Monthly Archives: December 2016

Text Mining for Claims Management in Insurance

Writing in the Claims Journal, a magazine dedicated to the claims industry, Judith Vaughan and Michael .W. Elliott, experts in claims and risk management, writes how text analytics is helping claims industry in claim processing, fraud detection and in increasing efficiency in claims process.

The authors quotes a report from Accenture which compares the revolution wrought by Big Data to business process to the way internet revolutionized lives in the 1990s. They quote from this same report to say 83% of the enterprises envisage using big data to gain business advantage.

Writing next on how insurance industry especially the claims department is in the forefront of development and deployment of data analytics to increase process efficiency, fraud detection and in claim assortment process. The major challenge for deploying this data analytical process is the prohibitive costs since only about 15% of the data may be converted into actionable information. And the other 85% data which is available in unstructured form is difficult to decipher and extract any useful business information. And this percentage jumps even higher in the claims department with witness reports, claimant statements and other notes being verbose and unstructured in nature.

The insurance industry and the data engineers are looking to harness Text mining or Text analytics to extract valuable data from this 85% of unstructured data. As we know text mining in its basic level is scanning for keywords or phrases and looking for relationships within data but with advances in natural language processing (NLP) and decision logic, it now has the capability for sentiment analysis thru’ which enterprises can discover customer’s opinion about a particular product. In its application for the insurance industry, text mining algorithm can scan the claimant’s social media content in real time to validate information provided in the claims.

Text mining also helps in streamlining the claims processing, thereby increasing process efficiency, and it can also be useful in the development of new products and overcome lacunae in the present offerings.

Another major application of text mining in the insurance industry is in the identification of fraud potential, as majority of frauds are known to happen during claims, Text mining   for example may red flag an instance where several claimants are using the exact same words or sentences in their claims, a sure sign of suspicious activity.

In this way and in many other ways insurers can harness text mining and other data analysis tools to increase process efficiency and bring innovation in terms of product offering or market reach.

For more on this, visit: