Name of participant: Jan Gronewald
Project’s name: Digitaler Assistent für das Journal Entry Testing (DA4JET)
Project description:
As part of the annual audit for financial statements, one of the account’s central tasks is to doublecheck systemically the financial accountings data to see if significant misstatements and misrepresentations—whether due to human mistakes or with intent to defraud—have been made. The accuracy of this examination builds the foundation of trust into companies’ annual audits, and simultaneously secures the economy’s stability. In the face of the examination becoming more demanding—due to exponentially growing data volume and modern ERP- and accounting systems gaining more complexity—an optimized use of resources is needed. Simultaneously, new data analysis approaches are possible due to technical progress and cost-efficient access to highly productive computing capacities. However, particularly processes based on machine learning (ML), such as deep learning, haven’t been implemented into examination processes comprehensively.
The aim of the project DA4JET (Digital Assistant for Journal Entry Testing) is to research on and conceptualize a sociotechnical system that combines traditional rule-based test procedures with the controller’s experiential and contextual knowledge by merging it together with a targeted application of ML methods into a hybrid approach. Therefore, ML methods are being analyzed to understand how they can contribute to detecting anomalies within accounting data and how the expert’s knowledge of users can ideally support ML models. Central research questions are related to the selection and adaptation of applicable ML methods, the integration of these technologies in already existing examination processes, as well as the outcomes’ evaluation. This project offers specific potentials to improve the preciseness and economic efficiency of the annual audit.
Software Campus Partner: German Research Center for Artificial Intelligence (DFKI) and Datev eG
Implementation period: 01.04.2025 – 31.03.2027
