Exploring the Potential of Large Language Models in Supply Chain Management: A Study Using Big Data

Introduction

A supply chain, integral to modern businesses, facilitates the exchange of materials, information, and resources among interconnected organizations, ensuring the delivery of valuable products and services to consumers (Stadler, 2008). It operates within a complex web of suppliers, customers, and service providers, which demands intricate decision-making (Bag et al., 2023). While technological advancements have automated and optimized supply chain operations, the challenge of conveying optimization outcomes to stakeholders persists.¹

Furthermore, the exponential growth of Internet data presents both opportunities and challenges for organizations. Modern supply chain professionals must leverage big data analytics, including data science and big data, to enhance the supply chain’s processes and performance (Chatterjee et al., 2023; Waller & Fawcett, 2013). However, data quality remains paramount for effective decision-making in supply chain management (SCM), which emphasizes the significance of functional capabilities, information sharing, and data proficiency (Hazen et al., 2014). The successful management of operations and the supply chain hinges on a data-centric approach, which has evolved from traditional reporting to advanced analytics that encompasses statistical analysis, forecasting, and real-time optimization (Jacobs et al., 2014). Nonetheless, integrating big data analytics into supply chains remains a huge scope for research (Kache and Seuring, 2017).

Supply chain operations entail complex decision-making despite the automation and optimization achieved through advancements in computing (Oliveira & Pereira, 2023). Although optimization tools have enhanced the efficiency of supply chain decision-making, they frequently necessitate input from individuals lacking expertise, resulting in prolonged interactions with program managers and data experts (Lambert & Cooper, 2000). The emergence of large language models (LLMs) such as ChatGPT has spurred interest in the application of artificial intelligence (AI) within supply chains. These LLMs, including various deep generative models (DGM), play a pivotal role in deciphering intricate probability distributions (Li et al., 2023a). Within the retail sector, ChatGPT is used to enhance inventory management and detect trends in customer inquiries, seamlessly integrating with supply chain software and warehouse management systems (Kumar et al., 2023). Meanwhile, in the military domain, AI is revolutionizing equipment acquisition and sustainment, predicting demand, optimizing transport routes, and automating inventory management to trim expenses and enhance supply chain efficiency (Mikhailov, 2023).

Another critical aspect of the supply chain involves security breaches. Relying on software systems leads to increased vulnerability to supply chain breaches, resulting in significant financial and data losses. Prioritizing cybersecurity is crucial, yet traditional methods for analyzing past failures involve manual report reading and summarization. Automated support through natural language processing (NLP), including LLMs, can reduce costs and enhance the analysis of such incidents (Trappey et al., 2022; Singla et al., 2023).