As part of an article about the use of analytics in the insurance claims process, Insurance & Technology asked me about areas of claims that can benefit from analytics—but for various reasons, those analytics aren’t being implemented.
For example, analytics could greatly improve organizational performance, notably injury treatment, injury management and customer attrition response. These days, it isn’t enough to ask whether a claim is fraudulent. Insurers also need to ask if they are following the right intervention sets to drive the best possible outcome.
I think the primary challenges in implementing analytics in claims are due to poor or outdated infrastructure. For analytics to be effective, a few things need to be in place:
- Data quality must be high. Otherwise even the most robust, real-time model will be working with incomplete or ambiguous data.
- Insurers must be able to consistently access and match data from multiple sources, such as underwriting, policy and vendor files, as well as customer files.
- Insurers must have systems in place to quickly and efficiently access and analyze data. Otherwise, it is very difficult to invoke real-time analytics during claims processing.
Finally, the predictive models that are currently used in claims aren’t transaction-intensive. Rather than interacting with core systems and major data sources, current models are operating on the fringes of the business. To truly leverage the benefits of analytics in the claims process, insurers must integrate the models with real-time transaction processing.