Year: 2026
Type: article
Abstract: Financial processes must be managed in alignment with governance, risk, and compliance (GRC) requirements and audited in accordance with reporting guidelines and standards. Financial process mining addresses these requirements by reconstructing and analysing journal entries due to domain-specific quantitative and qualitative aspects. Event log data derived from enterprise resource planning (ERP) systems are underspecified and, consequently, do not necessarily reflect financial business processes accurately. To address this challenge, the paper proposes the idea of using a fine-tuned large language model (LLM) to interpret and classify journal entries. This fine-tuned LLM is intended to augment a rule-based algorithm, enabling the accurate specification of financial process activities based on posted journal entries. The paper outlines the planned methodological approach for the research and initial work items that have been developed.
Authors: Laerum, Niklas-Kim, Nüttgens, Markus
Institution: Universität Hamburg
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7162479823