About the Insurance Company
- One of the largest auto insurance companies in the U.S.
- $35B+ revenue in 2021
- 45,000+ employees
The company, one of the largest auto insurers in the world, processes tens of thousands of claims each day. Most claims begin with a First Notice of Loss (FNOL), which is the initial report made to an insurance provider following a loss, theft, or damage of an insured asset. Typically, these reports contain specific information including, but not limited to, date and time of theft or damage, police report (if one was filed), location of the incident, and a personal or eye-witness account of the loss.
The First Notice of Loss is not only important to the claims process lifecycle, but also to the insurance customer lifecycle. It is critical that all pertinent details are accurately captured and recorded from an FNOL both to ensure accurate claims processing and to satisfy the expectations of insurance customers—who likely had a recent, emotional experience prior to filing their insurance claim.
In an effort to further optimize operations, the company sought to automate data extraction from FNOLs. Due to the high volume of FNOL submissions, as well as variability in the documents themselves, manual processing proved to be a time consuming, expensive, and error prone step in their FNOL workflow. From the formatting used to write dates and addresses to hand written descriptions of how a loss transpired, FNOL reports are not only a challenge for human workers, but also for artificial intelligence (AI).
The company experimented with in-house and third-party solutions for automated document processing, but struggled to achieve highly accurate data extraction and levels of automation.
The company sought a FNOL report processing automation solution that would:
Super.AI’s Intelligent Document Processing (IDP) offered the ideal mix of features and performance to resolve this FNOL report automation challenge. Our next-generation IDP leverages the latest AI and Data Processing Crowd to extract information from any FNOL report (or other document) quickly and with guaranteed quality. This is accomplished by leveraging a pool of on-demand, crowdsourced human workers to train and improve the machine learning models used to for information extraction.
Our solution was able to process more than 99% of the FNOL reports submitted to the insurance company with guaranteed accuracy. This was made possible thanks to the flexible nature of our IDP offering, which is built on top of a unified AI platform for Unstructured Data Processing (UDP). The auto insurance company quickly realized additional opportunities once their data was made available to the super.AI platform, and is now pursuing several additional use cases.