Year: 2024
Type: review
Study design: Systematic Review
Abstract: Process discovery is a fundamental phase of Business Process Management that aims to document the current state of organizational processes. It employs various methods such as interview-based, workshop-based, and evidence-based approaches to gather information, model processes, and ensure model quality. Generative AI (GAI) is emerging as a significant area of interest for organizations, promising innovative solutions for process discovery. The present study through a systematic literature review investigates the intersection of process discovery and GAI, providing an overview of GAI applications in this context. The study aims to understand the problems GAIs address, the types of generative technologies employed, and its applications within different discovery methods. To achieve this, an extensive search was conducted across academic databases, covering generative models and general deep learning models used for generative tasks. The results indicate that general deep learning models are utilized for anomaly detection and reconstruction, while Large Language Models and GPTs are used within a range of tasks, from information gathering to validating model quality, due to their proficiency in processing natural languages and long sequences, such as event logs. The literature shows a significant representation of LLMs and GPTs, especially in recent years following the advent of ChatGPT, highlighting the growing importance of these technologies in process discovery. In conclusion, this review demonstrates that GAI usage in process discovery can lead to promising results in a diverse range of tasks and hope that this review inspire the adoption of GAI methods and foster more interest for future research and developments in this area.
Author: Schmidt, Luiza
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7047355350