old legacy systems that do not have a great API because let us be honest if you had an application programmers interface you have got a way to directly integrate with a software program. You would not use an RPA bot. I mean, that is silly. If you do not understand any of the words coming out of my mouth right now, I recognize I just geeked out a little bit. But RPA is a common topic and what I am hearing, senior–level C level executives and insurance companies talk about RPA regularly because they recognize that it can eliminate all this mundane, menial work that is such a drag on productivity and morale. It allows them to elevate human workers.
We use a system at JB knowledge called EOS Entrepreneurial operating system. It is a way of running a business, not a software product. It is a methodology. And they have a tool they call the delegate and elevate chart. Whenever I think of RPA, I think of delegate and elevate. You are picking the things you do not like doing or hate doing, and you are not getting good at or you are bad at and you are saying, look, I want to delegate all this menial stuff, like down to all this repetitive task work, down to an RPA bot. You can use UI path, or you use like Selenium for testing, where you can repeat the same process over and over and over and over again, so a human being does not have to, and it can be effective and consistent, and it can make a big difference.
We are seeing our clients and others in the space have a huge impact. So, we are going to talk to some RPA folks, we are going to bring some of our RPA folks in, and we are going to talk about robotics and RPA because there’s a space in both areas to have a pretty substandard discussion. Both in hardware robots that you can use for underwriting and claims, and you can use them in all kinds of loss control. There are all kinds of great ways you can use robots, but also software RPA, neat stuff coming down.
Telematics and wearables. This is common. You know, it is interesting. We saw this hit the consumer market probably most heavily in the auto insurance space when they started rolling out these OBD two–port. If you did not know that little port under your steering wheel is called your OBD two–port. We saw them roll out dongles for the OBD two–port and then those dongles have self-cards that report your driving behavior. It was kind of the beginning of mass consumerization of telematics. And then, of course, the Apple Watch and the Samsung gear. And of course, the Apple Watch is still in the lead on market share. It has changed the game of people wearing computers. And so, you are looking at personal wearables and then telematics for vehicular, road transport, electrical engineering, sensors, wireless comm, computer science. There is a lot to telematics. And it phases broad, but when I am talking about telematics, I am talking about connecting mechanical and electrical objects and then buildings up to sensor networks and giving the ability to read data off and potentially even influence settings and things on that device.
There are interesting consequences to that. There are insurance companies that have rolled out and said, hey, if you use this OBD two–port device for a period of 3 or 4 months, we can measure your driving habits, just putting it in there and being monitored to get 10% savings. And if you are a good driver, you can get more than that. That is kind of a carrot and a stick. It is not much of a stick, but carrot and the carrot methodology to get people to allow themselves to be tracked. But let us be honest, how many consumers legitimately want their insurance company to know everywhere they go in their car and how they drive? Because a lot of people do not drive like angels and they do not want to know that they gas the vehicle all the time and they have been lying about how many miles they drive on all their insurance applications.
So, we are going to talk about the implications, both social implications and technology, and telematics and wearables. There are much more attractive wearables now that do amazing stuff, from, detecting your heart rate to looking at your skin temperature, to your emotional activity. It is neat. You can also see if you are a fib now on the new, Apple watch. That is pretty cool. So, we are going to talk about telematics and wearables, and we are going to deep dive into how we think this could be a game–changer and help us to stop guessing so much on what these insureds are doing and then start making more calculated decisions. So interesting stuff on telematics and wearables.
Also going to talk about exchanging data. This has been something that I have spent the last decade and a half doing a whole lot of, and that is just exchanging data between all the different players in the insurance space. Believe it or not, everybody has a lot of proprietary databases or they have their big enterprise databases and they do not have simple, easy to understand API’s. They do not have a lot of documentation. I remember going to one insurance company and heard them talking about the series 900 file and the series 1200 file and it ended up being, that was the bite length of the file and it was a fixed delimited width text file, which is a very, very, very old way of storing and transmitting files. It is very error–prone and, that those systems still exist. CSV transfers, a lot of people in the insurance space that connect applications by exporting into Excel or CSV or Lord help us even flat fixed with files and then transfer those files around.
And so, this is in a time when insurance carriers and insurance commissioners and owners and insureds are getting much savvier and they want a lot more data. So, we are seeing an order of magnitude more data flowing between all the different players, whether it’s brokers or carriers, or their claims administrators, there’s just a lot more information flowing back and forth because everybody needs to make better decisions and they’re starting to get analytical tools that let them make better decisions if they have data. And of course, there is this big thing in the insurance business for takeover data. When you change from one carrier to another and you are trying to transfer historical information. It is complicated. So, we are going to talk about how people are doing this with portals. We are about what is old school EDI and what is new, what’s new school.
And then of course you cannot talk about this without talking about blockchain. That is right. Blockchain got to be discussed when you are talking about this, and there are some big implications for blockchain. Blockchain is the underlying technology behind Bitcoin. We are not going to talk about Bitcoin. There was just another Bitcoin heist as well if you did not know that. I am waiting for a movie to be made about Bitcoin heist. Wonder if they are going to call it a BitHeist. Either way. Blockchain is a fascinating concept. It is fascinating and execution as delivering real value in a lot of industries. Decentralized, distributed, public, immutable, digital ledger. It could be big, or it could be bad, and I think there is a lot more good in this than bad. What it is going to most likely expose is people who have been covering up specific aspects and transactions and do not want activity posted to a public blockchain. And so, there is a lot of room for improvement here, and there are some things to do with blockchain and I am excited about it.
Also going to talk about, one of my favorite topics, drones. I am a commercial drone pilot. Love flying, drones. I am also a private pilot. Love flying airplanes. I just love flying things. Love being flying in things like flying things. They are incredibly enjoyable, and they deliver a lot of value. And drones do too. Drones are already being heavily utilized in the insurance sector and the construction sector. And in the industrial sector, there is a lot of industries that are jumping on board in particular because the capabilities are so incredible, and the prices are so low. So, we will be talking about drones, we are going to interview some folks in the drone space, and some of my old friends there, we have been testing drone software for years, and so we will talk about some of our favorite apps and how they can impact constructions about drone deploy. We will talk about sky cats, we will talk about, Pix 4-D. We will go through all those. And there are some neat consequences and implications there for what they do in this space.
More than any of this, I want all of you to think about something that Peter Diamandis calls abundant thinking. It is an abundant future. Because it impacts this whole podcast, it impacts everything that I look at. It impacts how I approach and look at the world and technology and innovation. There is a significant number of people whose livelihood depends on pushing the negativity agenda, whether it is a mass media outlets, like the constant negative news network, or whether it is online sites or whether it is people that have formed nonprofits that are there to fearmonger and drive, donations into whatever causes they are trying to scare you into giving money to. There is a lot of people who would have you believe that we have a scarce future. That if you win, they must lose. They believe in a zero-sum game, and I just do not buy that theology. I do not buy that methodology of living. I believe, as does Peter Diamandis, that we have a very abundant future, that this is arguably the best time, statistically speaking, to be alive and longest average life expectancy. The best quality of life, the least number of working hours, ever, per week. The most vacation time humans have ever had, the most disposable income, the lowest percentage of people that are in extreme poverty. Literacy rates up to 90%. I mean, there is just so much to look at and go, wow, technology and innovation have had a massive impact on the world and life is significantly better.
And so, as you are listening to this and you are looking at the technology and topics we talk about, I want you to think abundantly, I want you to think abundantly because abundance thinking and looking at how good everything is, how much better it is getting, shapes how you innovate in your own company and your life. So, I have some abundant thinking, we will be talking about that too. It is an important topic. The way you approach innovation, the way you approach change, impacts everything that you do in your life, whether it is work or play or anything else. So that is what we are going to be covering in this podcast. I hope you will continue to join me. I hope you will hang out with us. We are going to have some great, great interviewees. I have already lined up a few that are coming up, in the coming weeks that you can go and check out. We are going to have some great topical discussions with our research team. I kind of laid out a lot of the different topics we are going to touch on over the next year.
The InsureTech Geek Podcast is going to be all about, helping you understand what is going on in the space. Helping you speak intelligently about it and, and hopefully connect you with people and players that are doing real, real, real innovation work. Not marketing, lip service innovation, but the innovation that counts. So that is all we have got for this week. I appreciate you listening in. Thanks for joining me. Do not forget the InsureTech Geek Podcast InsureTech Geek Podcast powered by JBKnowledge. That’s JBknowledge.com is all about technology that is transforming and disrupting the insurance world. I have been your host, James Benham, that is Jamesbenham.com. Please join us next week as we interview more guests and go into more deep dives in the technology, we see changing the insurance industry.
Taking you on a journey through insurance tech. So, enjoy the ride and geek out. See you next week.