Fine-Tuning a General-Purpose Large Language Model for Financial Process Mining
2 June 2026, by Viktoria Wrobel

Photo: base.camp
Financial process mining reconstructs business processes based on journal entries. This becomes problematic when system-specific terms lead to unclear or ambiguous activity descriptions. The aim of this project is to solve this problem by fine-tuning a general-purpose large language model (LLM) for financial process mining to correctly interpret and classify journal entries in a business context. Various LLMs are first finetuned and compared in a benchmarking evaluation to identify the best method. The selected LLM is then tested against other general-purpose LLMs to demonstrate the benefits achieved.

