How Policy Admin Has Evolved Like the Automobile
February 10, 2016 1:07 AM
With policy administration replacement by insurance companies at a fever pitch, insurance technologists can learn from the auto industry in terms of what to expect as core competencies evolve. Thats what Celent analyst Tom Scales asserts in the report Mainframe-based Core Insurance Systems: On the Road to Oblivion? Read on for his insight.
If you think of the first car as the barely-get-it-down-the-road model T kind of thing, the parallel to that in insurance is the CFO II system, which was first designed in 1955, he says. And variations of that are still in production today. But thats not what insurers need for the future, he explains.
The first cars were built with body-on-frame architecture, Scales explains, with a body and a chassis bolted together during assembly. And, the first insurance systems were equally primitive, with CFO II written in assembler, a language that basically communicates on the same level as the machines code instructions.
Eventually, the primitive design of early cars was enhanced with fenders, trunks, and hoods. And, in insurance, companies embraced the COBOL language and fed it into mainframe-based systems using stacks of cards.
But in the 1960s, things changed for the automobile. The new unibody design strategy created more room for passengers and lightened the load of the car. In insurance, this parallels the development of service oriented architecture in the early 2000s, Scales says.
COBOL systems weathered a new input device (the PC) in the 1980s, but the back end, still dependent on batch processing, was too slow for complex financial products or the demands of the Internet. If I wanted to put a new feature on a new product on COBOL, Id have to get a programmer to help, Scales notes.
SOA was the first attempt to blend the COBOL world with the new reality of the Internet. Companies wrapped their old systems in .NET- and C#-powered layers to bring new capabilities to market.
But the problem with SOA is that it doesnt address the underlying issues with the mainframe-based system. In fact, it increased the risk associated with changing the core system functionality, as any change could have a cascading effect on the layers. Scales compares this to automakers effort to market cars at different price points with only cosmetic differences in the grills or paint. This ultimately led to bloated product lines and failure.
In order to better differentiate among products, automakers turned to platform engineering, which meant more granular standardization in areas that arent noticeable by consumers. That gave them more flexibility to make changes.
So what can insurers do thats similar? Abandon the mainframe, Scales says. He predicts cloud-based systems will take over insurance in the near future as insurers get away from outdated systems and realize that its the capabilities that they need more than the ownership of the infrastructure.
The core competency of an insurer isnt running a data center -- theyre all good at it only because they had to be, Scales says. Rather than slow evolution of the old systems, whats ultimately going to revolutionize is moving from the data center into the cloud, he adds.