過去需要製造零件業客服人力去花大量時間去做重複性的料號資訊查詢，若沒相關經驗新人要花更多時間，隨著AI的普及，本研究旨在探討導入開源低代碼（Low-Code）聊天機器人平台 Botpress 作為產品資訊AI聊天助理的可行性，透過 Botpress平台建構一套針對零件規格、料號查詢之 AI聊天機器人系統。技術層面採用 F1分數進行績效評估，量化分析系統在結構化與非結構化語意識別上的表現；心理接受度層面，則透過 AHP分析決定 TAM3模型中各子因子（如結果可證明性、客觀可用性等）對使用者採納意願的影響權重。

研究結果顯示，在技術績效方面，Botpress 處理結構化數據（如料號與明確規格查詢）展現極高可靠性，F1分數達到可用標準；然而在開放性問題處理上，F1 分數不到可用標準，顯示系統在模糊語意理解上仍有優化空間。在使用者採納意願方面，AHP 分析證實：在「知覺有用性」維度中，「結果可證明性」權重最高；在「知覺易用性」維度中，則以「客觀可用性」為重點。這也反映出製造業使用者對 AI聊天機器人的採納關鍵，在於答案的精準可驗證性與可信任，而非是介面操作的使用上的接受度。

In the manufacturing competent industry, customer service personnel traditionally spend significant time on repetitive part number and specification queries, a challenge that is particularly pronounced for inexperienced staff. Amid the rapid proliferation of AI, this study explores how open-source platforms can be leveraged as 'AI Product Information Assistants' by investigating the feasibility and successful implementation of Botpress. The methodology is two-fold Technically, the F1-score was used to evaluate the system's accuracy in recognizing both structured and unstructured semantics. Psychologically, AHP was applied within the TAM3 framework to identify the key factors influencing user adoption intention.
The research findings indicate that in terms of technical performance, Botpress demonstrates high reliability for structured data (e.g., clear part numbers and explicit specifications), with F1-scores reaching usable standards. However, the performance on open-ended queries fell below these standards, highlighting a need for further optimization in ambiguous semantic understanding.
Besides, the study reveals that for manufacturing users, the adoption of AI chatbots depends heavily on the precision, verifiability, and trustworthiness of the output rather than the simplicity of the user interface. These insights provide a practical reference for enterprises seeking to integrate low-code AI solutions into specialized industrial workflows.

一、外文部分(依作者姓名開頭字母排列)
[1] Accenture Company (2021) "Customer service has had a long, successful run in its traditional role"
[2] C. Tumelero Martins De Andrade, (2022), "Increasing customer service efficiency through artificial intelligence chatbot," Revista de Gestao (2022) 29 (3): 238–251.
[3] Dhruv Kikani, Sunny M. Ramchandani. (2025) "A Comparative Analysis of Automation and Human Interaction in Customer "Support. 2025 IJRTI | Volume 10, Issue 3 March ISSN: 2456-331.
[4] Bryan James Dsouza, Kashif M D, Nagesh K N (2025), Ask me A Customer Service Chatbot, IEEE Xplore Part Number: CFP25AV8-ART; ISBN: 979-8-3315-0967-5
[5] Cui, S., et al. (2017). “A Context-aware approach for FAQ question answering chatbots.” IEEE International Conference on Web Services (ICWS).
[6] Chi-Hsun Li, Ken Chen, Yung-Ju Chang (2019) "When There is No Progress with a Task-Oriented Chatbot: A Conversation Analysis" ACM ISBN 978-1-4503-6825
[7] Koh Mitsuda, Ryuichiro Higashinaka, Tingxuan Li, Sen Yoshida"Investigating person-specific errors in chat-oriented dialogue systems " Association for Computational Linguistics
[8] R. Perera and P. Nand, (2017) “Recent Advances in Natural Language Generation: A Survey and Classification of the Empirical Literature,” Comput. Inform., vol. 36, pp. 1–31, doi: 10.4149/cai_2017_1_1.
[9] A. Zheng and A. Casari, Feature Engineering for Machine Learning: Principles and Techniques for Data Scientists. Sebastopol, CA, USA: O’Reilly Media, Inc., 2018. [Online]. Accessed on: Apr. 19, 2021.
[10] Bocklisch, T., Faulkner, J., Wilde, N., & Nichol, A. (2017). Rasa: Open source language understanding and dialogue management for chatbots. arXiv preprint arXiv:1712.05181.
[11] Mingyue Cheng, Yucong Luo, Jie Ouyang, Qi Liu*, Huijie Liu, Li Li, Shuo Yu, Bohou Zhang, Jiawei Cao, Jie Ma, Daoyu Wang, and Enhong Chen (2025) “A Survey on Knowledge-Oriented Retrieval-Augmented Generation” arXiv:2503.10677v2
[12] Linnéa Olofsson & Heidi Patj, (2024) “Understand me, do you?- An experiment exploring the natural language understanding of two open source chatbots ” Software Engineering for Blekinge Institute of Technology Thesis[43] M. Zubani, L. Sigalini, I. Serina, and A.
[13] Xiao, J. (2025). Academic Level as a Moderator in University Students' Acceptance of Educational AI Chatbots: An Extended TAM3 Model. Applied Sciences, 15(19), 10603.