In May 2019, Cheryl Leong, Head of Fraud Analytics and Data Management at Moss & Associates, a mid-sized New York accounting firm, was tasked with fraud detection in annual reports.

Besides helping clients present financial information to stakeholders, accounting firms had to ensure there were no misrepresentations.

The incidence in companies reporting material falsehoods had risen in recent years and regulators were pushing accounting firms to detect those instances earlier.
Companies had been using qualitative text to mislead stakeholders of their financial wellbeing.

Leong was working on a data analytics platform that would replace the time consuming task of manually going over executive statements and management discussion & answers (MD&A) sections. Several text mining techniques were available within the system to break text down for classification.
With the fraud detection tool ready for launch, she wondered whether she had adequately addressed the challenges of analysing text in annual reports.

What steps should she programme the tool to take? How successful it would be? The case study provides students with an opportunity to discuss a number of analytical concepts related to applying text mining to solve a business problem in finance.

Students who have studied the case should be able to justify important reasons for initiating the analytics project; identify and solve the possible data challenges; develop the high-level solution design; discuss analytics project risks, benefits and hurdles?