How actuarial rigor, combined with technology, communication, and risk-taking, is opening doors to nontraditional careers.
By Barbara Bryant
For many actuaries, a typical day involves analyzing and predicting risks, helping set insurance premiums, or calculating pension plan reserves for an insurance company.
These skills not only support actuaries in traditional roles but also enable them to expand into new positions, sometimes in industries where actuaries have rarely, if ever, had a formal presence.
The trailblazers featured in this article attribute their success in seizing such opportunities to mastering skills beyond the standard curriculum, taking risks to chase their passions, and learning from others. They include Chelsea Adler, senior state product manager and actuary at Root Inc.; Frank Chang, Uber’s vice president of Applied Science; Ian Duncan, adjunct professor of actuarial statistics, University of California, Santa Barbara; Kyle Bartee, founder of Aviation Ops Ally; and James Guszcza, cofounder and a principal at Clear Risk Analytics LLC. They also emphasize that their actuarial skills have been vital in helping them pursue new opportunities and overcoming unfamiliar challenges.
Tackling Technology
“I learned Python early on the job at Uber,” recalls Frank Chang, who joined the company nearly 12 years ago. “I needed to process a lot of data to calculate mileage, which you can’t do in Excel.” That’s just one example of the skills Chang had to acquire after leaving his lead actuary job at Google, where the company had a far smaller automotive footprint.
As Uber’s director of actuarial services, he focused heavily on risk assessment related to the significant insurance coverage Uber purchases from large carriers. He then moved into increasingly varied roles, first as head of Insurance and Safety Analytics and later as head of Core Safety Analytics. Now, as vice president of Applied Science, Chang manages large actuarial and data science teams that focus not only on safety and financial risk, such as fraud, but also on applying data science across a wide range of operational functions. These functions include customer service reporting, identifying customer payments within the Uber app, managing worldwide payment flows, and ensuring reliable messaging across Uber’s ride-sharing platform.
Ian Duncan drew on his graduate degree in economics, a Ph.D. in actuarial science, and experience in management consulting, predictive modeling, and value-based contracting to found or co-found several health care data companies, as well as an employee benefits decision-support company. He has worked for a diverse set of employers, ranging from Aetna and PwC to Walgreens and the University of California, Santa Barbara, where he serves as an adjunct professor of actuarial statistics.
In 2023, Duncan sold his last company, Santa Barbara Actuaries, to Arbital Health, a software company focused on the health care industry’s transition to value-based care.
While using his insurance expertise to advise clients on lines of business strategy and product development, financial projections, and enterprise-wide systems for PwC’s managing consulting practice, Duncan also worked with SAP, a German company that develops enterprise software to manage business operations.
“That was unusual for an actuary—helping to implement big systems, including finance, HR, and supply chain operations for insurance companies. This is how I learned about technology and systems,” he explained.
Duncan began using the underwriting systems he developed and claims data to build predictive models that could identify which individuals were likely to incur high medical costs.“What I knew about predictive modeling was in demand; there was a need to find a way to deliver predictions to clinicians who could manage patient risk, so I helped a few health plans implement these systems. My approach was to take predictive modeling and build workflows and systems around it.And now,actuaries are beginning to be hired to do predictive modeling in-house for physicians.”
That led him to co-found a company focused on health care predictive modeling and workflow systems technology, which he and his partners sold at the height of the dot-com bubble.
“Doctors and engineers know how to build technology and about health care, but [they] don’t know how to get paid for it,” Duncan said. “In health care engineering startups, actuaries understand the payers and payment mechanisms. That’s what I’ve built business on in the last 25 years; it’s interesting non-traditional work and I get to see a lot of startups.”
For Kyle Bartee, founder of Aviation Ops Ally, stepping into a hybrid technology-and-actuarial role helped propel his career forward. As an intern studying actuarial science, he learned how to use Excel hotkeys to create macros, spreadsheets, and templates, including a rater template product that reduced the time required to process 100 policies from eight hours to just 10 minutes. The process efficiencies he developed quickly elevated him from the company’s “Excel guru” to a project manager, setting him up for continued success. Later, while working at a 1-year-old startup, he helped engineers develop rules for a program that enabled the company’s internal policy management system to handle renewals more efficiently.
“It was an IT project, not actuarial, but from there I became the company’s underwriting and product manager. I was also the IT and business analyst, translating business requirements into IT requirements.” Bartee also handled pricing, reserving, and reinsurance; managed relationships with program administrators; and took on reporting responsibilities. “After a while, I realized I could do all of this on my terms, so I started my own aviation insurance company.”
James Guszcza, a principal at Clear Risk Analytics LLC, was the first person to serve as Deloitte’s U.S. chief data scientist. While working at the Allstate Research Center in 1999, he built the company’s first continuous credit scoring model. He first learned what would later be called “data science” through a short professional development course on statistical learning by Trevor Hastie and Rob Tibshirani, several years before their (and co-author Jerome Friedman’s) seminal textbook, The Elements of Statistical Learning,was published.
After moving to Los Angeles, Guszcza accepted a predictive modeling role with Deloitte (then Deloitte & Touche), using artificial neural networks—a simpler form of deep learning that today is the cornerstone of AI engineering.
“I suspected that neural networks weren’t the right tool for messy and sparse commercial insurance data and, indeed, we quickly embraced linear models,” Guszcza recalled. “Deep learning is a big deal today because text, films, conversations that are digitized and stored in the cloud, can be treated as training data for machine learning algorithms. And this gives us things like image recognition and generative AI. In contrast, commercial insurance data is thin, heterogeneous and noisy—so [it] calls for simpler models.
“Part of the joy for me was the ability to experiment and the feeling of creating a new sub-discipline,” Guszcza explained. “The work we were doing wasn’t mainstream or taught in universities, which focused mostly on principles of statistics but not on how to apply them—and certainly not in commercial insurance.”
Leaning in Hard on the Soft Skills
Technology’s usefulness in helping actuaries collect, organize, and analyze vast amounts of data—and streamline processes and procedures—are indisputable. But many of those featured in this article pointed to another type of expertise as equally essential and often harder to master. They stressed the importance of developing soft skills: empathy—the ability to understand other people’s perspectives, priorities and concerns—as well as clear, persuasive communication and presentation skills, and the ability to lead and motivate others.
“Learning technology isn’t that hard,” Chang said. “It just requires dedication. Mastering soft skills is harder. You need to gain trust from your colleagues and clients to win business. It requires networking, storytelling with data, and understanding your audience, which I learned the hard way by doing post-mortems on bad meetings.”
Chang recalled a sales forecast presentation he made to Google executives after the company bought Motorola’s phone line. “I rearranged data into triangles—a format actuaries understand—to examine variables that could affect phone sale volume,” he said. “When I presented the results to leadership, I pegged sales at the 75th percentile, but they didn’t understand what that meant until I reframed it as our having a 1-in-4 chance of hitting our sales target. That, they understood.”
Chang learned a similar lesson while pitching a proposal to executives at Uber, where he highlighted the accuracy of his phone sales forecast to demonstrate his skill set. “I failed by just focusing on my successes. They weren’t that interested. If I’d taken the time to read the room, I could have provided more background that mattered to them. I should have prepared by showing more curiosity and doing more fact-finding. You also need to know how to collect intel by holding back-channel conversations.”
Bartee stressed the importance of efficiency and brevity. “One of my biggest barriers I had to break down as an actuary was overexplaining when talking to executives or decision-makers. Actuaries tend to want to display the depth of their analysis to add credibility to what they’re proposing. It took me 10 years to realize that I overexplained my pitch, which was why executives weren’t adopting my proposals. I was so far in the weeds that I was losing my audience.”
He advised paring that 20-slide presentation you worked so hard on down to three, and using plain language to get the point across.
“Executives need to move fast, so the longer you take to give them the information, the greater you’re at risk of losing them,” Bartee said. “Show the meat and potatoes in the first graph, with a little context. Follow that up with one [slide] on implementation strategy, KPIs, or metrics you’re going to influence, and finish with the estimated impact—all translated into language they’ll understand.”
Duncan also drove home the importance of developing effective communication skills.
“I advise actuaries to practice making public presentations and to write,” he said. “I never refuse an opportunity to speak into a microphone. Some years ago, I interviewed former Academy and SOA [Society of Actuaries] president Barbara Lautzenheiser, about how she grew her practice. She explained that she never turned down the opportunity to get behind a microphone. She once said, ‘I turned from mousey to mouthy.’”
Accentuating Actuarial Acumen
Although this article focuses largely on nonactuarial skills, experiences, and roles, all of those interviewed emphasized that the rigorous training the profession provides has been central to their success.
After working as a pricing actuary at State Farm, Chelsea Adler joined Root Inc., an auto insurance startup in Columbus, Ohio, where she moved into state management. In this role, “I’m almost a mini-CEO for the markets I’m responsible for,” she explained. “I own profit and loss and am responsible for every decision in the states I’m assigned to. It’s a unique and exciting opportunity because it gives me exposure to every aspect of insurance.”
She said that her work overlaps somewhat with her previous actuarial work, including decisions on rate calibration, implementing models for filings, and maintaining relationships with regulators.
“But now I’m involved much more broadly with our insurance products, such as underwriting and claims,” Adler said. “Getting to use data to drive those decisions leverages my actuarial skill set, but I’m applying it to solve different problems.”
Adler also credits actuarial training with opening doors to a variety of other professions, noting that she has seen former co-workers transition into product management, software engineering, and data science careers, and points out that Root’s CEO is a credentialed actuary.
“The training ensures you have the necessary rigor from a statistical standpoint, so decisions are data-driven, and it pairs that with an understanding of the business you’re operating in. We’re strong analytically and can apply that appropriately and recommend sound business decisions because we understand how insurance works and the fundamentals of what we’re building models for.”
Duncan explained how he has used his actuarial skills and predictive modeling to work with health care providers to help them track health insurance costs. “The industry needed someone who knew technology and the data side of things, and I developed a consulting practice on that, working with a now-defunct industry trade group, Disease Management Association of America, which managed chronic populations.” As the only actuary in the group, which was composed mainly of doctors and businesspeople, he was able to explain how money flowed from insurers to doctors and hospitals. “If you want to demonstrate that you reduced costs, you need an actuary to assess whether you’ve been successful at lowering cost per member per month,” he explained.
Guzscza foresees a broader, more ambitious role, and sphere of influence for insurers—and actuaries—because of their ability to draw on statistics, data science, and behavioral science to improve human judgement and decision-making. He believes this could be accomplished by integrating machine learning methods, Internet of Things and digital technologies, artificial intelligence (AI), and behavioral economics to help insurers encourage policyholders not just to insure against risks but actively reduce them.
“This might take the form of giving policyholders timely reminders, social-norm nudge messages, customized checklists on how to fortify their homes, instructions for improving fire safety by clearing brush or weatherproofing windows before storm season, and guidance on improving their driving habits and safety behaviors on the job,” Guzscza explained. To this end, he envisions the emergence of “a breed of business- and social-minded actuaries [who] will help broaden our focus beyond traditional actuarial segmentation and profitability to create business models that improve the customer experience and create new forms of value for policyholders. This could be a way of making a kind of ‘behavioral actuarial science’ mainstream.”
Taking on Risks—and Creating Opportunities
Transitioning to new roles often requires the ability and willingness to step out of one’s comfort zone—to react in unfamiliar ways to new situations and make decisions and predictions based more on current operational factors and less than on historical data.
Chang described a credit card project that Google wanted to offer to its ad buyers. “The company didn’t know how to set the risk threshold, how much to charge, or what the underwriting risk was likely to be—they had no data. I pulled hundreds of prospectuses for buying securitized debt. I used this information to simulate debt outcomes for default and used it to launch the card. My predictions weren’t quite as accurate as they were for the Motorola phones, but you have to solve a problem. I don’t take a strictly actuarial approach to everything I do, but I never stray far from my roots; my work has to be quantifiable and accurate.”
Duncan believes his academic background helped equip him to succeed as an entrepreneur, but he felt he couldn’t take on the professional risk of launching his first startup until one of his children was in college and the other had graduated. While acknowledging the uncertainty taking such chances poses, he suspects others are simply reluctant to venture out of their comfort zones.
“The money is so good in traditional roles that stepping into the unknown and taking a huge pay cut in hopes there’ll be a big payoff in the future is the type of risk not many are willing to take on.” But, he added, “On the other hand, I also know that I don’t make a good employee; I learned that I needed to work for myself.”
Duncan believes that the actuarial profession should encourage its members not to stay in their lane. “The biggest need isn’t education; it’s to get actuaries to become more entrepreneurial, innovative, and continue to learn. We’re trained to say no, to follow regulations, that’s much of what we do. But many actuaries actually are entrepreneurial and we need to find ways to encourage that.”
“Always think about how you can improve something and, if you’re interested in entrepreneurship, develop something new yourself. That’s what differentiates you: what you can turn into a product or service. And everyone should start a company at some point in their career,” he added. “You’ll learn how difficult it is and will be less likely to complain about your boss. It’s not easy, but it is rewarding, and you learn a lot quickly.”
Duncan has helped young and future actuaries do just that. He has hired a number of former students to work for him. He recalled one who joined him as a business partner and later turned an idea into a new startup the two of them launched, which is now run by that protégé.
Adler said she was grateful to take on the unfamiliar and more ambitious leadership and planning roles her state management position entails, which she believes helped her gain a deeper and more holistic view of the company she joined.
“Mine is less of an execution role and more of an influence role now. I’m setting the priorities rather than having them handed to me and I’m executing on them. It’s important to invest in relationships so I can get priorities I think are important onto a roadmap to help make them happen.”
Adler also has better insight into how and why upper management decisions and plans are made. “I understand now that this is why we do things a certain way and that there’s so much more to the picture than I could previously see. It gives me a valuable perspective on pain points that I hadn’t known about. If I stepped back into a traditional actuarial role, I could use this perspective.”
Paying it Forward
Several of these trailblazers are thankful to role models and helpful colleagues who supported their progress, and they have committed to mentoring others in turn. Adler said her father, an agent at State Farm, modeled how to help others and encouraged her to develop leadership skills. She recalled accompanying him and her sister on trips to policyholders’ homes after storms to help them clean up their yards. She has sought to pave the way for others to enter and excel as actuaries through her blog, Inspiring Actuaries (inspiringactuaries.com), and by recruiting volunteers to write and grade actuarial exams. And she’s found that she gets as much from volunteering as she gives.
“Most people want to pay it forward, but I’ve never met someone who didn’t get something in return: gratitude or new friendships they never anticipated. And the people you meet or reconnect with can open the door to new opportunities. It’s a powerful way to project your career forward.”
Bartee said he started his company to use his actuarial skills and insurance industry knowledge to support the goals of those who share his passion. “I’m not in it for the money. I want to help all general aviators get from the hangar to the horizon with fewer of the pain points they experience before they get to fly.”
Duncan, who is the Society of Actuaries’ president-elect, said, “I’ll also put in a word for academia.” The university professor, who has taught and guided more than 1,000 students over the years, said, “It isn’t for everyone, but for an actuary who wants to give back to the profession and is able to afford a pay cut, teaching, training, and mentoring young actuarial students is extremely rewarding.”
Barbara Bryant is a technical writer at the Academy.
Actuaries and AI: Opportunity with Guardrails
Actuaries are finding many interesting ways to use artificial intelligence (AI) to advance their careers and help clients. That said, they also recognize that “guardrails are needed,” says Chelsea Adler, senior state product manager and actuary at Root Inc.
Frank Chang, Uber’s vice president of Applied Science, treats AI chatbots as a teaching—and testing—tool. “I ask the product I use to employ the Socratic method to teach me about a subject. I ask it for useful information and resources and then tell it to ask me questions to see if I understand what I’ve learned. It helps me to figure out what questions to ask, what problems to solve, and who to talk to about a particular space.”
James Guszcza, a principal at Clear Risk Analytics LLC, believes it’s a mistake—and an underestimation—to view AI as simply a “labor substitution technology” or a way of using an algorithm to automate decision-making. He asks, “How do you create a human–algorithm system that produces the right outcome?” He’s also excited about the prospect of using generative AI to revise human workflows and urges exploration into “when and how to design and deploy AI to streamline ratemaking, rate filings, competitive analysis, customer support, claims adjudication, marketing, underwriting, nudging safer risk behaviors, and so on. It’s a great set of issues for actuaries to get involved in.”
Guszcza believes that actuaries have a valuable role to play in promoting responsible AI use. “If the AI world had something analogous to the actuarial profession, its practitioners would be more than machine learning engineers. They would also need to understand different ideas of algorithmic fairness; they’d work to understand how different stakeholders are affected by AI algorithms, and to design and deploy them in ways that respect their safety and autonomy. The current landscape creates a need for people who both understand the technology and are willing to be socially responsible actors. To me, that sounds reminiscent of the actuarial profession.”
Kyle Bartee, founder of Aviation Ops Ally, recalls that when he was working as a pricing and reserving actuary, he used AI to master a specialty he knew nothing about: reinsurance. In less than a month, he learned enough to help conduct a $2 million reinsurance negotiation on the fly. “After three weeks of self-training using AI, I was able to dig into the contract, review previous contracts and get a reduction in the capital requirement for the older contracts to offset the $800,000 in capital the company was asked to put up for the new one. I learned how to wield AI like Excalibur. But I don’t use it in legal contracts or to understand FAA regulations because I don’t have that subject-matter expertise to identify hallucinations or outdated information. I only use it in areas where I’m fluent.”
Adler sees AI as a useful tool but stressed that rules and training are needed to ensure the technology is used responsibly and accurately. “I’ve heard from people who have hired interns or entry-level actuaries that they’re lacking the ability to think deeper, to question things, and to consider whether the output they’re receiving is reasonable,” she said. “They’re just taking it at face value, so training on how to interact with tools as they evolve will be important for actuaries.”