# A Study on the Implementation of Generative AI Services Using an Enterprise Data-Based LLM Application Architecture

**Authors:** Cheonsu Jeong  
**arXiv:** [2309.01105](https://arxiv.org/abs/2309.01105)  
**Date:** 2023-09-03  
**Summary page:** https://www.summarizepaper.com/en/paper/2309.01105v2/

## Key points

- Focus on implementing generative AI services using Large Language Models (LLMs) in enterprise data-based application architecture
- Addressing information scarcity through tailored solutions leveraging LLM capabilities
- Exploration of strategies like fine-tuning techniques and direct document integration to mitigate inadequate data
- Development of Retrieval-Augmented Generation (RAG) model for enhanced information storage and retrieval processes
- Emphasis on key phases of the information storage and retrieval methodology supported by the RAG model
- Illustrative examples showcasing applicability of proposed method
- Contribution to advancing generative AI fields by enhancing data-driven content generation and promoting active utilization of LLM-based services within corporate settings

## Summary

In this study by Cheonsu Jeong, the focus is on implementing generative AI services through the utilization of Large Language Models (LLM) within an enterprise data-based application architecture. With the rapid advancements in generative AI technology, LLMs have emerged as key players in various domains. The research specifically tackles the issue of information scarcity and proposes tailored solutions by leveraging the capabilities of LLMs. Strategies for mitigating inadequate data are explored, including the effectiveness of fine-tuning techniques and direct document integration to address data insufficiency. One significant contribution of this work is the development of a Retrieval-Augmented Generation (RAG) model, designed to enhance information storage and retrieval processes for improved content generation. The study delves into the key phases of the information storage and retrieval methodology supported by the RAG model, emphasizing their importance in overcoming data scarcity challenges. Through a comprehensive analysis,  of the proposed method is highlighted with illustrative examples showcasing its applicability. By implementing  for information storage and retrieval, this research not only deepens our understanding of generative AI technology but also facilitates its practical usability within enterprises utilizing LLMs. The study holds substantial value in advancing generative AI fields by offering insights into enhancing data-driven content generation and promoting active utilization of LLM-based services within corporate settings. Overall, this work contributes significantly to bridging gaps in data availability and improving content generation processes through innovative AI methodologies.