Accessing complex information to prosecute fraud

A software solution combining artificial intelligence, machine learning and pattern recognition software helped the Riverside County District Attorney’s Office in California prosecute seven people in a scheme allegedly involving the fraudulent billing of $98 million.

In early June, Riverside County District Attorney Michael Hestrin announced that seven people were indicted for insurance fraud and conspiracy resulting from one of the largest workers’ compensation fraud investigations in the county. The indictments, which stemmed from a joint investigation by the Riverside County District Attorney's Office and the California Department of Insurance, alleged that $98 million was fraudulently billed, resulting in $12.4 million being paid by 18 insurance companies defrauded in the scheme.

The prosecution was able to access complex data in millions of handwritten and unstructured documents through the use of smartC big data analytics technology from Infinilytics, which incorporates handwriting recognition and document indexing toolkit A2iA DocumentReader from A2iA. Once the data was accessible, according to A2iA, the software was able to analyze the content and help the district attorney file charges against the seven people accused of filing false workers’ compensation claims.

A2iA reports that the integration of its A2iA DocumentReader with smartC, combined with Infinilytics’ predictive modeling, enabled the District Attorney’s Office to gain a clear and concise pathway to winning its case. With smartC, the prosecutors were said to be able to gather data and virtually remove all manual document interaction, including document sorting and search. A2iA says that the automated solution allowed law enforcement to:

• automatically sort all seized documents, including the intricate and unstructured documents;
• locate and extract the data automatically, including cursive handwritten paper notes, columnar information and alpha and numeric content;
• quickly analyze the extracted data and provide it to the predictive data model; and
• issue arrest warrants and file charges based on the results.