An Overview of Approaches for Converting Textual Descriptions to BPMN
An Overview of Approaches for Converting Textual Descriptions to BPMN
Fatemeh Ebrahimi1
1) Isfahan Municipality Employee
IOI Code
:
AAHD-ABKDZF
Publication :
National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
Converting Textual Process Descriptions into BPMN is one of the key challenges in process engineering and enterprise systems analysis. The importance of this problem stems from the fact that many real-world business processes are initially documented not as formal process diagrams, but rather as procedural guidelines, user narratives, operational scenarios, or semi-structured textual documents. In such settings, the accurate transformation of textual descriptions into process models can facilitate process analysis, standardization, and automation.
Nevertheless, this task remains inherently complex and multidimensional due to linguistic ambiguity, omitted implicit information, semantic dependencies, verb polysemy, and the difficulty of identifying exceptions and special-case behaviors.
This paper reviews the major approaches to transforming textual process descriptions into Business Process Model and Notation (BPMN), drawing on both seminal studies and recent advances in the field. The review shows that the evolution of this research area began with rule-based and linguistic analysis methods, progressed toward semi-automated approaches and interactive tools, and, in recent years, has shifted toward hybrid frameworks, incremental model generation, and multidimensional evaluation methodologies driven by the emergence of large language models (LLMs).
The findings indicate that no existing approach has yet achieved a level of maturity capable of consistently ensuring structural accuracy, semantic accuracy, explainability, and ease of use simultaneously. It can therefore be argued that the future of this field lies neither in relying exclusively on rigid rule-based techniques nor in delegating the entire modeling task to general-purpose large language models.
Keywords : bpmn textual descriptions process mining natural language processing large language models