Authors:
Hanmin Jung
1
;
2
and
Athiruj Poositaporn
1
;
2
Affiliations:
1
University of Science and Technology, 217, Gajeong-ro, Yuseong-gu, Daejeon, Gyeonggi-do, Republic of Korea
;
2
Korea Institute of Science and Technology Information, 245, Daehak-ro, Yuseong-gu, Daejeon, Republic of Korea
Keyword(s):
Client Engagement, Retrieval-Augmented Generation, Large Language Model, Q&A System.
Abstract:
Client engagement refers to the process of companies and customers building and maintaining relationships through communication, personalized marketing, and value-added services. This often results in analysis reports, consulting services, and strategic planning documents. Tools like GPT-4o have significant potential to support these interactions in sectors such as meteorological organizations. However, standalone generative models like GPT-4o face challenges in accessing external datasets and often produce generic outputs. To overcome these limitations, this study introduces a chat-based Retrieval-Augmented Generation (RAG) system integrated with a pattern prediction framework. We demonstrate our RAG system in analyzing air pollution pattern prediction results from our prior study and compare its generated answers with a standalone GPT-4o model. Experimental results show that the RAG system delivers actionable recommendations and contextually enriched outputs grounded in domain-spec
ific data. In future work, we aim to explore the potential of RAG in real-world applications, such as improving client engagement by generating client-focused reports.
(More)