Using Large Language Models to Generate Process Knowledge from Enterprise Content

Franzoi, S.; Delwaulle, M.; Dyong, J.; Schaffner, J.; Burger, M.; vom Brocke, J.


Zusammenfassung

Large language models (LLMs) have disrupted knowledge work in many application areas. Accordingly, the Business Process Management (BPM) community has started to explore how LLMs can be leveraged, resulting in a variety of promising research directions across the BPM lifecycle. Despite rapid adoption in practice and strong research interest, however, little is known about the actual design of BPM systems that leverage LLMs in organizational contexts. In this paper, we report on design science-based research in collaboration with a large multinational company to design a BPM system that leverages LLMs for process knowledge extraction from diverse enterprise content. Based on the development of our prototype, we observe that LLMs provide the means to organize and generate process knowledge independent of specific forms of representation. We present a conceptual framework that describes the role of LLMs in generating process knowledge from diverse input formats and, in turn, making it available in diverse output formats via prompting, resulting in representation-agnostic process knowledge. We also highlight implications of our study for BPM research and practice.

Schlüsselwörter
Large Language Models; Process Knowledge; Generative Artificial Intelligence; Business Process Management



Publikationstyp
Forschungsartikel in Sammelband (Konferenz)

Begutachtet
Ja

Publikationsstatus
accepted / in press (not yet published)

Jahr
2024

Konferenz
International Conference on Business Process Management (BPM)

Konferenzort
Krakow

Buchtitel
Lecture Notes in Business Information Processing (LNBIP)

Herausgeber
Resinas M., van der Aa H., del Río-Ortega A. & Leopold H.

Erste Seite
1

Letzte Seite
12

Verlag
Springer Publishing

Ort
Cham