Skip to main content

Intelligent automation in loss reserving

Why actuaries need help from robot analysts

Insurance Consulting and Technology
|Insurer Solutions|COVID 19 Coronavirus

By Jamie Mackay | April 26, 2021

Automation can help reserving teams to focus on the critical work that requires planning, decision-making and vision — a blend of art and science.

Among the many things that have changed how the actuarial field functions is an underlying movement to automate. By integrating cognitive computing capabilities into their actuarial processes, reserving teams have a new and powerful tool to better equip them to eschew repetitive tasks, like data cleaning, validating and loading, and analysis preparation. Instead, they can focus on the critical work that requires planning, decision-making and vision—a blend of art and science, so to speak.

Automation is not a goal, rather a means to achieving a business objective.

Automation is not a goal, rather a means to achieving a business objective. It is a tool to make people more productive. This is reflected in a recent Willis Towers Watson reserving survey, which probed chief actuaries and reserving actuaries among insurance companies about their satisfaction with their existing processes and strategic goals. Only 25 percent said they were fully satisfied, and 84 percent said that not having enough time for detailed, value-added analysis was a big concern.

Transition from manual work to adding value

Two of the more striking areas where satisfaction decreased over the last four years centered on “effective use of available actuarial staff” and “speed/timeliness.” In other words, there’s a perception the team is working too slowly of doing things that do not add value. This isn’t a harsh criticism, rather more reflective of the sheer volume of tasks actuaries are being asked to complete while also having to stake their professional integrity on the results produced.

Actuaries are expected to produce hand-painted works of art at the rate of assembly line paintings.

This has been even more apparent with disrupters like COVID-19. Actuaries are expected to produce hand-painted works of art at the rate of assembly line paintings, and this aligns more with what we are hearing as to the purpose of using automation. In fact, when probed about the overall objective of investing in automation, respondents who prioritized “adding value” have consistently far outweighed those looking to “save money” or “reduce team size” across several years of surveys.

As insurers seek to rely more heavily on data analytics, how can they best deploy these “robot reserving analysts” where they can add the most value to the company and the C-suite, and what does this “added value” actually look like?

This is where an “optimization” goal surfaces: In a field where data and granularity are growing at enormous rates, how do we focus the actuary’s attention where it matters the most and when it matters the most? How do we leverage technology to provide the actuary with the information to apply their art, deploy their judgment and develop their opinion?

Better and deeper analysis

The first rather mundane but crucial part of leveraging automation is identifying the critical path of data and information through a reserving process. How do we get the information loaded into the tools used for analysis and the results that feed back into the downstream data repositories? This really is automation in its simplest form: getting things organized and done expeditiously.

Before we can add value, the core routine needs to be taken care of quickly with ease and confidence. This is where use of automation and extract, transform, load (ETL) tools are critical in communicating information and preparing working files so actuaries spend less resources running queries and inserting rows but instead spend more time performing critical thinking.

But today’s actuaries aren’t just tasked with leveraging more data from claims or new techniques. Actuaries are under increasing pressure in an environment where claims experience is rapidly changing to prepare more regular analysis and provide more updates, not just quarterly but perhaps monthly or weekly—or even on-demand. Automation is key to making analysis efficiently re-runnable. If teams can perform analysis quicker, then it could become cost-effective to do it more regularly.

This past year has demonstrated the value in regular—and deeper—insights more than ever, especially in short-tailed lines like personal auto. With the COVID-19 pandemic causing dramatic and unprecedented changes in driver behavior, a quarter is a long time between analyses. Leveraging automation to take more regular cuts of the data and efficiently re-apply and update assumptions has been critical in providing insights in a quickly changing environment.

In reserving, we’ve long seen the value of doing isolated scenario analyses or running thousands of simulations through individual models. But what if we quickly and easily could define multiple sets of alternative assumptions, running each through the entire analysis automatically when the data is made available? Before even starting an analysis, the actuary would be able to identity where their insight and judgment is needed most, which is critical information on Day 1 of a short turnaround time.

More effective communication

Reserving teams can be guilty of generating and disseminating enormous amounts of data in often indigestible formats. We’ve historically had a tendency to produce so much data and information that oftentimes we are unable to see the wood for the sheer volume of trees involved in producing the technical appendices that accompany our reports.

There may be a lot of important analysis that you want to present to your stakeholders, but how effective is it if they can’t visualize it? How does the brain analyze numbers?

Will your CEO be able to focus on the message or declutter that message?

When dealing with datasets that may include hundreds of thousands of data points, automation, combined with the interactivity provided by data visualization tools, can be very useful. Using engaging elements like charts, graphs and maps, data visualization tools ensure that reserving teams, and their CEOs, are able to see—rather than read—analytics to help them understand trends, outliers and grasp difficult patterns in data. Intelligently designed data visualization can influence and inform a decision based on the data analyzed. As the amount of data generated increases radically, intelligent data visualization will become more and more critical. And automation is the key to enabling the gathering, organization and clear communication of insights.

Actuaries should perform creative problem solving, critical thinking and decision-making rather than data wrangling.

Actuaries should perform creative problem solving, critical thinking and decision-making rather than data wrangling. Automation is crucial in taking the structured information from the core reserving platform and delivering it to the right people at the right time. More and more, intelligent automation is being used by carriers to enhance their reserving function. They are identifying areas that provide the greatest opportunity and ease of entry to enhance processes so that reserving has become a work of art.

This article first appeared in the March 23 issue of Carrier Management.


Director, Insurance Consulting and Technology

Related Capabilities

Contact Us