“77% of risk professionals are negatively impacted by the process of managing and maintaining SOV data, especially as insurer expectations continue to increase.”
Early stages of Data Science
The manner in which our species collects data today is analogous to the ways in which Paleolithic people obtained their food some 13,500 years ago. Simply put, humans today are still in their hunter-gatherer phase when harvesting data.
With the advent of Machine Learning and Artificial Intelligence, however, we’re at an inflection point, learning how to farm the data we need. Still, we’ve only just scratched the surface of this new era in data evolution.
In Insurance pricing and Underwriting, the adoption of AI and ML to streamline day-to-day operations is subpar at best. SOV scrubbing was a bottleneck in the underwriting process three decades ago and is still today. In 2022, “77% of risk professionals are negatively impacted by the process of managing and maintaining SOV data, especially as insurer expectations continue to increase.” quotes Archipelago.
Despite attempts from several InsureTech firms, the AI & ML solutions available today are mainly effective on the more standard formats, with some obligatory mundane pre-processing. Yet, most SOVs are still cleaned manually by offshore teams. Why?
This is because there is currently no way to know if an SOV scrubber was 100% accurate without comparing the input and output, line by line, thereby defeating the purpose of using an automated solution in the first place. This strongly indicates that existing SOV solutions are ineffective against the most non-standard SOVs.
In the past, numerous Insurance brokers simultaneously tried to establish their own SOV standards. As a result, we now have hundreds of unique standard formats. The lack of conformity, however, resulted in a majority of the SOVs seen by Underwriters today remaining non-standard. All existing solutions are not Excel-native. Not to mention that an XLSX file is bound to be misunderstood when read as any form of TXT file.
An Excel-native solution with AI & ML capabilities will most effectively sift non-standard data. But, it’s not enough to have a solution that just sorts and organizes your data. What is really needed is a tool that can validate and augment incomplete or inaccurate data. Such a solution would truly alleviate the frustrations experienced by 77% of Risk Managers in 2022 and will be the first step in getting us to the Late Modern era of Black Mirror Insurance.