Scholarlink

Business Process Management (BPM) is important for organizations, with process modeling being a key component. However, creating accurate and consistent models is challenging and time-consuming. As large language models (LLMs) offer potential solutions, this article explores using LLMs to generate business process models that follow company internal style guides. Interviews with process engineers at a company provided insights into current practices and challenges. A prototype using OpenAI’s GPT-4 API was developed based on these insights. This tool enables non-experts to generate business process models in BPMN 2.0 notation through a user-friendly interface. Evaluation using the PET dataset demonstrated that the LLM-generated models are comparable to those created by human experts regarding recall and precision.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Large Language Models for Democratizing Business Process Modeling: BPMN Model Generation and Style Guide Adherence

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Seline Wenger,
  • Maja Spahic-Bogdanovic,
  • Andreas Martin

摘要

Business Process Management (BPM) is important for organizations, with process modeling being a key component. However, creating accurate and consistent models is challenging and time-consuming. As large language models (LLMs) offer potential solutions, this article explores using LLMs to generate business process models that follow company internal style guides. Interviews with process engineers at a company provided insights into current practices and challenges. A prototype using OpenAI’s GPT-4 API was developed based on these insights. This tool enables non-experts to generate business process models in BPMN 2.0 notation through a user-friendly interface. Evaluation using the PET dataset demonstrated that the LLM-generated models are comparable to those created by human experts regarding recall and precision.