Robotic Process Automation (RPA) tools provide end-to-end automation of repetitive business processes. They pay back many times over the set-up cost in time and money savings by automating time-consuming and often costly functions.
RPA tools work best as no-code platforms, meaning they should be managed by business users, rather than IT specialists, to:
- Set up workflows
- Access multiple data sources
- Prepare data
- Create reports
The leading RPA tools also provide access to statistical toolkits that are ideal for technical business users such as actuaries and data scientists. But what are these users to do when these toolkits don’t meet all their requirements, or can’t perform a critical function?
When it comes to spreadsheet-based logic, like complex calculations and models, this is where Coherent Spark comes in. Spark is the perfect tool for when a business process includes taking information from large complex sets of calculations. In the insurance industry, these are often actuarial models which provide outputs for management information, or as part of a finance closing process or regulatory filing.
When these types of mathematical models run in an actuarial system and connect seamlessly to the RPA tool, then true end-to-end automation is achievable.
However, these actuarial models are often large Excel spreadsheets and, unfortunately, Excel and RPA tools are not always best friends! In particular:
- The complex Excel model could be recoded directly into the RPA tool, but only with a large investment of time in setup and testing.
- The Excel model cannot be readily integrated with the rest of the workflow process, hence blocking end-to-end automation.
- The more complex the model, the slower the runtime, especially when large data sets must be run through a large model.
However, Spark can be the bridge between Excel and RPA tools. The Spark platform automatically transforms the Excel model into fast running code with API connectivity. The RPA tool is readily configured to make API calls to Spark to:
- Feed the necessary inputs into the model, potentially from multiple data sources and after completing required data management routines.
- Pull the required outputs from the model to be delivered in subsequent stages of the process, including a dashboard or reporting tool.
Spark becomes the connector that enables the model to integrate into the entire process, achieving true end-to-end automation.
Case study – Automated Actuarial Valuation report generation
An Asian insurer was running a valuation process which:
- Ran detailed cash flow projections over 100 years.
- Required alternative scenarios to calculate provision for adverse deviation.
- Covered a million policies, grouped into 10,000 data points.
- Produced reserve and capital requirements as outputs.
The process was run in Excel as a validation check on the same process running in the commercial actuarial software used by the insurer. It took many hours to run in Excel. Several runs were required, and the actuarial team had to manually undertake routine data manipulation and reporting tasks for every run needed.
The insurer used an RPA tool for other processes and was reviewing how best to apply it to the valuation process. The tool was readily configurable for data handling and reporting but the actuarial model itself was more of a challenge. The chosen approach was to:
- Set up the RPA tool to extract the required policy data.
- Transform the actuarial model from Excel into a running API using Coherent Spark for run speed and API connectivity.
- Set up the RPA tool to run policy data through the model in Coherent Spark.
- Set up the RPA tool to automatically identify any breaks in the process and provide error warnings.
- Set up the RPA tool to feed results into an automated report and automatically email them to relevant management.
For further details of how Coherent Spark can fix a hard-to-automate process involving a complex Excel model, schedule a discussion with our actuarial and technology teams.
