Automating Business Process Modeling: An LLM-Based Approach with Situation-Based Modeling Notation
Resumo
Research Context: Although business process modeling is beneficial for clarifying organizational processes, it remains a task with little automation and requires significant human effort. Scientific and/or Practical Problem: To reduce this effort, research has advanced toward automating the task, mainly focusing on extracting imperative models from textual descriptions. However, imperative models may lead to a single solution that does not necessarily represent the best fit. Proposed Solution and/or Analysis: The declarative Situation-Based Modeling Notation (SBMN) allows modelers to generate multiple modeling structure alternatives, enabling them to choose the one that best fits the context. This work investigates the use of Large Language Models (LLMs) to identify core business process entities and constraints represented in SBMN from textual descriptions. Related IS Theory: This research is grounded in the Task-Technology Fit (TTF) theory, applied to evaluate the adequacy of LLMs in supporting business process analysts during modeling activities. Research Method: A dataset of 133 textual descriptions paired with SBMN models was employed. Each description was processed by three medium-scale LLMs and three extra-large LLMs using zero-shot prompts to identify domain situations and Active Flow Objects. The outputs were semantically compared with the corresponding SBMN models to measure their similarity. Summary of Results: Large-scale LLMs (e.g., Claude Sonnet and Qwen3-80B) were able to identify flow objects and situations in SBMN with relevant accuracy, demonstrating the feasibility of their use to support the modeling task. Contributions and Impact to IS area: This research provides: (i) empirical evidence highlighting the strengths and limitations of current LLMs in the process modeling domain; (ii) a dataset containing aligned pairs of textual descriptions and SBMN models; and (iii) a semantic–structural similarity metric enabling the quantitative evaluation of SBMN component extraction accuracy from text.
Referências
Araujo, R. (2016). Information systems and the open world challenges. In I GranDSI-BR: Grand Research Challenges in Information Systems in Brazil (2016–2026), chapter 4, pages 42–51. Brazilian Computer Society (SBC), Porto Alegre, Brazil.
Avila, D. T., de Moura, V. C., and Thom, L. H. (2023). Using machine learning to classify process model elements for process infrastructure analysis. In Proceedings of the XIX Brazilian Symposium on Information Systems, pages 45–52.
Balabko, P., Atkinson, C., and Tucci, C. (2005). Situation-Based Modeling Framework for Enterprise Architecture. PhD thesis, École Polytechnique Fédérale de Lausanne (EPFL).
Beerepoot, I., Di Ciccio, C., Reijers, H. A., Rinderle-Ma, S., Bandara, W., Burattin, A., Calvanese, D., Chen, T., Cohen, I., Depaire, B., et al. (2023). The biggest business process management problems to solve before we die. Computers in Industry, 146:103837.
Candido, D., Lima, J., Oliveira, H., and Costa, M. (2024). An annotated dataset for automatic extraction of entities and restrictions from business process models. In Anais do XXI Encontro Nacional de Inteligência Artificial e Computacional, pages 978–989, Porto Alegre, RS, Brasil. SBC.
Costa, M. B. and Tamzalit, D. (2017). Recommendation patterns for business process imperative modeling. In Proceedings of the Symposium on Applied Computing, SAC ’17, page 735–742, New York, NY, USA. Association for Computing Machinery.
De Marneffe, M.-C. and Manning, C. D. (2008). The stanford typed dependencies representation. In Coling 2008: proceedings of the workshop on cross-framework and cross-domain parser evaluation, pages 1–8.
de Santana Candido, D., Alves de Oliveira, H. T., and Costa, M. B. (2025). Unveiling business processes control-flow: Automated extraction of entities and constraint relations from text. In Proceedings of the 27th International Conference on Enterprise Information Systems - Volume 2: ICEIS, pages 771–782. INSTICC, SciTePress.
Dumas, M., Rosa, M. L., Mendling, J., and Reijers, H. A. (2018). Fundamentals of Business Process Management. Springer Berlin Heidelberg, 2 edition.
Ferreira, R. C. B. and Thom, L. H. (2016). Uma abordagem para gerar texto orientado a processo a partir de texto em linguagem natural. In Simpósio Brasileiro de Sistemas de Informação (SBSI), pages 585–588. SBC.
Friedrich, F., Mendling, J., and Puhlmann, F. (2011). Process model generation from natural language text. In Mouratidis, H. and Rolland, C., editors, Advanced Information Systems Engineering, pages 482–496, Berlin, Heidelberg. Springer Berlin Heidelberg.
Kojima, T., Gu, S. S., Reid, M., Matsuo, Y., and Iwasawa, Y. (2023). Large language models are zero-shot reasoners.
Kourani, H., Berti, A., Schuster, D., and van der Aalst, W. M. (2024). Process modeling with large language models. In International Conference on Business Process Modeling, Development and Support, pages 229–244. Springer.
Licardo, J. T., Tanković, N., and Etinger, D. (2024). A method for extracting bpmn models from textual descriptions using natural language processing. Procedia Computer Science, 239:483–490. CENTERIS – International Conference on ENTERprise Information Systems / ProjMAN - International Conference on Project MANagement / HCist - International Conference on Health and Social Care Information Systems and Technologies 2023.
Motta, E., Rajan, T., and Eisenstadt, M. (1989). A methodology and tool for knowledge acquisition in keats-2. In Studies in Computer Science and Artificial Intelligence, volume 5, pages 297–322. Elsevier.
Nivon, Q. and Salaün, G. (2025). Automated generation of bpmn processes from textual requirements. In Gaaloul, W., Sheng, M., Yu, Q., and Yangui, S., editors, Service-Oriented Computing, pages 185–201, Singapore. Springer Nature Singapore.
Pedroso, B., Pereira, M., and Pereira, D. (2025). Performance evaluation of llms in the text-to-sql task in portuguese. In Anais do XXI Simpósio Brasileiro de Sistemas de Informação, pages 260–269, Porto Alegre, RS, Brasil. SBC.
Pesic, M., Schonenberg, H., and Van der Aalst, W. M. (2007). Declare: Full support for loosely-structured processes. In 11th IEEE international enterprise distributed object computing conference (EDOC 2007), pages 287–287. IEEE.
Reimers, N. and Gurevych, I. (2019). Sentence-bert: Sentence embeddings using siamese bert-networks.
Silva de Castro, D., Fantinato, M., and Costa, M. B. (2025). Business process design support with automated interviews. In Proceedings of the 27th International Conference on Enterprise Information Systems - Volume 2: ICEIS, pages 734–745. INSTICC, SciTePress.
Song, K., Tan, X., Qin, T., Lu, J., and Liu, T.-Y. (2020). Mpnet: Masked and permuted pre-training for language understanding.
van der Aa, H., Di Ciccio, C., Leopold, H., and Reijers, H. A. (2019). Extracting declarative process models from natural language. In Giorgini, P. and Weber, B., editors, Advanced Information Systems Engineering, pages 365–382, Cham. Springer International Publishing.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., and Polosukhin, I. (2017). Attention is all you need. Advances in neural information processing systems, 30.
