Articles - Introducing IFRS 17 Software with Machine Learning

 3Blocks has been known for providing IFRS 17 and Solvency 2 reporting advisory services, but now it has launched its own IFRS 17 software. They have been working on IFRS 17 projects for insurance companies and other IT firms since 2016. Their solution has been built on the back of these experiences, taking into account lessons learned from them. 3Blocks PAA is intended for small and medium-sized non-life insurance companies that do not have the time and budget to implement expensive alternatives offered by competitors. 

 Algorithms designed by 3Blocks learn what patterns are exhibited by claims, expenses, premiums, and other reporting data. Then, users can check if the actual amounts fit the pattern. Significant variances may indicate data issues that require fixing.

 Machine learning is not the only trick that 3Blocks has up its sleeve. Most IFRS 17 software solutions available on the market require data on a defined level of granularity. For example, they require that all data should be by group or by policy. 3Blocks software gives full flexibility in this respect, as data and calculations can be arranged by policy, group, or portfolio, with a possibility to mix levels.

 Secondly, the solution supports three different accounting rules sets, so whatever accounting rules are used by an insurance company, one of these sets should fit in. Lastly, numerous business validation rules and data completeness checks go a long way to ensure data quality.

 From a technical perspective, 3Blocks software works with the SQL Server database with all the calculations being executed by a code written in Python. This is not a black box though, as the software produces an Excel replica of the calculations, so there is full visibility in this respect.
  

 YouTube video

 Website