Implementasi Retrieval-Augmented Generation untuk Chatbot Layanan Pelanggan PT. GAWS Inti Solusi

Verio Renedominick; Simon Prananta Barus

doi:10.55382/jurnalpustakaai.v6i2.1768

Penulis

Verio Renedominick Matana University
Simon Prananta Barus Universitas Matana

DOI:

https://doi.org/10.55382/jurnalpustakaai.v6i2.1768

Kata Kunci:

Chatbot, RAG, LLM, Layanan Pelanggan, ChromaDB, PT Gaws Inti Solusi, Retrieval-Augmented Generation

Abstrak

Abstract

The advancement of Artificial Intelligence (AI), particularly Large Language Models (LLMs), has enabled the development of more intelligent customer support chatbots. However, conventional LLM-based systems often suffer from issues such as hallucinated responses and limited context-awareness. This study aims to implement the Retrieval-Augmented Generation (RAG) model in a customer support chatbot to address the limitations of the existing customer service system at PT Gaws Inti Solusi. The system was developed using an Agile methodology through four iterative sprints, covering the setup of the development environment, RAG pipeline integration, UI/UX improvements, and system optimization. The proposed solution integrates LangChain, ChromaDB, and a locally deployed instruction-tuned LLM (mistral:instruct), along with a user-friendly chat widget and administrative dashboard. Evaluation results demonstrate that the RAG-based chatbot significantly improves response accuracy, reduces reliance on company leadership, and accelerates response times. The system effectively enhances the automation and efficiency of customer service processes, offering practical value for enterprise-level deployment

Keywords: Chatbot, Retrieval-Augmented Generation, RAG, LLM, Customer Support, LangChain, ChromaDB, PT Gaws Inti Solusi

Abstrak

Perkembangan teknologi Artificial Intelligence (AI), khususnya Large Language Models (LLM), membuka peluang baru dalam pengembangan chatbot untuk layanan pelanggan. Namun, LLM konvensional masih menghadapi permasalahan seperti halusinasi informasi dan keterbatasan konteks. Penelitian ini bertujuan untuk mengimplementasikan model Retrieval-Augmented Generation (RAG) dalam sistem chatbot guna mengatasi keterbatasan sistem layanan pelanggan di PT Gaws Inti Solusi. Metode pengembangan menggunakan pendekatan Agile dengan empat sprint iteratif yang meliputi setup lingkungan pengembangan, integrasi pipeline RAG, peningkatan UI/UX, serta optimasi sistem. Sistem yang dibangun memanfaatkan LangChain, ChromaDB, dan model LLM lokal (mistral:instruct) yang diintegrasikan dengan dashboard admin dan widget chatbot interaktif. Evaluasi menunjukkan bahwa pendekatan RAG mampu meningkatkan akurasi respons, mengurangi ketergantungan pada pimpinan perusahaan, serta mempercepat waktu tanggap layanan pelanggan. Dengan demikian, sistem ini memberikan kontribusi nyata dalam otomatisasi dan efisiensi layanan informasi di lingkungan perusahaan

Kata kunci: Chatbot, Retrieval-Augmented Generation, RAG, LLM, Layanan Pelanggan ,LangChain, ChromaDB, PT Gaws Inti Solusi

Unduhan

Data unduhan belum tersedia.

Referensi

Adamopoulou, E., & Moussiades, L. (2020). An Overview of Chatbot Technology. IFIP Advances in Information and Communication Technology, 584 IFIP, 373–383. https://doi.org/10.1007/978-3-030- 49186-4_31

Al-Saqqa, S., Sawalha, S., & Abdelnabi, H. (2020). Agile software development: Methodologies and trends. International Journal of Interactive Mobile Technologies, 14(11), 246–270. https://doi.org/10.3991/ijim.v14i11.13269

Bonechi, S., Palma, G., Caronna, M., Ugolini, M., Massaro, A., & Rizzo, A. (2024). Enhancing Customer Support in Banking: Leveraging AI for Efficient Ticket Classification. Procedia Computer Science, 246(C), 128– 137. https://doi.org/10.1016/j.procs.2024.09.235

Brown, S., Pereira, M., & Guvlor, I. (2024). Implementation of Artificial Intelligence Framework to Enhance Human Resources Competency in Indonesia. International Journal of Cyber and IT Service Management, 4(1), 65–71. https://doi.org/10.34306/ijcitsm.v4i1.154

Firdaus, D., Sumardi, I., & Kulsum, Y. (2024). Integrating RetrievalAugmented Generation with Large Language Model Mistral 7b for Indonesian Medical Herb. In Jurnal Informatika Sunan Kalijaga) (Vol. 9, Issue 3).

Lubis, A. T. U. BR., Harahap, N. S., Agustian, S., Irsyad, M., & Afrianty, I. (2024). Question Answering System pada Chatbot Telegram Menggunakan Large Language Models (LLM) dan Langchain (Studi Kasus UU Kesehatan). MALCOM: Indonesian Journal of Machine Learning and Computer Science, 4(3), 955–964. https://doi.org/10.57152/malcom.v4i3.1378

Mahardika, B. T., & Hasan2, A. M. (2025). APPLICATION OF GPT IN CHATBOTS TO FACILITATE KNOWLEDGE MANAGEMENT SYSTEM INTERACTION USING LANGCHAIN (CASE STUDY: PT SOFTBLESS SOLUTIONS). Eduvest-Journal of Universal Studies, 5. http://eduvest.greenvest.co.id

Miklosik, A., Evans, N., & Qureshi, A. M. A. (2021). The Use of Chatbots in Digital Business Transformation: A Systematic Literature Review. In IEEE Access (Vol. 9, pp. 106530–106539). Institute of Electrical and Electronics Engineers Inc. https://doi.org/10.1109/ACCESS.2021.3100885

Pujiono, I., Agtyaputra, I. M., & Ruldeviyani, Y. (2024). IMPLEMENTING RETRIEVAL-AUGMENTED GENERATION AND VECTOR DATABASES FOR CHATBOTS IN PUBLIC SERVICES AGENCIES CONTEXT. JITK (Jurnal Ilmu Pengetahuan Dan Teknologi Komputer), 10(1), 216–223. https://doi.org/10.33480/jitk.v10i1.5572

Sanny, L., Susastra, A. C., Roberts, C., & Yusramdaleni, R. (2020). The analysis of customer satisfaction factors which influence chatbot acceptance in Indonesia. Management Science Letters, 10(6), 1225– 1232. https://doi.org/10.5267/j.msl.2019.11.036

Setyawati, E., Sarwani, H., Wijoyo, H., & Soeharmoko, N. (2020). RELATIONAL DATABASE MANAGEMENT SYSTEM (RDBMS). CV. Pena Persada.

Slawek-Polczynska, A. (2020). Is Agile always the best solution for software development projects? Soldevelo Blog. Retrieved from https://soldevelo.com/blog/is-agile-always-the-best-solution-forsoftware-development-projects/

Suryotomo, A. P., Muhammad Akbar, B., & Husaini, R. (2024). Performance Analysis of FastAPI Framework on Lost Circulation Handling Management Application in Oil Well Drilling Analisis Performa Framework FastAPI pada Aplikasi Manajemen Penanganan Lost Circulation di Pemboran Sumur Minyak. Jurnal Informatika Dan Teknologi Informasi, 21(1), 110–121. https://doi.org/10.31515/telematika.v21i1.13259

Syaputri, L., Gunawan Putra, E., Syahrani, E., Dwian, E., Purwani, F., informasi, S., dan Teknologi, S., & Islam Negri Raden Fatah Palembang, U. (2024). PERBANDINGAN EFEKTIVITAS METODE WATERFALL DAN AGILE DALAM PENGEMBANGAN SISTEM INFORMASI SEBUAH SYSTEMATIC LITERATURE REVIEW. Journal of Scientech Research and Development, 6(2). https://idm.or.id/JSCR/in