AI Applications in Supply Chain Management: A Survey

Abstract

The advent of Industry 4.0 and the integration of Artificial Intelligence (AI) is transforming supply chain management (SCM), improving efficiency, resilience and strategic decision-making capabilities. This research study provides a comprehensive overview of AI applications in key SCM processes, including customer relationship management, inventory management, transportation networks, procurement, demand forecasting and risk management. AI technologies such as Machine Learning, Natural Language Processing and Generative AI offer transformative solutions to streamline logistics, reduce operational risk and improve demand forecasting. In addition, this study identifies barriers to AI adoption, such as implementation challenges, organizational readiness and ethical concerns, and highlights the critical role of AI in promoting supply chain visibility and resilience in the midst of global crises. Future trends emphasize human-centric AI, increasing digital maturity, and addressing ethical and security concerns. This review concludes by confirming the critical role of AI in shaping sustainable, flexible and resilient supply chains while providing a roadmap for future research and application in SCM.

1. Introduction

In the midst of the fourth Industrial Revolution, where transformative technologies such as Artificial Intelligence (AI), advanced robotics and Internet of Things (IoT) are reshaping production and business overall, the lines between humans and machines are becoming less distinct. And what is more, Industry 5.0 brings humans and technology even closer [1,2]. In general, AI involves the use of computers to perform tasks that are typically run by humans, such as learning, understanding, recognizing, reasoning, and adapting [3]. The economic contribution of AI technologies will reach around USD 13 trillion by 2030 [4], while the supply chain management market is expected to surpass USD 62.2 billion by the same year [5]. Considering the increasing role of AI in supply chain management (SCM), this study seeks to investigate its primary applications, benefits and challenges. To structure this investigation, we pose the following research question: How are AI techniques being applied in key SCM processes, and what are the emerging trends and challenges associated with their adoption?

Supply chain management (SCM) is the monitoring and optimization of the production and distribution of products and services of a company [6]. Supply chains are cause for a lot of concern for and are the focus of many enterprises, governments and consumers and gain broader media coverage [7]. In today’s major competitive environment, a cloud supply chain based on Industry 4.0 technologies and digital platforms transforms into the “supply chain-as-a-service” paradigm [8,9]. The benefits of employing AI in SCM are resilience improvement [10,11,12] and risk management [13,14,15], end-to-end visibility and transparency [2,16,17], information and process resilience [18], the real-time tracking of goods in every part of the supply chain [19], sustainability [20], and food safety and public health [21]. It is important to note that all these features are possible through the integration of the digital twin concept into SCM processes, leveraging real-time data and AI-based algorithms to interpret them [9].

Past academic work concerning AI in SCM emphasized the need to lessen the current hype [22]; brought to the surface the lack of know-how in organizations [23] and the lack of understanding in ordinary decision makers [24]; proposed an AI taxonomy relevant to SCM [4]; described the most frequent AI techniques and SCM subfields [25,26]; utilized a bibliometric review to trace the evolution of AI research in SCM [2]; critically reviewed the drivers, practices, benefits, barriers and consequences of AI adoption [27,28,29]; proposed an AI integration framework [30]; and reviewed the effect of AI on operational efficiency, strategic innovation and sustainability [31]. Moreover, the following studies offered a practical framework of AI applications in SCM [32], reviewed the role of AI in the supply chain analytics [33,34], examined the organizational and behavioral factors that promote AI adoption in supply chain [35], introduced a research framework for AI in supply chain resilience [36] and explored the relationship between AI applications and supply chain concentration [37]. Finally, these studies investigated the potential of AI in supply chain visibility [38], evaluated the benefits and challenges of AI in the agricultural supply chain [39], examined the relationship between AI and supply chain finance [40,41] and proposed specific officer roles in overseeing AI utilization in SCM [42].

The scope of this research study covers a thorough analysis of AI applications in SCM, with a focus on the key processes of the chain. Although our paper encompasses a broad and representative range of studies, it is not intended to be an exhaustive review of all existing research, since our objective is to present relevant studies that emphasize the integration of AI techniques in SCM, focusing on their practical applications and emerging trends. Chapter Two and Three cover relevant information on supply chain and AI, respectively. Also, Chapter Three describes specific AI applications in SCM, more specifically in the areas of customer relationship management, inventory management, transportation, procurement, demand forecasting, resilience and risk management. Finally, Chapter Four highlights future trends, challenges and threats of AI presence in SCM, and in Chapter Five, conclusions are presented. To guarantee a thorough and representative selection, we performed searches in prominent academic databases such as Scopus, Web of Science, IEEE Xplore, Google Scholar and ProQuest. We used the search phrase ‘Artificial Intelligence’ AND ‘Supply Chain Management’ to identify current and pertinent studies. Although this initial query yielded a large number of results, we implemented additional selection criteria—such as peer-reviewed sources, publication date (2015–2025) and relevance to AI-driven SCM applications—to narrow down the final set of studies analyzed in this survey.

2. Supply Chain

A supply chain can be understood as a network that creates value through interconnected elements. It involves various business sectors working closely together, including production, services, funds and information. The main objective is to ensure that products are available to consumers [43]. It is vital for contemporary businesses to build integrated supply chains with infrastructure and networking features [44]. The SCM research landscape has evolved significantly over time. Initially, the focus was on freight transportation. However, subsequent studies have brought to light other critical areas, such as risk [45], performance [46] and integration [47]. In addition, there is an increasing focus on the flow of information within the network of organizational relationships [48], both internally and externally, alongside the management of supply networks. Given these research challenges, the main objective of SCM is to enhance four key flows: goods, information, cash flows and overall processes [49]. A major obstacle to managing the supply chain as an integrated system is the lack of appropriate information technology. Despite the emergence of Industry 4.0 technologies, the significant costs associated with their installation and operation in each stage of the chain have not yet been fully explored [50]. Moreover, there are specific barriers and drivers that allow for sustainable SCM [51], namely, there are social, economic and environmental concerns on one hand and technological, competitive and operational benefits on the other hand.

Key components that are typically included in an SCM framework (Figure 1) are customer relationship management (CRM), inventory management, transportation networks, procurement, demand forecasting, resilience and risk management. These principal activities play a critical role in both the internal and external aspects of a business, which can be understood as a complex system involving inputs, internal processes and outputs. There are quite a few challenges and opportunities to implementing AI technologies in the above processes, both of which pave the way for a cautious but fruitful transition [22].

3. Artificial Intelligence

AI is defined as intelligence through which machines can display and emulate all human cognitive functions, such as problem solving and decision making, to the benefit of organizational optimization and automation. The presence of AI is evident in the fields of agricultural management, education, healthcare, fashion, e-commerce, gaming and the military [1]. The roots of AI go back to the 1950s, and its key techniques include Machine Learning (ML), Deep Learning (DL), Neural Networks (NNs), Natural Language Processing (NLP), Computer Vision (CV), Knowledge Representation and Reasoning (KR&R), Recommender Systems (RSs) and optimization (OP) [3,52]. Many of these techniques incorporate simulation as a fundamental approach to modeling, testing and optimizing AI-driven decisions, particularly in complex and dynamic supply chain environments. A branch of AI is Generative AI (GEN AI), which can produce a variety of different forms of content, namely, text, graphics, audio, video or other data forms, and leverages Machine Learning models [53]. A prime example of Generative AI is ChatGPT (Version 4.0) which offers an adequate number of benefits but also various threats and challenges to SCM [54,55]. The IoT has the leading role of generating data for AI to analyze. Integrating the IoT and AI within the SCM context provides notable advancements in supply chain transparency, agility and overall functionality [56]. Artificial Intelligence of Things, which refers to utilizing the IoT to execute smart tasks with the aid of AI integration, is one of these upcoming developments that can convert a complex supply chain into a unified process [57]. Figure 2 depicts various AI manifestations as they have been AI-generated by the Microsoft Designer program (Version 19041.0).

Nowadays, it is more than evident that AI leads the way in revolutionizing every aspect of SCM and is employed to make educated predictions on demand and transportation routes, suggest innovative solutions and optimize operations costs [58]. Moreover, the need to make ends meet in a highly competitive environment sets AI as the main force to succeed. Even when faced with a global pandemic crisis, such as COVID-19, AI is a crucial factor that helps management achieve its goals [59]. Furthermore, although there is a hesitation in managers to utilize AI [59], neurosymbolic AI is explicitly suitable to give explanations for every decision based on AI models [60].

Managers’ perceptions and intentions in utilizing AI in decision making are still in a contemplating phase, even though the COVID-19 crisis has strengthened business operations through AI [59]. AI-driven SCM optimization delivers a range of benefits that markedly enhance inventory management, improve demand projection, optimize logistics, increase efficiency and productivity and upgrade decision-making policies [61]. The adoption of AI in supply chain decision making is also triggered by the Environmental, Social and Governance framework. These triggers are product waste reduction and greenhouse gas emission reduction, product security and quality, and agile and lean practices, respectively [62].

3.1. Customer Relationship Management

Customer relationship management comes to further addressing and nurturing customer needs, boosting the role of classic SCM. An agent-based model that tracks customer experiences in a social network can build upon word-of-mouth reputation to accelerate revenue generation [24]. Chatbots and virtual assistants powered by AI can address customer concerns, manage order fulfillment, support tickets, improve response time and ensure live shipment tracking [16,61].

3.2. Inventory Management

One of the biggest challenges in SCM is inventory management and its relevant cost, as demand patterns change to facilitate diverse needs [63]. AI-powered systems can optimize inventory, when taking into consideration factors such as demand, storage costs, lead time and even supply chain constraints [16]. Integrating AI in inventory management offers numerous advantages, namely, reduced stock-outs, minimized overstocking, strategic clearance sales and improved profit margins [19]. AI techniques provide new, innovative ways to inventory control and planning challenges by capturing inventory patterns naked to the human eye [24]. Machine Learning techniques, like reinforcement learning and anomaly detection, are capitalizing on data insights to fine-tune inventory levels. The analysis of historical stock quantities and abrupt changes in trends and the handling of large volumes of data produce informative reports on estimated inventory status [64,65]. Robotic systems assisted by AI and drones can rationalize warehouse operations and automate functions, such as picking, packing [66] and inventory counting [67], leading to accuracy improvement and enhanced use of human resources [16]. In all available picking scenarios, there have been numerous proposed solutions based on simulation and mathematical models. An intelligent agent-based method could handle the added complexity driven by the growing adoption of new and premium services [24]. Large Language Models have played a crucial role in automating inventory management and order fulfillment by enabling advanced data analysis and decision-making capabilities in areas such as historical logistics data (delivery logs, transportation trends, climatic conditions and consumer demand predictions) [68].

3.3. Transportation Networks

The supply chain network is the backbone of SCM, as it brings together suppliers, manufacturers, distributors and customers [69]. AI can streamline delivery routes, vehicle loads and logistics timelines to cut back fuel consumption and save work hours [16]. The use of network theory and graph algorithms allows for a better understanding of the key features of such a network by identifying bottlenecks and facilitating the flow of goods and information [64]. Identifying optimal areas for logistics, storage facilities and retail operations requires intricate planning since costs and performance depend on these decisions. AI is a valuable assistant in facility location planning by taking into consideration multiple factors, for example, customer demographics, land costs, transportation infrastructure and regulatory environment [70]. Dynamic route scheduling constitutes a major challenge for SCM, notably the last-mile delivery task. Although there are many heuristic techniques to handle everyday transportation issues, AI really shines in solving this vehicle routing problem in the forms of genetic algorithms, ant colony optimization algorithms [24] and reinforcement learning [65]. On top of that, Generative AI can be employed to create backup plans to neutralize disturbances such as traffic jams and severe weather [71].

3.4. Procurement

Procurement and resource planning departments have to make everyday decisions concerning volume, capital and risk related to obtaining items necessary for all operational purposes. Intelligent agent-based systems take up the role of a human decision maker or provide aid to the purchasing manager with the sequence of strategic and operational procurement choices [24]. Intelligence process automation can refine routine operation, in cases of data entry, order management and invoice processing, while data collected from appliances and machinery are used to predict equipment failure and program predictive maintenance [61]. Generative AI analyzes a lot of parameters, like financial viability, product excellence, dependability, operational effectiveness and green practices to put together an optimal portfolio of potent suppliers [71].

3.5. Demand Forecasting

The integration of AI into demand projection presents many benefits, such as improved production planning, strategic inventory allocation, risk mitigation and new-product development [19]. Machine Learning techniques, like support vector machines, Neural Networks and decision trees, utilize data-informed insights to produce more precise forecasts and give companies tools to develop dynamic inventory policies to satisfy ever-changing demand [64]. In another research work [65], the AI techniques used for demand forecasting are the auto-regressive integrated moving average, long short-term memory networks, gradient boosting machines, support vector machines and Deep Neural Networks.

Supply chain demand prediction and administration is a primary focus for SCM, where Artificial Neural Networks, data mining and fuzzy models are employed to foresee consumer consumption [25]. Demand forecasting is solely based on historical data for existing products and services. In cases where there is a new product or an innovative service, the absence of any chronological records remains deterring for any estimated demand. Precisely for these instances, AI is a viable alternative for sales projection and planning [24]. AI-based analysis of social media, like sentiment analysis [72], can help gain deeper insights into customer behaviors and preferences, allowing predictive models to identify potential markets and profit margins [73].

The application of AI in demand forecasting for irregular demands [63] was investigated, and the most effective approaches were identified. The Neural Networks that were employed improved demand forecast accuracy, especially in cases of intermittent demands, lowering overall financial burdens, such as higher stock levels. Large Language Models have shown noteworthy progress in enhancing demand prediction accuracy by processing and analyzing large volumes of data, in both text and numeric forms [68].

3.6. Resilience and Risk

Supply chain resilience is improved with the help of AI in turmoil times. Moreover, firm performance is related to AI and supply chain resilience [74]. Especially, AI improves transparency by employing continuous monitoring, handling last-mile delivery and tailor-made demands, and lessens the negative impacts of global crises, like COVID-19 [10,11]. A quality measure of resilience is also important, not just a quantity one. While it is crucial to understand the disruption and recovery actions a firm must take to roll back to normal proceedings, it is even more intuitive to boost the quality of network health by investing in prevention policies [75]. There is empirical evidence that the level of visibility maintained and shared with all supply chain partners defines the impact of disruption events. Thus, AI fosters resilience in the supply chain [76].

Artificial Neural Networks are used to recognize and lessen risks in everyday operations in supply chains. These risk cases are disruptions in supply, variations in demands and shifts in market conditions [77]. Once again, analysis of historical data, patterns and trends can recognize and create risk management policies that alleviate disruptions’ effects and fortify resilience [64]. The integration of deep reinforcement learning and predictive analysis further enhance decision-making processes in real time and anticipate disruptions [78]. The most used AI technique in supply chain resilience is Bayesian Networks, as a thorough approach to evaluating risk, analyzing uncertainty and deciding amid structural dynamics [79]. There is concrete evidence demonstrating that the data processing functionalities of AI impact supply chain performance by enhancing supply chain resilience overall [80]. The numerous successive crises and supply chain instabilities attest to the need for further utilization of AI to bolster resilience.

Additionally, resilience necessitates a forward-thinking approach to risk management, namely, the identification of potential threats, the assessment of likelihood and impact and the implementation of prevention and/or mitigation policies [81]. Risks in SCM can be categorized into supply and demand risk, process and control risk, and environmental and information risk [14] and are quantified as the multiplication of their likelihood and effects. Predictive risk management is based on AI methods and instruments, like ensemble learning and Neural Networks to anticipate risks and take actions to mitigate them [65]. Specific AI techniques, including fuzzy logic programming, Machine Learning, Big Data and agent-based systems, could be used to promote resilience in supply chain management [82]. A proactive approach to risk reduction in SCM can be achieved with Generative AI. The continual analysis of specific supplier performance indicators, data obtained from markets and IoT devices, and AI algorithms can identify potential emerging threats and facilitate documented decision making [71]. Deep Convolutional Neural Networks (DCNNs) demonstrate remarkable efficiency stemming from calculating intricate and nonlinear correlations among variables. The utilization of DCNNs bolsters predictability and robustness in the global supply chain [83]. Large Language Models have demonstrated their worth as essential assets for improving risk assessment and mitigation by analyzing various data sources and delivering actionable intelligence [68].

Table 1. Key SCM activities and their corresponding AI apps.

SCM Activities	AI Apps
Customer relationship management	Agent-based models, chatbots and virtual assistants
Inventory management	Machine Learning, robots, drones, agent-based models and Large Language Models
Transportation networks	Network theory, graph algorithms, genetic algorithms, ant colony optimization and reinforcement learning
Procurement	Agent-based models, process automation and Generative AI
Demand forecasting	Machine Learning, support vector machines, Neural Networks, decision trees, Deep Neural Networks, data mining, fuzzy models, sentiment analysis and Large Language Models
Resilience	Artificial Neural Networks, deep reinforcement learning and Bayesian Networks
Risk	Ensemble learning, Neural Networks, fuzzy logic programming, Machine Learning, Big Data, agent-based systems, Generative AI, Deep Convolutional Neural Networks and Large Language Models

summarizes key SCM activities with their corresponding AI applications.

4. Future Trends, Challenges, Threats

The future of AI is promising across all domains of human endeavor and innovation [1], garnering significant academic attention [33,84]. Significant disruptions, such as COVID-19, have commenced a discussion on reversing outsourcing in supply chains and further developing local and in-house manufacturing, utilizing the Manufacturing-as-a-Service model (MaaS) [4]. A shift towards human-centered AI in SCM is expected to manifest in the form of training and specializing human workforce while boosting AI familiarity among the general population [58]. There is a clear trend in utilizing AI applications to tackle both traditional and fresh problems [26].

Nowadays, AI faces many challenges, namely, data privacy [68] and algorithmic bias [85], cybersecurity [68] and ethics [86], the absence of transparency and documentation, the lack of specialized professionals [33] and implementation challenges [1,65]. Also, there is a lack of evidence to measure its Return on Investment (ROI) in supply management enterprises, thus making it difficult to appraise its value [58]. Managers’ attitudes toward adopting AI for decision making remain in a deliberative stage, despite the COVID-19 crisis highlighting AI’s potential to enhance business operations [59]. Additional challenges for AI in SCM are strong dependence on computer software, difficult implementation and inapplicability in cross-function and cross-border supply chain decision frameworks [24]. On top of that, companies must advance and elevate their degree of digital maturity in their organizational culture context before trying to implement AI solutions in their SCM [87]. Over and above technical challenges, there are organizational hurdles to overcome, namely, redefining processes, roles and responsibilities [85].

The main threats of AI presence in human societies are the disruption of the current job landscape, use for nefarious purposes, weaponization and the increase in inequalities [1].

Industry 4.0 buzz is not consistent with companies’ readiness to adopt its technologies, since they lack the expertise, specifically tied to their existent business domain [23]. Managers should be cautious and not hold lofty expectations [88] in regards to the implications of AI concerning performance, since organizational features are also critical [22]. Besides AI in SCM, Big Data, the IoT, advanced Natural Language Processing, blockchain [89], robotics, autonomous vehicles and drones are all crucial fields to explore scientifically for refining even more the management of the supply chain [52,90].

5. Conclusions

This paper highlights the transformative role of AI in revolutionizing SCM within the evolving contexts of Industry 4.0 and the emerging paradigms of Industry 5.0. By leveraging AI technologies, such as Machine Learning, Natural Language Processing and Generative AI, SCM processes can unlock unprecedented levels of efficiency, resilience and adaptability. Moreover, this study provides a thorough analysis of AI applications across key supply chain processes, including customer relationship management, inventory management, transportation networks, procurement, demand forecasting, resilience and risk management. Notable advancements from the existing literature are presented, including enhanced demand prediction accuracy, optimization of logistics operations, improved supplier and customer relationships and reduced operational risks, all contributing to a more agile and responsive supply chain ecosystem. Moreover, AI’s critical role in fostering supply chain visibility and resilience has been underscored, particularly in mitigating challenges posed by global crises, disruptions and uncertainties.

However, despite these promising developments, several barriers impede the full-scale adoption of AI in SCM. These include technological implementation obstacles, disparities in organizational digital maturity and concerns around ethics, transparency and cybersecurity. What is more, this study has some limitations, as the literature was selected based on specific search terms and databases, potentially excluding some relevant works. Additionally, with the fast-paced evolution of AI, newer advancements may arise that fall outside the scope of this review. Addressing these challenges is essential to unlocking AI’s full potential for creating dynamic and sustainable supply chains.

This comprehensive review outlines guidelines for future research to address these gaps by focusing on optimizing AI integration strategies, fostering digital readiness across industries and addressing ethical and security-related concerns to ensure equitable and responsible AI deployment. To conclude, embracing a human-centric approach to AI paired with innovations in AI-driven tools will be vital to shaping sustainable, resilient and adaptable supply chains.