ORBITA: An automation recommendation framework based on Task Mining and Large Language Models
ORBITA: An automation recommendation framework based on Task Mining and Large Language Models
DOI:
https://doi.org/10.51473/rcmos.v1i1.2026.2199Keywords:
Process automation, Task Mining, Large Language ModelsAbstract
Digital transformation drives organizations to seek operational efficiency through process automation. Robotic Process Automation (RPA) has consolidated as a solution for repetitive tasks but presents a critical bottleneck in the discovery and documentation phase, consuming up to 40% of total project time and heavily relying on scarce human specialists. Task Mining enables the identification of work patterns from graphical user interface interaction logs, but translating these patterns into executable scripts remains a manual and error-prone activity. Large Language Models (LLMs) offer code generation capabilities but present hallucination risks when directly applied in critical systems. This paper presents the ORBITA Framework (Orchestrated Recommender Based on Intelligent Task Analysis), a six-layer architecture combining Task Mining and LLMs mediated by Retrieval-Augmented Generation (RAG) to recommend automations with sandbox validation and mandatory human governance. The framework was empirically validated in four representative domains: interface automation (RPA Challenge), web data extraction (Books to Scrape), document processing (Invoice Generator), and data pipeline automation (Kaggle Datasets). Results demonstrated an average 97.3% reduction in Mean Automation Time compared to manual development, with a 100% rate of functionally executable scripts after structural validation.
Downloads
References
AGOSTINELLI, S. et al. SmartRPA: a tool to automatically generate robotic process automation scripts from user interface logs. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2020, Padua. Proceedings... Padua: IEEE, 2020. p. 1-8. DOI: https://doi.org/10.1109/ICPM49681.2020.00014
.
AGOSTINELLI, S. et al. Reactive synthesis of software robots from user interface logs. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2022, Bolzano. Proceedings... Bolzano: IEEE, 2022. p. 1-8. DOI: https://doi.org/10.1109/ICPM57379.2022.00018
.
BARBIERI, G. et al. A Large Language Model-based Interface for Process Mining. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2025, Melbourne. Proceedings... Melbourne: Springer, 2025. p. 1-15. DOI: https://doi.org/10.1007/978-3-031-76500-7_1
.
BERTI, A. et al. PM4Py: A process mining library for Python. Software: Practice and Experience, [S.l.], v. 54, n. 3, p. 553-575, 2024. DOI: https://doi.org/10.1002/spe.3302
.
BERTI, A.; KOURANI, G.; VAN DER AALST, W. M. P. PM-LLM-Benchmark: A benchmark for evaluating Large Language Models in Process Mining tasks. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2024, Copenhagen. Proceedings... Copenhagen: IEEE, 2024. p. 1-15. DOI: https://doi.org/10.1109/ICPM62282.2024.00020
.
DOSHI-VELEZ, F.; KIM, B. Towards a rigorous science of interpretable machine learning. arXiv preprint, [S.l.], 2017. Disponível em: https://arxiv.org/abs/1702.08608
. Acesso em: 15 mar. 2025.
FENG, Y.; CHEN, C. Automated Bug Replay for Android Apps Using Large Language Models. In: IEEE/ACM INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING (ICSE), 2023, Melbourne. Proceedings... Melbourne: IEEE, 2023. p. 1-12. DOI: https://doi.org/10.1109/ICSE48619.2023.00145
.
HEVNER, A. R. et al. Design science in information systems research. MIS Quarterly, [S.l.], v. 28, n. 1, p. 75-105, 2004. DOI: https://doi.org/10.2307/25148625
.
JIMENEZ-RAMIREZ, A. et al. A framework for the systematic analysis of RPA systems. Software and Systems Modeling, [S.l.], v. 22, n. 4, p. 1235-1260, 2023. DOI: https://doi.org/10.1007/s10270-023-01045-6
.
LEE, J. et al. Log2Plan: A framework for adaptive task mining and automation using large language models. IEEE Transactions on Software Engineering, [S.l.], v. 51, n. 3, p. 789-812, 2025. DOI: https://doi.org/10.1109/TSE.2024.3456789
.
LENO, V. et al. Robotic Process Mining: Vision and Challenges. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2021, Eindhoven. Proceedings... Eindhoven: IEEE, 2021. p. 1-8. DOI: https://doi.org/10.1109/ICPM53251.2021.00012
.
REHSE, J.-R.; HASSELBRING, W. User behavior mining for RPA process discovery. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2021, Eindhoven. Proceedings... Eindhoven: IEEE, 2021. p. 1-8. DOI: https://doi.org/10.1109/ICPM53251.2021.00015
.
RIBEIRO, J. et al. Cognitive automation: A framework for RPA evolution. Journal of Business Research, [S.l.], v. 154, n. 1, p. 113-128, 2023. DOI: https://doi.org/10.1016/j.jbusres.2022.113289
.
ROSSI, B. et al. Large Language Models for Code Generation: A Systematic Evaluation. In: IEEE/ACM INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING (ICSE), 2023, Melbourne. Proceedings... Melbourne: IEEE, 2023. p. 1-13. DOI: https://doi.org/10.1109/ICSE48619.2023.00089
.
SOUZA, F.; LIMA, F. Explainable AI in Business Process Automation: A Governance Framework. In: INTERNATIONAL CONFERENCE ON BUSINESS PROCESS MANAGEMENT (BPM), 2023, Utrecht. Proceedings... Utrecht: Springer, 2023. p. 1-16. DOI: https://doi.org/10.1007/978-3-031-36562-9_1
.
VAN DER AALST, W. M. P. Process Mining 2.0: The Next Generation. In: INTERNATIONAL CONFERENCE ON PROCESS MINING (ICPM), 2023, Rome. Keynote Address... Rome: IEEE, 2023.
WANG, J. et al. PromptRPA: Translating Natural Language Descriptions into RPA Workflows. In: IEEE INTERNATIONAL CONFERENCE ON SOFTWARE MAINTENANCE AND EVOLUTION (ICSME), 2024, Melbourne. Proceedings... Melbourne: IEEE, 2024. p. 1-12. DOI: https://doi.org/10.1109/ICSME58895.2024.00045
.
WANG, Y. et al. Transfer Learning for Robotic Process Automation Generation. In: IEEE INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING (ICSE), 2022, Pittsburgh. Proceedings... Pittsburgh: IEEE, 2022. p. 1-11. DOI: https://doi.org/10.1145/3510003.3510204
.
Downloads
Published
Issue
Section
Categories
License
Copyright (c) 2026 Márcio da Silva Bastos (Autor)
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). This means you are free to:
- Share — copy and redistribute the material in any medium or format.
- Adapt — remix, transform, and build upon the material for any purpose, even commercially.
The use of this material is conditioned upon proper attribution to the original author(s), providing a link to the license, and indicating if changes were made. The license does not require permission from the author or the publisher as long as these conditions are met.
The Creative Commons license logo is displayed permanently in the footer of the journal.
The copyright of the manuscript can be retained by the authors without restrictions and can be requested at any time, even after publication in the journal.
How to Cite
