Leveraging big data to minimise claims fraud

Claims fraud is a major drain on insurance carriers.

Many agency estimates state that around 10 per cent of all property and casualty insurance claims are fraudulent equating to billions of dollars and a significant impact on carrier profitability. As fraud has become more sophisticated with the advent of newer technologies, with sophisticated criminals and organised crime rings in many markets, loss ratios have deteriorated. And, if fraud losses are not addressed properly, higher premiums, customer base erosion, and ultimately insolvency could result.

To combat fraud, insurers currently are using conventional techniques including training employees, setting up Special Investigation Units (SIU) and deploying specialised software programmes. However, times have changed and the magnitude of fraud’s impact on carriers’ bottom line demands that they step up their attack.

Insurers need a multi-pronged approach to prevent fraud and leveraging big data is their first line of defence.  First, insurers should analyse lifecycle checkpoints during underwriting, claim intake and adjudication/subrogation processes to prevent fraud.  They also could examine past claims histories to identify patterns of fraud for appropriate flagging and investigation. These analyses are done using a blend of business rules and database searches.

Big data is the answer to addressing the rapid change in the velocity, volume and variety of data coming into insurance companies. For example First Notice of Loss (FNOL), which was conventionally a telephone call, is now coming through multiple channels such as email, fax and SMS.  The insurer’s major challenge is to decipher disparate data sources, channels and formats to find meaningful information and patterns.

Big data technologies provide the capability to handle volume and variety (structured and unstructured) of data throughout the policy lifecycle and flag any deviation for possible fraud. Throughout the claims management lifecycle, insurers can keep receiving live data feeds, such as blogs, tweets and social media posts and scrutinise them to ascertain the veracity of the claim on an ongoing basis. This is a big advantage to carriers who were only using static scoring mechanisms until now to flag a suspicious claim.

Big data solutions also provide analytical capabilities such as inconsistency detection, predictive modelling, data mining, and social network analysis. These capabilities along with high-volume in-memory data transformation and processing capabilities will provide insurers with a state of the art armory to identify probable fraudulent claims.

Big data analytics solutions can be leveraged to analyse large volumes of insurance claims data from multiple channels, multiple sources and in multiple formats using real-time feeds and streaming through high speed networks. Big data analytics engines also offer in-memory computing and processing with configurable algorithms to reveal fraudulent claim patterns on a real-time basis.

Insurers need to take a holistic view of their technology and infrastructure investments in big data technologies. These expenditures must be supported by a strong business case linking the costs to the benefits and validating a positive return on these investments.  Implementing big data solutions is an effective way to combat fraud’s drain on carrier’s profitabilities.


Apparao Chathurvedula is Manager – Property and Casualty Center of Excellence, Capgemini Financial Services

+44 (0)203 202 3323