Supernova Award Category

The Problem

Claims and risk management departments at companies across industries are being asked to do a lot more with a lot less today. The volume of claims they must handle continues to grow year after year, yet their resources, both people and technology, have not kept up. As a result, many claims and risk management departments struggle to prioritize claims and make data-driven decisions about how to adjudicate them most effectively. They struggle to answer questions like, “Should we contest this claim or pay it out now?” “Are we more or less likely to win should we contest this or that claim?” “Which claims have the potentially to be the most costly for the company and, therefore, we should put our most experienced claims adjuster on the case?” The result of this challenge is millions of dollars in costs that could otherwise be avoided.

The Solution

Marsh ClearSight knew it was sitting on the client and financial data needed to help its clients better prioritize claims and determine the best actions to take for each claim. It knew it needed a way to analyze all that data to provide insights and recommendations to cleints, but traditional database tools wouldn’t cut it. The company needed a new approach to analyze the petabytes of data – structured and unstructured – to better serve its clients. After evaluating a number of different options, Marsh ClearSight made the decision to adopt Pivotal Greenplum, a massively parallel processing analytical database, to serve as the analytics engine supporting its platform. 

The results

Since implementing Pivotal Greenplum, March ClearSight has significantly improved its ability to predict the likely outcomes of individual claims and to provide data-driven insights and suggestions to clients about how best to proceed with each claim. For example, Marsh ClearSight is now able to accurately predict, within a couple of thousand dollars, the likely settlement outcome of a claim. With that insight, Marsh ClearSight clients can put their best claims adjusters on those cases that are likely to end up in large payments and more quickly adjudicate the less costly claims. The result is significantly better outcomes for clients.

Metrics

We are working on piloting this solution with a select set of clients to measure, refine and improve the predictability of the outcomes. Our goal is to open it up to all clients in an upcoming release.

The Technology

Pivotal Greenplum

Disruptive Factor

Marsh ClearSight has more claims and loss data than any other RMIS (Risk  Management Information System) provider, but did not have a way to harness that data and use it to provide deeper analytical insights to their clients beyond traditional reporting and business intelligence. The other players in the RMIS space also focus on traditional reporting and lack the data that we have. This predictive model is the first example in the industry that will combine traditional structured data with unstructured data to yield actionable insights for our clients. The industry has been asking for this level of analytics but has never had a tool to provide it until now. Clients will be able to optimize their workflows and resourcing to ensure that they are focusing their teams on the claims where they can make the biggest difference and potentially settle the rest of the claims automatically. Ultimately, Marsh ClearSight will look to provide future predictive and prescriptive analytics models to continue to help clients to understand the hidden stories, insights, and trends buried within their data.

Shining Moment

Through our partnership with the data scientists at Pivotal Labs, we were able to build a model that can be applied and trained to support multiple clients and multiple lines of coverage globally. Traditional actuarial models focus on one specific client or one line of coverage only.