TY - JOUR
T1 - A modular chatbot framework for assisting workers in diverse production tasks
AU - Kern, Thomas
AU - Stang, Julian
AU - Milde, Michael
AU - Hofer, Andreas
AU - Streibel, Lasse
AU - Zaeh, Michael F.
N1 - Publisher Copyright:
© 2024 The Authors.
PY - 2024
Y1 - 2024
N2 - In modern production environments, workers face increasingly demanding conditions that require skills such as navigating through extensive information repositories or operating complex machinery. This paper proposes a modular chatbot framework to address this challenge, aiming to assist workers in diverse production tasks. The study is based on a systematic literature review to gain an understanding of the current state of the art in chatbot technology, with the goal of identifying research needs in this field. Subsequently, interviews and an industrial use case were conducted with potential application partners to uncover the specific requirements of chatbots in production companies. These insights led to the conceptualization of a modular chatbot framework using modern natural language processing techniques tailored to integrate a variety of production use cases. In order to validate the presented framework's adaptability, it was instantiated with three distinct production use cases that simulate real-world operations. The resulting chatbot underwent testing and evaluation by applying a set of test metrics to conversation scenarios that simulate user requests to the chatbot. The outcome of this study presents a modular chatbot framework that allows for the integration of diverse production use cases as individual modules. Findings from the pilot implementation suggest that the proposed framework enables the development of chatbots to support workers in diverse production use cases, indicating the framework's adaptability. While the pilot implementation highlights notable benefits, certain limitations are acknowledged, including insufficient precision in the chatbot's responses and data security considerations. Future research should focus on fine-tuning the underlying Natural Language Processing techniques to enhance precision and on expanding the framework's integration into existing production workflows to assess its real-world effectiveness.
AB - In modern production environments, workers face increasingly demanding conditions that require skills such as navigating through extensive information repositories or operating complex machinery. This paper proposes a modular chatbot framework to address this challenge, aiming to assist workers in diverse production tasks. The study is based on a systematic literature review to gain an understanding of the current state of the art in chatbot technology, with the goal of identifying research needs in this field. Subsequently, interviews and an industrial use case were conducted with potential application partners to uncover the specific requirements of chatbots in production companies. These insights led to the conceptualization of a modular chatbot framework using modern natural language processing techniques tailored to integrate a variety of production use cases. In order to validate the presented framework's adaptability, it was instantiated with three distinct production use cases that simulate real-world operations. The resulting chatbot underwent testing and evaluation by applying a set of test metrics to conversation scenarios that simulate user requests to the chatbot. The outcome of this study presents a modular chatbot framework that allows for the integration of diverse production use cases as individual modules. Findings from the pilot implementation suggest that the proposed framework enables the development of chatbots to support workers in diverse production use cases, indicating the framework's adaptability. While the pilot implementation highlights notable benefits, certain limitations are acknowledged, including insufficient precision in the chatbot's responses and data security considerations. Future research should focus on fine-tuning the underlying Natural Language Processing techniques to enhance precision and on expanding the framework's integration into existing production workflows to assess its real-world effectiveness.
KW - Chatbot
KW - Conversational Assistance
KW - Natural Language Processing
KW - Production Industry
UR - http://www.scopus.com/inward/record.url?scp=85213063248&partnerID=8YFLogxK
U2 - 10.1016/j.procir.2024.10.238
DO - 10.1016/j.procir.2024.10.238
M3 - Conference article
AN - SCOPUS:85213063248
SN - 2405-8971
VL - 58
SP - 1268
EP - 1275
JO - IFAC Proceedings Volumes (IFAC-PapersOnline)
JF - IFAC Proceedings Volumes (IFAC-PapersOnline)
IS - 27
T2 - 18th IFAC Workshop on Time Delay Systems, TDS 2024
Y2 - 2 October 2023 through 5 October 2023
ER -