What I see is a technology driven mindset to adopt automation at key points in the actuarial process, with a keen eye for quality assurance and making sure that proper opinions are made by the actual actuary and not the software. Something that I see in a lot of the innovative actuaries that I talk with is a very keen mathematical curiosity that's a little bit deeper than the average actuary.
I was really intrigued when I first started my career back 11 years ago. After about a year in an actuarial analyst type role and being exposed to a number of actuaries, it was interesting to see how little innovation there was and little drive for innovation. There was a lot of very traditionalist mindset, staying in your box and not being very open minded to technology and advancements, even in probability theory and distributed computing. It was an intriguing experience because when I was going through college, I got into actuarial science because everyone was telling me “if you really like math and you're good at programming and you like solving interesting problems, then becoming an actuary is the right path for you.” So I was pretty excited about that. I had been in the engineering world in college for about three years, and I was getting bored of that after a couple of internships. I thought, “this isn't quite what I'm looking for.” So when I first ran into that after about a year into my career as an actuary analyst, it really made me think, “well, what can I do to be more innovative, bring some more of the modern technologies to the forefront, the more advanced analytical capabilities.”
I'm from the life insurance side, which is a lot less innovative than the casualty property P&C side. Looking at them as an example of how they progress to at least somewhat more innovative from a process and methodology standpoint, they've been much more innovative in the last 20 to 30 years than the life insurance side. Something that I ended up doing was taking a break from the insurance world to go work at some innovative startups in the non insurance world. One was in the architecture or construction industry for smart building architecture. Another one was for precision ag technology, collecting data, harnessing tons of data per day and at various other places learning what was working from a technology perspective and what wasn't working, what other people were doing with their data, their analysis, what kind of software tools they were using. There were a whole lot of things I was trying to figure out that other industries were doing so that eventually I could come back to the insurance world and be like, “Hey, this is what I found out.” That's pretty much what I'm doing now.
I think about that drive for using technology in a way to help us drive our processes more efficiently. I think the statistic they always say right now is 80% of our projects are data intensive preparation, pre-processing, gluing all the points together and the remaining 20% is the actual studying of the outputs and making your opinions. So as much of that 80% that we can automate, make more efficient, it may still be some processing time, but it can at least be done in parallel with other tasks, maybe more effectively.
One of the key influences we have as actuaries is that we tend to have direct access to top leadership, top executives, even the board of an insurance company. Being able to be aware and technically understanding of different technologies that are out there is important. That, combined with our ability to assess risk and measure it accurately and mathematically, is very important for us to be able to convey in effective and business layman terms how technology can help the business meet their goals, their vision, and get things done in a more effective way. We have an interesting influence in the company, especially an insurance company where we're the engines of the company in terms of making sure the company remains profitable and that we have the reserves up to par. There's a lot of weight on our shoulders, making sure those companies are staying afloat. To me, that’s one of the top influences we have.
Actuaries also tend to have an ability to be a bridge between the more technical people in a company, namely the software developers or the database administrators, and the business executives because we understand the data and the practices that maintain a good dataset. In the right circumstances, an actuary can be quite a good fit for that position in the company to make communication more efficient. The bigger the company, the more important it is to have a position that's a bridge or a doorway to take the technical information and convey it to leadership.
There’s data scientists, data engineers and data analysts. An actuary that has an understanding of all three of those, plus the in-depth actuarial understanding of the business, can be a real powerful force in terms of influence in the company.
One of the main things that is going to become more stable in the actuarial community is the use of open source actuarial toolkits being more acceptable and mainstream. I know at least one major actuarial firm that has people who, as part of their regular job, are contributing to open source actuarial software. People are starting to be more open to that. I've been seeing it a lot on LinkedIn and other places where actuaries are starting to contribute to the software community. I've been doing a little bit for the last six years. Six years ago the only other person I could think of at the time was just dabbling in stuff. It's come a long way since then and people are starting to use it more or at least starting to contribute.
As the younger actuaries are coming into the scene they have the foundational software background from university programs that are including software development as part of their curriculum in a more in depth way. As more of those people come into play, the more people are going to be contributing to the open source community. They're going to want an open system to be able to make the changes that they want to the model that they see are necessary, especially with some of the new regulations. It's a natural progression for people to start using systems that have more flexibility. In the long run, the more people that are using it, the more robust it becomes. There is an initial barrier there with less usage and having reliability but just like with everything, there's a transition period.
Something else that I see in the near future of the actuarial world is the need for a software credential, something an actuary can show. I think we’re moving towards having some kind of a credential, showing software understanding and usage and professionalism especially as open source becomes more used. The last thing I'm thinking of for the future of the actuarial world is an actuarial digital assistant. It's been Alexa and Siri and Google's assistant. They're useful in some ways but there is this big wave right now of software engineers and entrepreneurs coming out of big companies like Amazon and Google that are leaving to start their own digital assistant companies that are more useful in a specific domain. Instead of you asking an assistant to schedule something for you or some basic task, these digital assistants are going to be doing more side-by-side work with you. They will do tasks in parallel with you at the same time, helping you do what you’ve got to do. I think a digital assistant is going to be an eventual thing helping actuaries form opinions and analyze insurance contracts. Kind of like an actuary in a box. I did a talk at the Actuarial Research Conference 2020 about the concept of the artificial actuary, which is like this actuarial digital assistant, bringing up interesting questions like “would it be useful or even practical or maybe necessary to have those assistants pass a certain test or have a credential of their own, like an AI having credential? What kind of ethical questions come up with that?” There's going to be some very interesting conversations coming up in the next five to ten years.
If you're still in college or even if you're not still in college, be very curious about software and how to use it to automate tasks. Take a 10x software engineer, that some people claim doesn't exist. What makes them exist is that it's not that they're manually creating ten times as much work, a 10x employee has basically ten versions of themself doing work at the same time because they’re automating as much as possible. He or she gets more work done in the same amount of time simply because of automated tasks.
"I tend to be one that promotes humans; we're hired to be creative, we solve problems that require abstract thinking. Let's let the computers do the non creative, boring tasks."
The best example I can think of is if you’ve got to send emails out to a group of people and they need to be customized emails, you can do it all by hand, or you can make your own email system automation tool, or you could realize that there's already tools that automate it for you. You can have an understanding the software landscape both from a programming language perspective and from a tooling perspective, and how those various tools can help you get stuff done that many people probably find very monotonous and boring, that detracts from the more creative work, the more interesting work, and the more impactful work. I tend to be one that promotes humans; we're hired to be creative, we solve problems that require abstract thinking. Let's let the computers do the non creative, boring tasks.
I would really recommend to people to find a problem that you want to automate or program, any problem or anything you want to build some software tool for. Think of that thing that you want to do and just start doing it. Google is your best friend. Ask it any question. You may not know where to start but just think about how you might solve the problem and start asking questions to Google and you’ll eventually find things that lead you down the path that you want to take. As you do that, you will start learning what are the questions to ask. Then you'll know the questions in the future for a different programming language or different tasks. The more questions you ask, the more questions you have in your arsenal, the more effective you'll be in the future. So the sooner you start doing that, the more effective you're going to be.
Excel is a good prototyping tool to get things worked out and understood and communicate prototypes to people, but try to think outside of the Excel box. There's a lot of tools out there but use the right tool for the right job. Excel can sometimes be used as a jackhammer for putting in some quarter round around your house or something. Know Excel well, start using it if you haven't already. Generally people that want to be actuaries in college probably have at least a budget of their own they keep in Excel or do some kind of Excel work. If you're a college student and you feel like you're pretty comfortable with Excel, once you get your first job, you're going to realize that you probably don't know as much as you thought. So just keep trying to break the barriers and think of what you could try to do differently.
Innovation is something that reduces trade-offs. So the key trifecta of trade-offs is you've got either high quality or low cost or low time and you’re trading off between those three things. Some people tend to think it's a zero sum game and if you take something away from one, then you try to add to one, you have to take away from another. But innovation comes from being able to add to one of those without having to make a change to another, or at least a smaller change to another one.
"The impact of innovation is that it allows you to make the impact that you want as efficiently and effectively as possible. It allows people to spend more time on more meaningful tasks."
Innovation has an impact in that it makes better decisions possible, it makes processes more efficient. From that trade-off perspective, it really gives the company that you're working for, or the product you're developing with that innovation, a competitive advantage. At least for a period of time until everyone else is doing the same thing. Innovation drives the ability to keep making an impact on the world.
The impact of innovation is that it allows you to make the impact that you want as efficiently and effectively as possible. It allows people to spend more time on more meaningful tasks. It creates whole new spaces to live in; some new insurance product or something like that, creates a whole new realm, a new model, a new ecosystem of community collaboration and innovation just kind of blossoms everything else from it.
I've got a very deep technical background, so I'll kind of stick with the technical side of things. Computers were a huge innovation back in the day and allowed us to be more effective at our job. In today's day and age distributed systems and CPU farms, GPU processing of projection systems, have really changed the game in terms of being able to run projections in a very reasonable amount of time.
It may be fairly expensive now to do some of that, but it's decreasing in price as time goes on and as more and more of these costs get driven down. A projection that might take, in the past, a couple of days on a cluster of servers might be reduced down to just an hour or two, even minutes, on the right architecture. That allows for a lot more creativity, decision quality and sensitivity analysis. There's things like proxy models that try to speed up processing using a watered down way of estimating reserve, for example. There's also clustering of model points to make a small point compression so you reduce your millions of model points to a number of buckets that represent your model points in a relatively small set of points. Those things are necessary when processing speed and hardware requirements are such that you can't run things as fast and cheaply as you want to. As these distributed systems architecture become more prevalent, we'll be able to do less of the proxy modeling and the model point compression and run entire models on every point in the data set in efficient time with efficient cost at efficient quality.
Again, that three pronged tradeoff equation will really be impacted by what's happening right now in distributed systems. That's something that I think students and other people, even current actuaries, should really get knowledgeable about, how to effectively use distributed systems and decentralized programing techniques. Which brings in decentralized finance and solidity and smart contracts. Smart contracts is an innovation happening right now in terms of the crypto world. Anyone that doesn't know what a smart contract is, I highly recommend looking it up. It could have interesting impacts on insurance contracts in the future. It's developed using the solidity programming language on the Ethereum blockchain.
