It’s Day 0 of the month-end valuation process. As the actuary leaves the office for the day, an automated script, or bot, prepares and executes a series of model runs to calculate reserves. With the job complete and model results automatically stored in a shared location, the bot analyzes the results and recognizes that the change in reserves is out of tolerance relative to prior quarters. The bot completes an analysis of the underlying drivers of reserve change and discovers the number of terminations increased for policies with a loan, and the number of policies with a high loan-to-cash value ratio also increased. The bot summarizes its findings using natural language generation in the quarterly analytics package, highlighting the trend in reserves using a visualization dashboard.
At 8 a.m. on Day 1 of the valuation process, the actuary walks into the office, reads an email containing the analytics dashboard from the bot and starts the day by exploring the increase in loan utilization. After a quick conversation with the marketing department, the actuary learns that a campaign to promote policy loans to policyholders took place during the quarter. So, on Day 1 of the monthly close, the actuary explained the unintended consequences of the loan campaign. Prior to implementation of the bot, the actuary may have spent two days calculating reserves and two days performing analysis. The connection to the loans may have taken weeks or months to discover. This is the future of work; this is the reimagined world of the Exponential ActuaryTM.
The exponential actuary—an actuary who is augmented by emerging, exponential technologies—is poised to leave behind traditional tasks to focus on higher-value, strategic roles within the organization. This new breed of actuary should arise to navigate this shifting landscape by embracing technology and focusing on outcomes that require uniquely human skills.
Current Actuarial Challenges
Today, many actuaries find themselves overwhelmed by manual machine-type or lower cognitive work that inhibits productivity. As an example, some actuaries work with more than 200 spreadsheets that are linked together to calculate, capture and allocate investment income. Others find themselves waking at 3 a.m. to confirm the valuation process completed effectively and to execute the next job. Others spend countless hours executing controls or refreshing actuarial memos after completing analysis on the actuarial results. As actuaries become overwhelmed by these operational and stewardship activities, they may lose track of the overall outcomes that necessitated the task in the first place. Due to constrained capacity, actuaries may operate suboptimally or lack focus on strategic activities that provide powerful insights into the business. The actuarial workforce is valuable yet expensive, and these constraints prevent many companies from achieving a high-quality return on their human capital. Exponential technology creates an opportunity to transfer the “machine work” to machines and expand the role of the actuary.
Cognitive technology—defined as technology that can “perform and/or augment tasks, help better inform decisions and accomplish objectives that have traditionally required human intelligence, such as planning, reasoning from partial or uncertain information, and learning”1—is now affecting the insurance landscape. The rapid development of technology, especially cognitive technology, is one of the major disruptors of how work is completed for many back-office functions. The rate at which the maturity and availability of technology has increased now outpaces the rate at which many companies have adopted these technologies. Although the insurance industry—and the actuarial profession in general—is behind other industries and sectors in the use of these “exponential” technologies, all is not lost.2 Some of the most exciting and pertinent technologies with actuarial applications include data wrangling software, robotic process automation (RPA), natural language generation (NLG), natural language processing (NLP), machine learning and crowdsourcing.
Automation, Augmentation and Actuarial Modernization
Exponential technologies are capable of both automating actuarial tasks and augmenting actuarial outcomes. Many actuarial roles involve spending a great deal of time working with spreadsheets, drafting memos and performing basic review and analysis. Much of the effort expended by actuaries is prime for automation, especially tasks that are repetitive, routine, rules-based, computational and/or have low cognitive requirements. Bots can perform these tasks faster, cheaper and with few to no errors. The use of automation augments the role of the actuary by providing capabilities that allow the actuary to provide more value to the enterprise. At the same time, it also helps the actuary to focus on the pipeline of value-added activities that are often deprioritized because they are seen as “nice to have” rather than a “must have.” Decades ago, the introduction of computers gave actuaries a tool that could price products and calculate reserves more quickly and accurately. More recently, enterprises have incorporated predictive analytics to help drive informed decisions. Exponential technologies give actuaries a similar opportunity to explore uncharted waters.
With automation and augmentation by technology, actuaries can elevate themselves into a more strategic, valuable role by redesigning how their work is done and what they can accomplish. This involves analyzing how work is currently performed and what opportunities exist to enhance future business outcomes using technology. Based on such analysis, processes and outcomes can be redesigned to apply exponential technology to create a reimagined role for the actuary.
Exponential Technology Applications
The rapid development of technology, especially cognitive technology, is disrupting the insurance industry. Some of these technologies, including different types of language technology, machine learning and crowdsourcing, have actuarial applications. They include:
- Data wrangling software. Streamlines and automates data validation, manipulation and cleansing.
- Robotic process automation (RPA). Software or bots provide macro-like capabilities that can be deployed at an enterprise or business unit level. Bots can open emails and attachments, log into web/enterprise applications, move files and folders, fill out forms and perform calculations.
- Natural language processing (NLP). Ingests unstructured data (e.g., emails, tweets, memos) and creates data designed for machine consumption.
- Natural language generation (NLG). Ingests structured data and outputs data designed for human consumption. NLG can be used to generate a memo from data that mimics how a human would write by varying language to reflect a change in the underlying data.
- Machine learning. Uses algorithms such as neural networks to learn and become more accurate as the tool experiences more data. Such tools can be used to identify trends and anomalies within data or rank and identify the most relevant variables. Actuarial applications include both reserving and experience studies.
- Crowdsourcing. Leverages a large network of people to solve problems or provide services. Insurance organizations are exploring internal and external crowdsourcing applications.
Machine Learning in Action
Consider how technology could disrupt the experience studies space. Actuaries spend a great deal of time performing experience studies and profitability analysis through iterative processes to identify drivers of policyholder behavior and the impact of those drivers on the profitability of the enterprise. An experience study tool, with a machine learning algorithm at its core, can automate the actuary’s work and augment the outcome of the study. Machine learning algorithms such as neural networks can be trained using thousands of data sets to develop expertise in a particular product line. When such a tool is applied to experience data, it considers all possible variables and identifies drivers that the actuary might not have considered when doing a manual review. An actuary equipped with a machine learning experience study tool can perform more frequent experience studies that are coupled with greater insights due to the additional correlations identified in the underlying data. Moreover, in an inforce management context, the time saved in performing the study can help companies more quickly identify policies likely to lapse and create opportunities for actuaries to partner with the business to develop programs and new products to increase (or decrease) the persistency of the inforce population.
Evolution and Adaptation
Actuaries should focus their efforts on outcomes requiring uniquely human skills to complement the capabilities of machines. Machines are strong at executing routine tasks without bias. However, critical thinking, creativity, communication and resilience are uniquely human abilities that define the exponential actuary.3,4,5 According to one estimate, the half-life of an acquired skill has fallen to five years.6 Another suggests that approximately 50 percent of the material comprising the first year of a four-year technical degree is obsolete by the end of the program.7 The exponential actuary must engage in lifelong learning over the course of his or her career. While the typical actuary may not currently be adept in some of these areas, companies can help their actuaries develop these skills with the assistance of the right trainers and training programs.
Implication 1: Training and Education
The first implication of the evolved actuarial role is that training and education of students and actuaries of all levels will require updating. Today, many new actuaries grow in their careers through the apprenticeship model—where work experience is a large component of qualification. For example, entry-level actuaries master the occupation by painstakingly working through manual calculations that trace through the details of reserve calculations. In a future where machines manage many of these entry-level procedures and actuaries shift to augment the machine, new ways of learning will be required to train actuaries on the ins and outs of the business. The responsibility for this will fall on companies, professional organizations, universities and, of course, the individuals themselves.8,9
Automation in Action
Voya experienced the benefits of modernization by transforming its annuity experience studies process. The modernized process mapped data across multiple administrative systems to a central, policy-level database and moved the legacy process to an IT-controlled production process using an Oracle-based platform. This solved challenges by reducing rounds of querying, which previously included several different administrative systems and extensive manual data manipulation. With this transformation, Voya achieved several successes:
- One source. Actuaries can now quickly query the new one-stop seriatim database to satisfy ad-hoc data requests. Past requests required a painful process of querying several tables across several administrative systems, resulting in inefficiently combining multiple outputs in order to respond.
- Less error. An IT-controlled process requires far less manual manipulation, thus greatly decreasing the operational risk of human error.
- More time. The expected time to prepare studies went from weeks to hours. This provides actuaries with more time to explore and improve their analysis. For instance, they connected to data sets through a data visualization software, Tableau, to analyze dynamic formula assumptions via an interactive dashboard.
By transforming the experience studies process, Voya transformed the role actuaries play—focusing more on analysis and less on data processing—and ultimately benefitted from modernizing.
Companies may consider tasking junior-level actuaries with performing independent analysis to validate and understand the output the machines produce. Professional organizations and universities, on the other hand, can work together to help identify which enduring, uniquely human skills are most vital to the actuarial profession, adjust the curriculum accordingly and teach students how to apply these skills on the job. For example, in addition to solving problem sets, students could spend more time answering questions that engage cross-functional student teams and help identify risks and opportunities. Actuaries themselves should prepare for a career of lifelong learning to stay abreast of technological trends and quickly master cutting-edge technical skills. Such efforts from all parties could supplement learning lost from the automation of manual activities and could help employees enhance skills machines are not able to replicate.
Implication 2: Professionalism
A second implication is that with an elevated and transformed role in the business, exponential actuaries will need to adapt their professional judgment and conduct. On one hand, the actuary will need to learn to trust the work of the machine to benefit from the opportunities brought by using it.10 The machines lack human intuition to account for certain external factors and considerations of which actuaries are aware as a normal course of their roles. As such, actuaries will need to consider how and when to challenge the results of the machine. Today, it is common for actuaries to rely on the work of other actuaries and trust the work of other professionals. Actuarial Standard of Practice No. 23 discusses the use of data and reliance on data created by someone else, and the considerations one must take into account when utilizing work produced by others.11 To what extent would this change if the data supplier is a bot? Similar professional standards will need to be created for the human-robot relationship. In an increasingly automated and augmented actuarial profession, it is easy to visualize a greater proportion of the actuary’s time being spent on quality review and other professionalism activities—an important example of applying uniquely human skills. Undoubtedly, actuarial standards and the actuary’s duties to stakeholders must not be compromised.
Just as many stakeholders are responsible for the education of the exponential actuary, many stakeholders stand to benefit from optimization of the actuarial role. First and foremost, actuaries themselves should be excited by the opportunity to have their work automated and augmented by technology. Actuaries could gain a greater work-life balance and flexibility through reduced workloads. Fewer actuaries would need to work weekends or go out of their way to make sure their model runs are completed on time. Moreover, actuaries gain the opportunity to enhance their professional potential by spending less time on manual tasks, focusing instead on the work they are interested in and passionate about, and potentially expanding their position within their company.
Organizations and the actuarial profession stand to benefit as well. Actuarial functions within various organizations may be able to provide a higher return on their actuarial talent once individuals are enabled by technology. Prospective actuarial candidates may be naturally drawn to work at organizations that invest in cutting-edge technologies and strategies. Hence, organizations that apply this “future of work” framework for actuaries can potentially create a workplace that attracts skilled candidates. Similarly, for an organization’s current employees, individuals have an opportunity to complete more interesting work and reach their professional goals while being less encumbered by manual and mundane tasks. The actuarial profession may benefit from increased relevance as the profession becomes more outcomes-based and strategy-focused. As the day-to-day activities and the value the profession provides become more attractive, an increased number of individuals could be driven to explore careers in the actuarial field. This is why professional actuarial organizations like the Society of Actuaries (SOA) must prepare their members for their exponential role.
Discussions of automation in the workplace often evoke fear, as professionals worry that robots will take away their jobs and make them obsolete. However, the actuary’s responsibilities are defined in terms of outcomes and stakeholder impacts rather than tasks. Instead of worrying about machines taking away work, actuaries should embrace such change by transferring machine work to the machines and shifting to more strategic tasks that enhance the actuary’s role within the enterprise. Companies can facilitate this movement by identifying opportunities for automation and augmentation, investing in exponential technologies, defining new strategic actuarial positions and supporting employee learning. A workforce of exponential actuaries can rapidly accelerate productivity, capabilities and impact enabled by technology.
Copyright © 2018 Deloitte Development LLC. All rights reserved. Deloitte refers to one or more of Deloitte Touche Tohmatsu Limited (DTTL), a U.K. private company limited by guarantee, its network of member firms and their related entities. DTTL and each of its member firms are legally separate and independent entities. DTTL (also referred to as “Deloitte Global”) does not provide services to clients. In the United States, Deloitte refers to one or more of the U.S. member firms of DTTL, their related entities that operate using the “Deloitte” name in the United States and their respective affiliates. Certain services may not be available to attest clients under the rules and regulations of public accounting. Please see www.deloitte.com/about to learn more about our global network of member firms.
- 1. Renner, Ryan, Mark Cotteleer, and Jonathan Holdowsky. 2018. “Cognitive Technologies: A Technical Primer.” Deloitte Insights. February 6. ↩
- 2. Pelster, Bill, and Jeff Schwartz (eds). 2017. Rewriting the Rules for the Digital Age: 2017 Deloitte Global Human Capital Trends. Deloitte University Press. ↩
- 3. Bersin, Josh. 2017. “Catch the Wave.” Deloitte Review 21 (July): 73. ↩
- 4. Talent for Survival: Essential Skills for Humans Working in the Machine Age. 2016. Deloitte. ↩
- 5. Hare, Julie. 2015. “Communication, Business Skills Essential for Graduates.” The Australian. September 23. ↩
- 6. Deloitte. 2017. “Meet Your Future Workforce.” Quartz. November 28. ↩
- 7. “Skills Stability.” 2016. The Future of Jobs Report. World Economic Forum. ↩
- 8. Bersin, Josh, Dimple Agarwal, Bill Pelster, and Jeff Schwartz (eds). 2015. Global Human Capital Trends 2015: Leading in the New World of Work. Deloitte University Press. ↩
- 9. Engelbert, Cathy, and John Hagel. 2017. “Radically Open.” Deloitte Review 21 (July): 100. ↩
- 10. Guszcza, Jim. 2017. “Who Determines Ethics in a Machine-run World? A Case for ‘Society in the Loop Artificial Intelligence.’” Deloitte. May 19. ↩
- 11. Data Quality Task Force of the General Committee of the Actuarial Standards Board. 2016. Actuarial Standard of Practice No. 23. Actuarial Standards Board. December. ↩