Automation Matters under IFRS 17

Though automation and workflow improvements have long been on the insurers’ “to-do list”, they are often kept aside for a time when work eases off. Ironically, a firm’s very lack of workflow solutions can explain why this time never comes. Now, IFRS 17 makes it more important than ever to automate your actuarial processes for three key reasons.

First, the new accountancy regulation asks actuaries to apply the same rigorous standards of audit and control as the finance department to actuarial modeling. As a result, they must reduce their dependency on manual processes.

Second, the senior managers of today’s insurance companies are increasingly interested in the risk metrics of their business, which means more modeling runs, data analysis and pressure on actuarial processes.

And third, insurance margins continue to tighten – firms need to make existing teams more productive through automation rather than hire more actuaries to ease the pressure on processes.

There are also good reasons why actuaries – unlike their finance colleagues – have largely resisted adopting automated workflow solutions until now. In contrast with highly standardized financial close processes, business processes for actuaries vary widely from firm to firm. The models are different, and so are the data requirements, thus making solutions complex to implement.

Despite these variations, actuarial processes (particularly for period-end reporting) do follow a common sequence that lends itself well to automation and should be the focus of any workflow solution to IFRS 17:

• Data gathering and enrichment – collation of data from a range of input systems, both internal and external, and including policy administration systems, finance systems and economic scenario generators
• Assumptions setting – although sometimes part of the data-gathering process, assumptions often require formal approval before use and involve extra stages in the business process
• Model execution – running the model using the data and assumptions gathered (and optionally approved)
• Aggregation and contextualization – summing up the results of the modeling process at the level expected by the business, frequently combining data from other sources to add context
• Results analysis – checking the results against expected values or cross-checking one set of results against another, sometimes automated but often significantly manual
• Dissemination – publishing the results to the business and regulators, usually via a third-party tool such as a business intelligence engine

The actual business process surrounding each stage can vary, with some firms requiring manual approval at certain points or automated “sanity checks” to ensure that results are not wildly different from expectations. Physical implementation can differ too, with a variety of systems being used to carry out every stage, particularly data gathering and dissemination.

It is also vital to remember that the sequence described above represents an ideal world where everything goes according to plan. Any workflow solution needs the ability to follow not only this path, but also exceptions – when a stage is missing data, is overdue, or has failed. When setting up business process definitions, paying a lot of attention to both exceptions and escalations will ensure you never miss a reporting deadline again.

Lastly, nothing in the world of actuarial science is constant. The recent uptake of cloud technology has been incredibly fast – from flat refusal to whole-hearted acceptance in less than five years. Focusing entirely on building a process for IFRS 17 may lead firms to create a rigid workflow with hard-coded stages that hamper the adoption of emerging working practices and technologies. So, while automation matters hugely under IFRS 17, keeping your workflow framework flexible can be just as important in the long run.