What if you could transform your underwriting process to have complete risk data gathered from various insurance documents and delivered to your underwriters in as little as two hours? Not only would the data be provided quickly, but it would be 98% accurate. With Groundspeed, this transformation is possible. This blog will highlight the importance of loss runs and how Groundspeed uses human-in-the-loop AI to transform the commercial underwriting process, so carriers can unlock the value in their unstructured data and evaluate risk faster.

Loss runs are one of the most critical components of the underwriting review process and are typically one of the first items reviewed. High loss frequency or profitability concerns over the past 3-5 years may indicate future loss trends. A typical insurance carrier will receive and process thousands of loss runs weekly. Given the complexity of the documents, commercial underwriters spend hours and sometimes even days reviewing loss runs before a quote may be provided.

Why Is a Loss Run Important?

A loss run document lists the claims made on an insurance policy over a period of time. This document provides an account of the losses experienced (and reported) by the insured, along with details such as:

  • Description of the event 
  • Amounts paid/reserved by the insurer, including expenses
  • Current status of the claims 
  • Date and location where the loss occurred  
  • Parties involved

The historical losses experienced by an insured and described in a loss run are used by underwriters to inform expectations of future losses, which are used to assess the potential profitability of insuring them. 

The underwriter reviewing a potential insured will use the information in a loss run to decide how to price a new policy, limit or exclude different types of claims from coverage, and whether to offer a policy at all. It is essential for underwriters only to approve or renew applicants that should yield long-term profit for their company. The hard truth, however, is that reviewing loss runs is an arduous and often manual process. Every insurance carrier’s loss runs come in different formats, with data presented in various ways. Furthermore, loss runs are oftentimes sent as PDFs or JPEGs with no easy way for the underwriter to conduct an evaluation without manually reviewing and re-typing the information. After all of that time has been spent, the underwriter may need to decline the submission for unacceptable loss history anyway. Talk about a waste of time! 

Why Is Digitizing Loss Runs Challenging?

There is an incredible diversity of loss-run formats exchanged within the insurance industry, with each insurer producing a unique array of presentations to share information about the different products and lines of business they offer. 

While there are regulations in place with minimum requirements on data points to include, each format of the document will present the same essential information organized in a unique way, with different levels of granularity and arbitrary nested structures, labeled using ambiguous terms that carry different meanings depending on the company producing the document.

Since these documents are generally used to convey information to their competitors, there is little incentive for standardization among insurance carriers. The formats are highly divergent and idiosyncratic to each carrier and each insurance product they offer. However, the underwriting process at each carrier is dependent on bringing this loss data into a consistent schema to allow for fast and accurate analysis and pricing decisions.

In order to bring these different presentations into a desirable standard format, Groundspeed has developed a robust set of tools to solve the various problems present in digitizing loss runs.

These include:

  • An integrated state-of-the-art OCR preprocessing that extracts the words and basic structural elements of PDFs and similar documents into a consistent machine-readable representation.
  • An extensive library of hundreds of formats observed “in the wild” and effective systems to rapidly identify documents and either associate them to known formats or add them to the catalog.
  • Document parsing models that use the known data structure and naming schemes for each format to automatically identify and classify the data points in new documents with that format, as well as an efficient pipeline for constructing new models.
  • A comprehensive universal data schema that encompasses the entire space of information relevant to the commercial insurance domain and systems to accurately translate and structure the data points extracted from documents into this common schema.
  • Extensive business logic to infer data points that are implied but missing from the original documents, as well as normalization of the contents of data fields into standard data types and formats.
  • An integrated quality control and human-in-the-loop system to handle exceptions and ensure a very high standard of reliable and complete data outputs.

How Groundspeed Can Help

Groundspeed’s Loss Run Automation Technology captures, structures, and delivers complete data from loss runs in an enriched format that is easy for underwriters to evaluate. With Groundspeed, unstructured data from 8+ document types for 12+ commercial lines of business may be extracted and delivered to the underwriter in as little as two hours, allowing them to do what they do best: underwrite! 

Give your underwriters the good news: they no longer need to spend hours manually typing trapped information from loss runs into a spreadsheet or rating system. With Groundspeed’s Artificial Intelligence platform, that work is done automatically! It’s time to win more business, improve loss ratios, and increase efficiencies with Groundspeed’s AI platform!

If you would like to learn more about Groundspeed’s human-in-the-loop solution, schedule a call with our team today or request a meeting in person with our team at InsureTech Connect Vegas in September. Let us partner with your company and help you unlock the value in your unstructured data.

 

This blog was written in collaboration by:

    • Evan Daywell – Evan is the Senior Manager of Document Processing and leads the automation team at Groundspeed. His team works on building systems to accurately extract information from a variety of insurance documents to transform the data into a consistent format.
    • Melissa Lakin – Melissa is a Customer Success Manager. Melissa enjoys working with our clients to make sure we are meeting expectations and reaching their goals. Before Groundspeed, Melissa was a market-facing underwriter for over four years.
    • John Swasey –John is an Enterprise Customer Success Manager at Groundspeed. He works closely with clients to make sure our products are successfully transforming their underwriting processes. Before Groundspeed, John worked in the commercial underwriting space for over ten years.

Leave a Reply

By clicking "Continue" or continuing to use our site, you acknowledge that you accept our Privacy Policy. We also use cookies to provide you with the best possible experience on our website.
Continue