Content area
Purpose
Artificial intelligence (AI) usage improves e-commerce logistics efficiency. However, many actors can play significant roles, such as supply chain consistency (SCC), last-mile logistics (LML) performance and collaboration and coordination among logistics firms. This study aims to assess how SCC and LML performance mediate and collaboration and coordination moderate the relationship between AI usage and logistics efficiency.
Design/methodology/approach
A structured questionnaire was used to collect the data. A total of 245 valid responses were received from Indian e-commerce businesses. The data were then analysed using AMOS v25 and structural equational modelling using SPSS for regression, PROCESS macro for mediation and moderated mediation analysis.
Findings
The findings show that AI usage independently impacts logistics efficiency, with SCC and last-mile delivery performance as mediating variables. Collaboration and coordination among logistic firms are also critical moderators in enhancing AI’s efficacy in logistic operations. The study findings suggest the integration of AI into logistic operations and provide implications to managers on the urgency of fostering a collaborative and synchronised environment to utilise the full potential of AI in e-commerce businesses.
Originality/value
This study not only contributes to the field of logistics theory by presenting empirical data on the various ramifications of AI but also offers practical guidance for logistics firms, particularly those operating in developing economies, on how to strategically employ AI to enhance operational efficiency and attain a competitive advantage in the era of e-commerce logistics in the digital age.
1. Introduction
The arrival of digital technology is considered a primary force behind the rapid growth of e-commerce, raising it to new heights (Govindan et al., 2024). The sector is quickly growing and at the forefront of technological changes, boosting productivity and ease of use. The continual adaptation of new technologies has promoted societal progress and innovative management practices, creating new avenues for trade, communication and technological collaboration (Fang et al., 2023). The escalating growth of e-commerce businesses has highlighted the crucial role of logistics in ensuring operational efficiency and customer satisfaction (Hong et al., 2019; Kawa and Zdrenka, 2023). Consequently, the rising demand for prompt and reliable service has necessitated the logistics industry to adopt new technologies (Xia et al., 2020). For example, Grover and Ashraf (2023) and Chen et al. (2021) highlighted that using artificial intelligence (AI) in logistics, including quick delivery systems, warehouse automation and predictive analytics, significantly enhances operations excellence. This paper examines the intricate relationship between AI and the effectiveness of logistics in e-commerce. It examines how advanced technology, such as AI, creates new opportunities for success in a rapidly evolving digital marketplace (Micu et al., 2021; Javaid et al., 2022). The synergy of AI in logistics is considered crucial, underpinning the profitability and sustainability of e-commerce businesses in the digital age.
Many studies reveal AI’s incredible effect on logistics efficiency. For example, the study by Xia et al. (2020) shows the potential of AI-powered predictive analytics to optimise inventory management and predict demand, resulting in decreased waste and enhanced responsiveness. Similarly, the significance of AI in the automation of warehouse operations is reported by Chen et al. (2021), who demonstrate how AI-generated algorithms can simplify order processing and accelerate delivery durations. These innovations go beyond improving functional efficiency, resulting in cost savings and increased customer satisfaction. Prominent companies like Amazon have successfully integrated AI into their logistical processes by utilising advanced algorithms to optimise routes and make shipment predictions. This has resulted in substantial reductions in both delivery times and cost savings. Furthermore, a significant advancement in product return management using AI, where sophisticated algorithms examine return trends and suggest preemptive actions, has been observed (Fang et al., 2023).
Despite the extensive body of literature, further research is necessary to understand how AI might enhance operational efficiency in the e-commerce business (Xia et al., 2020; Chen et al., 2021). Given AI’s rising prevalence across various industries, it is crucial to explore the dynamic relationships between AI usage (AIU), supply chain consistency (SCC), last-mile logistics (LML) performance and collaboration and coordination (CC) within e-commerce logistics (Imam, 2024; Shkurenko et al., 2023; Neto et al., 2023). Several critical questions regarding the impacts of AI on SCC, LML and logistic efficiency (LE) still need to be addressed. This study initially investigates the roles of SCC and LML as mediators in the relationship between AIU and LE. It then examines the potential sequential mediation effects of SCC and LML on the relationship between AIU and overall LE within the e-commerce context.
Furthermore, this work also assesses the moderating effect of logistics coordination and collaboration (CC) on the enhanced performance of SCC and LML due to AI applications. Drawing upon recent studies by Calvo et al. (2023), which highlight AI’s role in enhancing omnichannel customer experiences, our research identifies strategic, relational and operational gaps in the current application of AI within the e-commerce sector. Additional analysis by Leung et al. (2020) into modelling near-real-time order arrival demand in an e-commerce context reveals opportunities to boost last-mile operational efficiency through improved data accuracy and consistency. Furthermore, findings by Pournader et al. (2021) demonstrate the effectiveness of AI in optimising supply chain processes, such as inventory management and real-time price adjustments, suggesting a broader impact on logistic operations. These insights lay the groundwork for addressing the research gaps our study targets, emphasising their significance in optimising LE and supporting the adaptive growth of the e-commerce sector in a digitally connected economy. By exploring these critical research gaps, this study seeks to elucidate the complex interplay between AI and logistic effectiveness, thereby shaping strategic decisions within the industry. The specific dynamics of these relationships are detailed through the following research questions:
The rest of the paper is structured as follows: Section 2 of the study delves into the concepts and theories found in existing literature examining the influence of AIU, SCC, LML and CC on logistics. Section 3 explains how data were gathered from managers and outlines the statistical analysis methods used. The results in Section 4 demonstrate the connections between AI usage and efficiency in logistics changes in supply chain operations and how CC are moderating factors. Discussion of these outcomes in Section 5 ties them to established frameworks and their practical implications. Section 6 addresses any limitations of the study. Suggests future research paths within AI and logistics. Finally, Section 7 wraps up by addressing the research inquiries.
2. Theoretical framework and hypothesis formulation
2.1 Constructs’ definition
2.1.1 Artificial usages (AIU)
AI is revolutionising e-commerce logistics by enhancing efficiency, automating operations and improving customer experiences (Belhadi et al., 2021). As the primary independent construct of this study, AIU encompasses the deployment of technologies such as chatbots, smart terminals and intelligent algorithms that streamline operations from inventory management to customer interactions, thereby facilitating a more dynamic e-commerce environment (Dash et al., 2019; Wang et al., 2021). AI’s capability to process and analyse large volumes of data is critical, enabling personalised customer recommendations, sales forecasting and optimised inventory management, which collectively enhance the efficiency of e-commerce operations (Jauhar et al., 2023; Lingam, 2018).
AI is crucial in improving logistics operations by forecasting inventory requirements and determining efficient distribution routes (Min, 2010; Xia et al., 2020). It uses algorithms to assess the connection between supply and demand, thus increasing the level of automation in sorting tasks within logistics (Konstantakopoulos et al., 2021). Further, incorporating AI helps improve the restructuring of order fulfilment procedures, leading to decreased order processing time and the distances travelled in logistics, both crucial factors in efficiency, according to Leung et al. (2020). Additionally, adopting Internet of Things (IoT) technology in e-commerce supply chains further enhances these optimisations, resulting in streamlined operations and cost reductions (Shao et al., 2019).
Furthermore, AIU fosters collaboration among various stakeholders in the logistics chain. It enhances information sharing and coordination, which are crucial for responding swiftly to customer demand changes and achieving cost efficiency (Jarrahi et al., 2023). Integrating storage and distribution processes through AI leads to direct cost savings and boosts overall supply chain competence, positioning e-commerce businesses for better competitive advantage and sustainability (Modgil et al., 2022). These comprehensive enhancements underscore AI’s transformative potential in e-commerce logistics, making it a key area of focus for businesses aiming to optimise their operational efficiencies and improve customer satisfaction.
2.1.2 Supply chain consistency (SCC)
According to Bendoly et al. (2018), SCC is characterised by the supplier’s ability to fulfil orders and recover reliably from service failures. This construct plays an essential role in ensuring the quality and dependability of supply chain operations, requiring all participants to adhere to performance standards that meet customer needs (Singhal and Singhal, 2012). Consistency extends beyond mere operational execution to encompass data management practices; data replication is crucial for maintaining coherence across supply chain applications (Hazen et al., 2014). The robustness of supply chain operations also depends on the quality of coordination agreements, assessed based on their reliability (Zhao et al., 2010). For global corporations, it is vital to maintain reliable processes in supply management, vendor relationships and procurement activities (Adams et al., 2014). Furthermore, the seamless integration of information and communication technologies enhances the efficiency of sales supply chain operations and optimises overall performance through simulation modelling and analysis (Dubey et al., 2020; Hu et al., 2024).
Establishing a resilient supply chain depends on building strong connections, creating value through continuous material and information flows and maintaining effective communication (Xia et al., 2020; Shkurenko et al., 2023). Incorporating cutting-edge technologies such as AI significantly enhances the resilience of supply chains (Sorooshian et al., 2022). AI tools enhance visibility, enable customised solutions and streamline decision-making processes and predictions, thus strengthening supply chain management’s preparedness, agility and recovery (Modgil et al., 2022; Imam, 2024). Additionally, AI drives supply chain resilience by automating processes, validating data and generating accurate forecasts, reducing operational latencies and improving performance (Belhadi et al., 2021). The synergy of AI with IoT and blockchain technologies augments resilience, optimises operational efficiency, enhances forecasting, planning and monitoring and improves communication and information sharing across the supply chain (Shao et al., 2019). These technological advancements underscore the transformative impact of AI on supply chain consistency, enabling businesses to move effectively navigate complexities and maintain a competitive advantage in dynamic market conditions.
2.1.3 Last-mile logistics (LML)
LML is a crucial element within the supply chain, essential in conventional commerce and humanitarian relief efforts, where timely and accurate delivery of goods to end-users is critical (Balcik et al., 2008). This stage of the supply chain involves transporting goods from local distribution centres to end consumers. It is significantly influenced by factors such as customer behaviour, fairness and coordination among various supply chain stakeholders (Jin and Yiyou, 2015). Adopting intelligent transportation systems and advanced technologies, including robots and drones, plays a pivotal role in optimising this crucial stage by enhancing efficiency and responsiveness (Simoni et al., 2020). Moreover, the proliferation of e-commerce has significantly impacted last-mile delivery, necessitating a focus on urban sustainability and the involvement of public authorities to support efficient urban logistics (Kawa and Zdrenka, 2023). Additionally, efficient route planning is a fundamental aspect of effective LML, supported by several theoretical frameworks to improve this process (Xia et al., 2020).
Evaluating the Effectiveness of Last-Mile Logistics: The effectiveness of LML can be assessed from multiple perspectives. Technological advancements significantly improve decision-making and operational efficiency within this field (Bányai, 2018; Lagin et al., 2022). The environmental and social impacts are equally important, with emerging technologies proposed as solutions to support more sustainable and socially responsible logistics practices (Melkonyan et al., 2020). Emphasising customer-centric factors such as behaviour and fairness enhances the effectiveness of last-mile operations and ensures that delivery processes align with customer expectations and ethical standards (Jin and Yiyou, 2015). The shift towards energy-efficient fulfilment options demonstrates an ongoing commitment to more sustainable logistics practices (Sureeyatanapas et al., 2018). Furthermore, understanding the role of customers in adopting innovations within LML is crucial, as it underscores the importance of adapting to consumer behaviour amidst evolving logistics technologies (Fang et al., 2023).
2.1.4 Coordination and collaboration (CC)
Effective CC are fundamental in enhancing supply chain efficiency, competitiveness and cost management (Carvalho et al., 2021). These processes facilitate real-time information sharing and joint planning, which is crucial for agile operations across supply chains (Ghasemi et al., 2023). Advanced technologies such as collaborative software and artificial neural networks play key roles in optimising these processes and improving decision-making and operational efficiency (Loske and Klumpp, 2021). However, achieving effective CC involves overcoming challenges like integrating diverse technological platforms and managing the complexities of sales forecasting and conflicting stakeholder incentives (Adams et al., 2014; Dubey et al., 2020). Despite these challenges, the strategic integration of CC significantly enhances supply chain performance, particularly in cost savings and operational efficiency (Aharonovitz et al., 2018; Vieira et al., 2009).
This research assesses CC as a moderator within the AI-enhanced supply chain management framework. It examines how alignment in collaborative practices can influence the effects of AIU on key mediators—SCC and LML—ultimately affecting LE. The effectiveness of CC in leveraging AI technologies highlights its crucial role in achieving a resilient and responsive supply chain (Kaynak and Tuğer, 2014; Park et al., 2016).
2.1.5 Logistic efficiency (LE)
LE is a pivotal measure within supply chain management that evaluates logistics operations' effectiveness and operational performance (Neto et al., 2023). This construct encompasses dimensions such as cost-effectiveness, operational leverage and the quality of logistics services. Dmuchowski (2021) emphasises the importance of employing robust methods to measure the effectiveness of logistics activities, highlighting the impact of logistic costs on operational and financial dimensions.
A comprehensive approach to measuring LE integrates economic, functional and quality dimensions, providing a holistic view of performance. Lichocik and Sadowski (2013) underscore the need for an efficient supply chain management strategy that covers both strategic and operational aspects. Additionally, applying sophisticated performance evaluation models, such as the Balanced Scorecard, is instrumental in quantifying various logistics operations and providing actionable insights into their efficiency.
In this research, LE is the dependent construct affected by innovations such as AIU, mediated by SCC and LML and moderated by CC. By comprehensively measuring LE, this study aims to ascertain how technological advancements, like AI, influence the key performance indicators of logistics operations (Belhadi et al., 2021; Zhou et al., 2023). Malhotra (2023) and Imam (2024) also highlight the significant impact of integrating advanced supply chain practices on overall performance, emphasising the need for effective management strategies in the rapidly evolving logistics landscape. A detailed description of all key constructs in e-commerce logistics is presented in Table 1.
2.2 Hypothesis formulation
2.2.1 Artificial intelligence usage and logistics efficiency
Integrating intelligence (AI) into the logistics sector has sparked a remarkable elevation in the e-commerce business, especially in tackling integration challenges and fostering sustainability (Themifstocleous and Irani, 2002). This transformation is mainly ascribed to AI’s swift data processing capabilities, which enable real-time decision-making (Wu and Lin, 2018). AI’s capability to forecast demand accurately, streamline inventory management and optimise delivery routes, as highlighted by Ma and Liu (2011), has been a game-changer. Further, the efficacy of AI-powered predictive analytics in reducing downtime, enhancing inventory supervision and optimising resource distribution has been observed recently by Belhadi et al. (2021), Ghouati et al. (2022), Javaid et al. (2022). Combining AI with cutting-edge technologies like the IoT has breathed new life into warehousing and logistics, ushering in a new era of efficiency and automation (Haldorai et al., 2020). This synergy is particularly evident in the progress made by AI through machine learning and robotic applications in logistics systems that revolutionise last-mile delivery processes (Lingam, 2018; Akbari et al., 2023). Automated AI-driven systems assure uniformity and enhanced output quality, uplifting overall operational proficiency. AI has played a role in achieving cost savings, increased profits and competitive advantages within the logistics industry (Kumar et al., 2017). The improvement of averting downtime and preventing outages, which are crucial for maintenance, is emphasised by Dash et al. (2019). The development of AI in logistics has fundamentally transformed the industry, elevating its management (inventory, routes) and operations to new heights of efficiency (Chen et al., 2015). Moreover, AI’s role in forecasting and demand management has dramatically enhanced accuracy and efficiency, as affirmed by Niranjan et al. (2021). A range of AI techniques, such as smoothing and Bayesian networks, have proven their potency in enhancing the precision of predicting road freight transportation volumes.
Building upon the transformative role of AI in logistics and supply chain management, its application extends to improving decision-making and human capabilities. The power of AI in predicting logistics demand is evident in the use of advanced neural network models, showcasing its wide-ranging impact. This emphasis on AI is crucial for achieving customer satisfaction in logistics, transcending traditional performance metrics, as Loske and Klumpp (2021) explored. AI-driven logistics systems significantly enhance customer satisfaction by optimising delivery schedules, providing personalised services and improving return and exchange management processes. The advancement in logistics efficiency is further exemplified by innovative solutions like self-driving delivery trucks and predictive analytics, as Talevski et al. (2005) pointed out. Moreover, the application of AI leads to improved performance by employing machine learning and multi-objective optimisation techniques for logistics engineering optimisation systems (Malhotra, 2023). The integration of IoT devices into logistics further enhances fleet management and route optimisation, as evidenced by Alavi and Buttlar (2019) and Shao et al. (2019). The impact of the Industrial Internet of Things (IIoT) in improving warehouse operations and process effectiveness is well documented in a recent study by Javaid et al. (2022). This underscores the importance of Iot and IIoT transforming fleet management and routing strategies, significantly advancing logistics and supply chain management. Therefore, theoretical discourse and practical evidence support AI’s crucial role in transforming logistics efficiency in the e-commerce industry. Thus, we propose the following hypothesise:
2.2.2 Mediation impact of supply chain consistency
SCC is crucial in managing logistics performance, ensuring reliability and dependability throughout the supply chain (Liu et al., 2013). Beyond operational efficiency, SCC maintains uniform data for prompt decision-making, using metrics like entropy to measure structure uniformity (Ruiz-Hernández et al., 2019). Alignment among partners and stakeholders is critical for competitive priorities (Saarijärvi et al., 2012). Maintaining harmony in internal and external relations is essential for an integrated supply chain (Krishnapriya and Baral, 2014), influencing the efficacy of AI in logistics.
Implementing AI in supply chain management substantially enhances consistency, reducing purchasing expenses and addressing the bullwhip effect (Kumar et al., 2017). AI’s ability to handle massive datasets facilitates adaptable decision-making (Belhadi et al., 2021). It contributes to robust supply chain frameworks, ensuring continuous operation (Modgil et al., 2022). AI strengthens transparency in last-mile delivery, offering personalised solutions and minimising disruptions (Dash et al., 2019; Zhou et al., 2023).
AI excels in analysing vast datasets, predicting potential issues and recommending proactive strategies, ensuring a consistent flow of information (Dash et al., 2019; Zhou et al., 2023). Using an artificial neural network, Dash et al. (2019) demonstrated an 88.7% gain in overall efficiency, emphasising AI’s transformative impact on supply chain associations and operational efficiency. Previous research by Belhadi et al. (2021), Modgil et al. (2022) and Loske and Klumpp (2021) confirms that AI-driven supply chains exhibit increased consistency and reliability. AI-powered decision-making and predictive analytics significantly reduce errors, supporting the assertion that AI enhances logistics efficiency by promoting a more consistent supply chain (Belhadi et al., 2021; Modgil et al., 2022; Loske and Klumpp, 2021).
In light of this, we put out the following hypothesis:
2.2.3 Mediation impact of last-mile logistics performance
The last mile, the concluding phase of the delivery process, is a crucial performance indicator that profoundly impacts logistics costs and service quality in the modern e-commerce landscape (Jin and Yiyou, 2015). As evidenced by Burlando and Vella (2021), the surge in last-mile traffic due to the growth of e-commerce has elevated delivery costs and inefficiencies, affecting metropolitan infrastructure and overall supply chain efficiency and recognised as a problematic and inefficient phase, LML can incur up to 28% of total delivery costs (Ranieri et al., 2018). Improving LML efficiency is crucial for urban sustainability, reducing costs, enhancing operational efficiency and delivering high-quality service (Jin and Yiyou, 2015; Dash et al., 2019). Given the contemporary demand for fast deliveries, last-mile logistics holds even more significance (Park et al., 2016; Ranieri et al., 2018). Existing literature explores strategies, including metaheuristic optimisation techniques (Silva et al., 2023), creative thinking and teamwork (Brach, 2018), to enhance operational efficiency, reduce energy and time consumption and promote sustainability in urban environments (Ye et al., 2023).
Integrating AI into LML has brought transformative changes, leveraging technologies such as robots, drones and autonomous vehicles to address efficiency, sustainability and cost challenges (Sorooshian et al., 2022). AI’s impact on customer satisfaction is notable, providing accurate delivery time estimates, real-time order status updates and enhancing service quality through delivery personnel (Bányai, 2018). The environmental aspect is also addressed as AI employs intelligent algorithms for order assignment and scheduling, reducing energy consumption (Chen et al., 2021). Early indications suggest that AI-driven predictive analytics optimise routing and scheduling, improving last-mile deliveries and cost savings (Chen et al., 2021). Additionally, the use of AI-powered robots with advanced security measures, such as biometrics, not only enhances accuracy but also significantly reduces last-mile delivery costs (Wang et al., 2022). AI advancements are pivotal, reshaping last-mile logistics to be more efficient, customer-centric and environmentally sustainable. Hence, we hypothesise:
2.2.4 Sequential mediation impact of supply chain consistency and last-mile logistics performance
To understand how AI influences efficiency in logistics, it is necessary to comprehensively explore the causal mechanisms involved, mainly through sequential mediation. Understanding this concept in the causal analysis is crucial for unravelling the complex dynamics that AI brings to improve LE. Sequential mediation is a process where two mediators work one after the other to transmit the impact of an independent variable to a dependent variable. In this context, AI usage (the independent variable) impacts LE (the dependent variable), with this relationship examined through a chain of reactions triggered by one factor leading to another (Pedersen et al., 2007).
Sequential mediation, as explained by Imai et al. (2010), examines the intermediate stages between a treatment (AI usage) and the outcome (improved logistics efficiency). In logistics, this approach is crucial for understanding how AI enhances the reliability of supply chains, positively influencing LML performance. The process involves utilising AI to ensure supply chain consistency, enhancing LML and ultimately contributing to overall LE.
In the sequential mediation model, the first mediator, supply chain consistency, is integral to the process. As highlighted by Ruiz-Hernández et al. (2019) and Saarijärvi et al. (2012), AI technologies enhance supply chain operations by ensuring data consistency and aligning stakeholder objectives. This uniformity positively influences LML operations. The reciprocal relationship between supply chain consistency and LML performance is supported by empirical studies (Bányai, 2018; Wang et al., 2022), indicating a sequential mediation effect. The Systems Theory, as per Nilsson (2006), underpins this model, emphasising that individual improvements in components like the supply chain and LML enhance the system’s performance. Hypothesis H4 addresses this interconnected relationship, highlighting the critical roles of AI usage, supply chain consistency and LML performance in optimising overall LE. So, we hypothesise:
2.2.5 Moderation effect of logistics coordination and collaboration
This section delves into the pivotal role of logistics CC in the nexus between AI usage and enhanced LE. Emphasising the synergy within supply chains, Logistics CC ensures seamless operations through synchronised and cooperative efforts of different stakeholders (Kaynak and Tuğer, 2014). Such collaboration facilitates quality supervision, strengthens partnerships and positively impacts transportation, coordination and delivery costs (Hu et al., 2024). The discussion illuminates how AI can improve supply chain consistency and optimise last-mile logistics performance through effective logistics CC. Moreover, Logistics Coordination and Collaboration encompass enhancing resource allocation and task assignment in collaborative logistics networks, employing advanced techniques such as the virus evolution genetic algorithm for greater efficiency (Ning et al., 2007). Collaboration between manufacturers and third-party logistics providers is a compelling CC example vital for successful production and distribution coordination (Jung et al., 2005). Inefficient logistics processes are often attributed to poor coordination, insufficient collaboration and a lack of trust, with trust being a pivotal factor in successful collaboration (Wang et al., 2022). This synthesis underscores the significance of these elements in realising the potential benefits of integrating AI into logistics operations.
Incorporating a central coordination system with multi-criteria genetic optimisation features, as suggested by Chan et al. (2004), enhances collaboration by solving problems and distributing resources effectively within the supply chain. Moreover, the synergistic impact on company performance is heightened when collaboration and integration are combined, especially with inter-firm coordination technologies (Adams et al., 2014). Recognising the link between AI, coordination and collaboration is crucial for enhancing supply chain consistency.
Hypothesis H5b explores the moderating impact of logistics CC on LML performance, particularly in a complex sector and real-time decision-making demands (Zhu et al., 2023). Souza et al. (2014) emphasise that coordinated collaboration is crucial for enhancing dependability, cost efficiency, resource planning and sustainability in LML. Collaborative initiatives in this sector contribute significantly to cost-effectiveness and pollution reduction compared to independent operations (Konstantakopoulos et al., 2021). Moreover, these efforts reduce travel distances and durations, enhancing delivery efficiency in the courier, express and parcel sectors (Park et al., 2016).
When AI is integrated into LML and coupled with effective CC, it enhances flexibility and operational efficiency. Feng et al. (2017) highlight that this synergy contributes to improved route planning, real-time schedule adjustments and enhanced customer interactions, collectively enhancing overall LE. Research indicates that logistics CC significantly amplify AI’s effectiveness in managing supply chains and LML. Therefore, hypotheses H5a and H5b underscore the pivotal role of logistics CC in shaping the relationship between AI usage and LE, influenced by SCC and LE performance. This underscores the importance of organisational capabilities in effectively leveraging technological innovations in the logistics industry. Thus, we hypothesise:
This article is guided by a research framework (see Figure 1) created to tackle the research questions mentioned at the end of the introduction section. The framework serves as a roadmap for the parts of the paper, offering a consistent exploration of how AIU and LE interact in the e-commerce industry, focusing on the Indian market.
2.3 Network theory
Our review of relevant literature delved into exploring the link between latent variables, psychometric models and behaviour intention design (Epskamp et al., 2016). Bititci and Coll (1998) examined how the effectiveness of Resource Dependence Theory (RDT) and Network Theory (NT) can explain the effect of interrelationships among latent variables. Van et al. (2017) applied RDT and NT to investigate factors affecting the quality of relationships within logistics. Kawa and Zdrenka (2023) shed light on the role of network design, network resilience and collaboration mechanisms in enhancing LE. Boccaletti et al. (2014) provided insights into the capacity of NT to define interactions within complex systems, considering their multi-layered and temporally variable characteristics. Ghomi et al. (2023) underlined the value of collaboration and coordination for logistical efficiency. Sarraj et al. (2014) introduced the idea of interconnected logistic networks and protocols to improve LE. These studies collectively underscore the significance of theoretical frameworks, particularly NT, in studying and enhancing complex logistic systems. NT is beneficial for modelling dynamic interactions in logistics, as it focuses on nodes and connections across diverse networks. It is observed to be effective in simulating these interactions using statistical models based on longitudinal network data and continuous-time Markov chain models.
With its complex web of interactions between suppliers, distributors, retailers and consumers, E-commerce logistics finds AI to be a transformative force due to its capacity to handle extensive data and optimise the flow of entities, information and services. The NT, which analyses interconnections among system elements, plays a crucial role in understanding how the integration of AI can affect information flow and enhance LE. By viewing e-commerce logistics as a network comprising interconnected elements, such as inventory management and delivery systems, NT helps effectively map the influence of AI on components and their interrelations. Our proposed research framework (Figure 1) explores how AI usages (AIU) affect SCC and LML performance, with CC acting as moderators within a networked model. NT focuses on relationships and flow dynamics within networks, which facilitates a comprehensive analysis of the e-commerce logistics network. The approach reveals the role of AI in isolation and highlights its function as part of an integrated system.
3. Research methodology
3.1 Data collection procedure
Information was gathered from managers in Indian logistics e-commerce companies through the survey method. A preliminary study was carried out to validate the structured questionnaire involving three colleagues from a business school in India. Additionally, three logistics professionals and two academicians with over five years of experience teaching and conducting research in the e-commerce logistics sector were consulted to review the questionnaire. After modifying the questionnaire regarding its readability and clarity, the study shared the survey questionnaire with the managers of the Indian logistics firms. Data were collected during four months between July to October 2023. The study selected Indian logistics firms as these firms are aware of adopting AI implementation in new business models to enhance the performance of these logistics firms. The workers were asked qualifying questions: “Are you aware of the logistics efficiency in Indian logistics firms? If yes, name a few companies you know about”; “Are you aware of the artificial intelligence used in this firm?” after obtaining their consent, the eligible workers were provided with a link to the questionnaire. To motivate the workers to participate in the survey, the researcher established rapport with the workers by asking them a few questions at the commencement of the survey, such as “Indicate your favourite logistics company and the reason for the same” and “How and in what position you are involved in the logistics firm”. The analysis has incorporated questions relevant to the study objectives. The study provided the introduction sheet and the questionnaire to explain the concept of AI in logistics e-commerce firms and how it can benefit and improve the performance of logistics firms in India. This information ensured that all the managers and people working in logistics firms had the same understanding. The study informed the participants regarding the anonymity, data privacy and confidentiality of their responses. The reminder emails were shared to encourage the timely completion of the sample questionnaire. Of 315 responses received from the managers in Indian logistics e-commerce firms, and after eliminating incomplete and missing responses, the study used 245 responses for the analysis (Table 2). Of 245 responses, 137 were males and 108 were females. Regarding work experience, 155 respondents had work experience between 10 and 20 years, and most of the firms' ages were more than 20 years (82.54%).
3.2 Measurement procedure
The constructs and scale items used in the questionnaire were obtained from the previous studies (Table 3). The questionnaire was divided into three parts: First, general instructions on how to use AI in the context of logistics firms were provided. Second, all items on the scale used in the study were listed. Third, the demographic profile of the respondents was collected. A five-point Likert scale from strongly disagree (1) to strongly agree (5) was used to measure all construct items. These scales have been adopted from numerous empirical studies, which demonstrated good reliability and validity for all variables (Douglas and Craig, 1983; Xu and Wang, 2024). AI usage was conceptualised as a second-order construct, whereas the other variables were considered single-order constructs. The AI usage construct included four dimensions, namely “Network design/optimisation,” “Purchasing spend analysis,” “Forecasting/Demand management,” and Warehouse operations improvement, with a total of 12 items adapted from Chen et al. (2015), Alavi and Buttlar (2019), Shao et al. (2019), Javaid et al. (2022), Lingam (2018). Four items of logistics efficiency were taken from Gryshko et al. (2018) and Bortolini et al. (2019). Four items scale for the LML performance scale were taken from Silva et al. (2023), Sorooshian et al. (2022), Burlando and Vella (2021) and Wang et al. (2022). Three items of the logistics CC scale were adapted from Dubey et al. (2020), Konstantakopoulos et al. (2021), Adams et al. (2014) and four items of supply chain consistency were obtained from Ruiz-Hernández et al. (2019), Bányai (2018).
3.3 Common method bias
The problem of common method bias (CMB) may arise due to the data collected for both dependent and independent variables simultaneously (Podsakoff et al., 2003). The research utilised Harman’s single-factor analysis with an unrotated factor solution to verify the absence of any CMB issue. The results indicate that the single factor explained 23.64% of the variance, falling below the recommended criterion of 50%, thereby suggesting that CMB is not a concern in this dataset (Podsakoff et al., 2003). In addition, the study conducted a full collinearity technique and found the results were between 1.21 and 2.36, less than the recommended criterion of 3.3 (Kock and Lynn, 2012), indicating CMB is fine in this study. Furthermore, the research employed a sophisticated model that integrated mediation analysis with two mediators and moderated mediation with one moderator. This approach reduces the potential for CMB, as individuals are less likely to employ cognitive maps to visualise such relationships (Podsakoff and Organ, 1986).
4. Results
The study examined the hypothesis of the measurement model through direct, mediation and moderated mediation analyses. Confirmatory factor analysis (CFA) was performed to evaluate the construct items' fit, reliability and validity. Mediation and moderated mediation models were analysed using the PROCESS Macro (Hayes, 2018), while AMOS v25 was employed for CFA. The study aimed not to identify the model that provided the best fit by investigating the covariance and correlations between the variables using a covariance matrix. Hence, PROCESS models were employed for mediation and moderated mediation models instead of AMOS models (Hayes, 2013; Bi and Zhang, 2022).
4.1 Assessment of measurement model
Table 3 indicates the standardised factor loadings, composite reliability (CR), average variance extracted (AVE) and Cronbach’s alpha values. We chose CB-SEM as it is better suited to factor-based models like the proposed conceptual framework and provides better model fit indices (Dash and Paul, 2021). Hair et al. (2016) suggest that all constructs exhibit Cronbach’s alpha values surpassing the recommended threshold of 0.7 (ranging between 0.807 and 0.935), signifying internal consistency for each construct. Subsequently, the study assessed multicollinearity among the constructs through the variance inflation factor (ranging from 1.176 to 2.029), all of which were below the recommended threshold of 5 (Hair et al., 2021). Tolerance values, falling within the range of 0.493–0.850, exceeded the recommended threshold of 0.2. This implies that multicollinearity was not an issue in this study. The research evaluated the construct validity within the comprehensive measurement model (Figure 2), encompassing both convergent and discriminant aspects. The AVE and CR values surpassed the suggested thresholds of 0.5 and 0.7, respectively, affirming convergent validity. For accessing discriminant validity, the square roots of the AVEs were examined (Fornell and Larcker, 1981), as presented in Table 4, with each latent construct demonstrating a greater value than its correlation with any other construct. Additionally, the study conducted the Heterotrait-Monotrait (HTMT) ratio of correlations for the measurement model, with all values falling below the recommended threshold of 0.85 (Henseler et al., 2015). Hence, discriminant validity was achieved. Last, the study demonstrated a good model fit, as recommended by Hair et al. (2016). The results of the CFA indicated a satisfactory fit for the model with CMIN/df = 1.478, CFI = 0.972, TLI = 0.963, IFI = 0.973, NFI = 0.921 and RMSEA = 0.044. Hence, the conceptual framework meets the criteria for reliability, displaying satisfactory fit indices and demonstrating both convergent and discriminant validity.
4.2 Hypothesis testing
4.2.1 Direct and mediation effect
The study applied SPSS and PROCESS Macro (Hayes, 2018) to investigate the hypothesis formulated. Table 5 provides results for direct and mediation analysis. The direct effect was examined using regression analysis. The findings suggest a statistically significant and positive association between the independent variable AIU and the dependent variable LE (effect = 0.744; t = 15.265; p-value = 0.000; R2 = 0.49). Thus confirming H1 and stating that LE increases with increased AIU. Next, the research employed Model 4 of the PROCESS macro to analyse mediation (Hayes, 2018). The mediating variables SSC and LML were analysed (one by one) between AIU and LE. First, a mediation analysis was conducted for SCC, revealing significant direct effects (b = 0.643; t = 11.630; 95% CI [0.534, 0.752]) and indirect effects (b = 0.101; BootLLCI = 0.053, BootULCI = 0.156). This confirms a partial mediation impact between AIU and LE. Thus, H2 validates the mediation effect of SCC.
Next, a mediation analysis was conducted for LML in the relationship between AIU and LE. The results revealed a significant direct effect (b = 0.562; t = 9.103; 95% CI [0.440, 0.684]) and an indirect effect (b = 0.182; BootLLCI = 0.099, BootULCI = 0.273). Consequently, LML partially mediates the connection between AIU and LE, thus confirming H3. Third, a sequential mediation using Model 6 was performed for SSC and LML between AIU and LE. The findings indicate that when SSC and LML mediators are employed simultaneously, the direct effect (b = 0.513; t = 8.056; 95% CI [0.388, 0.639]) and indirect effects are statistically significant. The values for three indirect effects were AIU → SSC → LE was 0.075 (BootLLCI = 0.029, BootULCI = 0.129), AIU → LML → LE was 0.130 (BootLLCI = 0.061, BootULCI = 0.208), AIU → SSC→ LML → LE was 0.026 (BootLLCI = 0.009, BootULCI = 0.049). Hence, H4 is supported in the presence of SSC and LML as sequential mediators between AIU and LE.
4.2.2 Moderated mediation effect
The study examined the moderating effect of logistics CC on the path (a) AIU and LE via SCC and (b) AIU and LE via LML. The research examined the moderating effect using Model 14 of the PROCESS Macro (Hayes, 2018). First, the interaction terms were identified, and then the conditional indirect effect was examined. The moderated mediation results are depicted in Table 6. The moderating variable CC was introduced on the paths (a) AIU → SSC → LE and (b) AIU → LML → LE. The results reveal that the interaction term SCC x CC (with LE as the outcome) was statistically significant (b = −0.651, t = −11.0). Similarly, the interaction term LML x CC (with LE as the outcome) was also significant (b = 0.578, t = 9.025). Consequently, both H5a and H5b were supported. Notably, the values of the conditional indirect effect of CC through the mediator SCC on LE were found to be statistically significant when the CC is low, average or high across all three levels of CC. Therefore, all values were statistically significant, decreasing from 0.121 at a low level to 0.078 at a high level. This suggests that CC weakens the relationship between AIU and LE through SCC. Similarly, the values of the conditional indirect effect of CC through the mediator LML on LE at different CC levels varied across all three levels, and all values were positive and statistically significant, increasing from 0.173 at a low level to 0.199 at a high level. When the CC is low, average or high, the impact of AIU on LE is significantly increased. This implies that CC strengthens the relationship between AIU and LE through LML. Hence, the moderated mediation analysis indicates that moderating variable CC moderate the mediated path AIU on LE via LML. Figure 3 shows the moderation effect of CC on the path (a) AIU→SCC→ LE (weakens) (Figure 3) and (b) AIU→LML→ LE (strengthens) the relationship between AIU and LE (Figure 4).
5. Discussion on results
The study investigated the usage of AI at different levels in logistics e-commerce companies. Multiple hypotheses were formulated and empirically examined using data from companies in India’s e-commerce sector, as discussed below.
Based on the study results, it can be concluded that H1 is supported, indicating that AIU is a significant predictor of LE. In other words, when the level of AIU increases, LE significantly improves. This result aligns with previous research highlighting AI’s critical role in improving several facets of logistics. Hong et al. (2019) and Wu and Lin (2018), for example, highlight the significance of AI in enhancing e-commerce logistics through improved demand forecasting. Similarly, Shao et al. (2019) and Tao et al. (2017) show how using AI for route planning and distribution optimisation may considerably improve retail LE. Our examination shows that these analyses collectively support AI’s observed positive impact on LE.
Hypothesis H2 investigates whether SCC mediates the impact of AIU on LE. This suggests that SCC partially mediates the impact of AIU on LE, meaning that SCC plays a role in conveying a portion of AIU’s influence on LE. The study observed that AIU has dual effects: a direct impact on LE and an indirect impact through enhancing SCC. AIU has a dual impact, directly improving LE and indirectly enhancing SCC, consistent with Niranjan et al.'s (2021) observations on AI’s positive influence on logistics productivity and its cascading effect on supply chain stability. Ghouati et al. (2022) also highlight AI’s role in boosting supply chain efficiency and consistency, reinforcing the benefits of AI in logistics and supply chain management.
In investigating Hypothesis H3, we discovered that LML mediates the AIU–LE dynamic. In addition, the results confirm that LML acted as a partial mediator. This mediation stresses how AIU has a two-fold impact: it directly improves logistics efficiency and indirectly enhances it by optimising LML operations. Both direct and indirect impacts demonstrate AIU’s broad impact on logistics, emphasising the need for direct AI applications and advances in areas such as LML for overall efficiency increase. This result is consistent with other studies that highlight the potential of AI in logistics. Studies conducted by Bányai (2018) discuss how AI can improve the quality of delivery and optimise routes. Additionally, Park et al. (2016) and Wang et al. (2022) have looked into how AI can be applied to forecast delivery times and improve logistics engineering. Loske and Klumpp (2021) also present research findings that support the advantages of AI-based route planning in retail logistics. Wang et al. (2022) and Belhadi et al. (2021) mention that AI significantly enhances service performance and supply chain logistics, supporting our findings and reinforcing this idea.
Our study examined serial mediation, including SCC, LML and AIU, and their combined effects on LE to investigate Hypothesis H4. The analysis showed that there is a strong relationship between AIU and LE. This result shows AIU’s noteworthy direct impact on enhancing logistics efficiency, indicating a sequential mediation route across SCC and LML. This means that AIU directly affects LE and has a ripple effect by improving SCC, LML and LE. The point that there are both direct and sequential indirect effects indicates an intricate interaction of factors where AIU plays a role in improving efficiency through various logistic components. This layered impact of AIU, both direct and cascading via SCC and LML, highlights the varied function of AI in optimising LE. It emphasises the significance of considering both the direct uses of AI and its ripple effects via various logistical operations to achieve complete advancements in the sector.
In order to evaluate Hypothesis H5a, we investigated the moderating effects of logistics network CC on the effectiveness of AI-driven SCC in improving LE. This result emphasises that the effectiveness of AIU in enhancing LE, as mediated by SCC, depends on the degree of CC present in the logistics network. Additionally, an examination was carried out into the conditional indirect effects of AIU on LE via SCC at various levels of CC. When the CC level was one standard deviation above the average (+1 SD), the effect was 0.121. This means that the mediation effect of SCC was stronger when CC levels were higher. The BootSE was 0.034, and the 95% confidence interval ranged from 0.056 to 0.194. The average effect of CC was slightly lower at 0.099. On the other hand, when the CC level was one standard deviation below the average (−1 SD), the effect decreased even more to 0.078.
These results show that when greater CC exist in the logistics network, the impact of AIU on LE through SCC becomes more apparent. This emphasises the importance of effective communication and coordination to make the most of the benefits of AI in logistics. It suggests that improving coordination and fostering collaboration is crucial to achieve the greatest efficiency gains when integrating AI technologies into logistics systems. The results align with the findings of Loske and Klumpp (2021) and Konstantakopoulos et al. (2021).
In order to investigate Hypothesis H5b, we looked at the moderating influence of logistics CC on the performance impact of AIU on LML and, ultimately, LE. This theory is founded on the knowledge that the degree of CC can significantly impact last-mile logistics complexity and the need for quick, real-time decisions (Zhu et al., 2023). The results confirmed hypothesis H5b, showing that CC significantly interacts with the AIU–LML–LE pathway. The study highlights that for AI to improve the efficiency of LML effectively, it is crucial to have efficient coordination and cooperation within the logistics network.
Furthermore, it was evident to observe the conditional indirect effects of AIU on LE via LML at different levels of CC. The impact was 0.173, which is one standard deviation above the average level of CC. This suggests an improved mediation effect of LML when CC is higher. The impact was slightly reduced to 0.186 when considering the average level of CC. In addition, the value was decreased to 0.199, which is one standard deviation below the average level of CC (−1 SD). The BootSE is 0.054, and the 95% confidence interval ranges from 0.095 to 0.309.
The results show that using AI in LML can make logistics operations more efficient. However, this is most effective when there is CC among those involved. This means that although AI is a powerful tool for improving LE, it works best when there is a strong emphasis on CC efforts. Therefore, logistics and supply chain managers need to focus on the technological aspects of integrating AI and developing a company culture emphasising CC. This will help maximise the benefits of AI in complex logistics operations like last-mile delivery.
6. Study implications
6.1 Theoretical implications
This research makes a substantial theoretical contribution to our knowledge of how AI fits into logistics, especially considering the Indian e-commerce market. First, it highlights the vital role that AIU plays in improving LE, confirming the significant impact that AIU has on operational processes, as discussed in previous studies (Wu and Lin, 2018; Ghouati et al., 2022). The research shows that using AI has a positive impact on LE. This builds on previous studies that focused on how AI helps with demand forecasting and optimisation in logistics. Second, the study examines how SCC and LML play a role in the relationship between AIU and LE. This strengthens the current supply chain management theories and makes them more intricate. It shows how the influence of AI is not only direct but also affects supply chain and logistics performance (Javaid et al., 2022). Third, the research’s findings on sequential mediation introduce a nuanced understanding of the cascading effects of AI through various components of logistics operations – SCC and LML, providing a more comprehensive theoretical model of how AI can simultaneously enhance and be enhanced by different elements of the logistics process (Haldorai et al., 2020). Lastly, investigating the moderating impacts of logistical CC in this study provides fresh insight into the circumstances in which AI integration works best. The study shows that the advantages of using AI in logistics depend on the level of CC (Adams et al., 2014). This adds an essential aspect to the theory of implementing AI in logistics, indicating that organisational and relationship strategies should support technological progress. It highlights the importance of organisational capabilities in making the most of technological resources such as AIU. This research offers a comprehensive and interconnected understanding of how AIU influences LE. It emphasises technology’s complex and multidimensional effects on this industry, making valuable contributions to AI and logistics management theory.
6.2 Implications for practice
The results of this study have important implications for the logistics and e-commerce industries, especially in the growing Indian market. The positive impact of using AI in logistics has shown how it can improve efficiency. This emphasises the importance of logistics companies investing in AI technologies. This means using advanced AI tools to optimise routes, predict demand and manage inventory. These tools make operations more efficient and stay ahead of the competition. The relationship between AIU and LE is influenced by SCC and LML performance. Companies should prioritise implementing AI solutions and improve their supply chain networks and last-mile delivery systems. This goal can be realised by conducting routine evaluations and improvements of the supply chain processes to ensure their compatibility with AI capabilities, which allows for a more streamlined and effective logistics operation.
In addition, the study’s results suggest that the level of CC within a company and across the supply chain plays a significant role in the success of implementing AI in logistics. Therefore, to fully take advantage of the potential of AI, logistics companies should cultivate a culture of collaboration and establish robust coordination mechanisms. One way to address this could be by offering training programmes to help staff improve their skills in managing AI-driven systems. Additionally, creating platforms facilitating better communication and information sharing among logistics stakeholders could be beneficial. It may also be helpful to establish partnerships with AI technology providers to ensure continuous support and updates.
Furthermore, in light of the rapidly expanding and dynamic Indian logistics industry, businesses must maintain flexibility and agility in their AI strategies to stay abreast of technological advancements and responsive to shifting customer demands and market dynamics. This research urges logistics companies, especially in emerging markets like India, to actively incorporate AI into their operations. It is not just about upgrading technology but about making comprehensive changes to the organisation, operations and strategy to enhance efficiency and competitiveness in a digitally focused market.
6.3 Limitations and scope of future work
This study illuminates AI integration in Indian e-commerce logistics yet carries limitations and indicates avenues for future research. First, this study focuses on the Indian logistics sector, which may limit generalisability and is culturally different from the western markets. This may lead to differences in the logistics efficiency related to AI in the Indian logistics sector. Hence, the findings cannot be extrapolated to diverse regions and sectors for comparative analysis. In the future, researchers may consider including multiple countries or different sectors in their studies. Researchers may implement longitudinal studies to test the conceptual framework in other geographical or cultural regions. This would allow for a comparative analysis to validate or contrast the findings in various contexts. Second, we rely on data reported by logistics managers, which could introduce subjective biases and restrict our understanding of the actual situation. In future studies, researchers could use quantitative data and qualitative insights from frontline employees or customers to better understand AI’s practical implications in logistics. This mixed-methods approach would provide a deeper, more nuanced understanding of the topic. The use of a 1–5 Likert scale assumes equal variance between scale points, which may not reflect the actual situation (Tanujaya et al., 2022). This scale produces qualitative data, limiting analytical options (Li, 2013). There is also an ongoing debate about whether Likert items should be treated as ordinal or interval data, affecting the choice of statistical analysis (Harpe, 2015). Additionally, issues like response bias and central tendency bias pose further challenges to the validity and reliability of the findings (Kusmaryono et al., 2022). Third, the study mainly focuses on how AI affects logistics efficiency but does not detail the specific AI technologies or their specific contributions. Future researchers could explore the specific roles of different AI technologies, such as machine learning, natural language processing and robotics, and how they affect different logistics and supply chain management areas. Furthermore, given the fast-paced advancements in AI, it is crucial to conduct ongoing research to stay informed on the latest developments and understand how they impact the field of logistics. Fourth, this study brings attention to the importance of CC. However, it needs to delve deeper into how these elements can be effectively fostered or the difficulties that may arise during their implementation. In the future, researchers could explore how logistics firms can create a more collaborative and coordinated environment. This would involve looking at their strategies and challenges, particularly when integrating AI. It would be helpful to have case studies or action research that offer practical insights on effectively managing the organisational change that comes with adopting AI in logistics.
7. Conclusions
The study on the application of AI in e-commerce logistics, particularly in the dynamic landscape of Indian trade, yields significant insights. It establishes AI not merely as a supplementary element but as a potent force revolutionising logistics efficiency through advanced route optimisation, dynamic inventory management and real-time monitoring. SCC and LML performance are pivotal components, akin to crucial parts in a large machine, enhancing AI’s capabilities by ensuring a stable and flexible supply chain. The impact of sequential mediation reveals the intricate layers of AI’s influence, showcasing how improving supply chain consistency with AI creates a ripple effect, leading to superior LML performance and unparalleled efficiency in logistics operations. This understanding provides a new perspective on maximising the potential of AI in reshaping logistics operations by integrating and connecting strategies. The study underscores the importance of collaboration and coordination, highlighting how a human-centric focus can significantly amplify the impact of AI-driven technological advancements. Beyond its academic contribution, the study is a strategic framework for advancing the e-commerce logistics industry, poised for substantial development. It emphasises AI’s critical role in enhancing operational efficiency and transforming the logistics landscape, offering valuable insights for industry practitioners, policymakers and researchers. The research results extend far beyond academia, representing a significant achievement in envisioning a more intelligent and interconnected e-commerce logistics ecosystem.
Figure 1
Conceptual framework
[Figure omitted. See PDF]
Figure 2
Measurement model
[Figure omitted. See PDF]
Figure 3
Moderation effect of CC on the path AIU → SCC → LE
[Figure omitted. See PDF]
Figure 4
Moderation effect of CC on the path AIU → LML → LE
[Figure omitted. See PDF]
Table 1
Key constructs in evaluating the power of AI usages in e-commerce
| Construct | Description | References |
|---|---|---|
| Artificial Intelligence Usages | ||
| Network design/optimisation | Effective network design and optimisation powered by AI plays a role in improving efficiency and responsiveness. This key component showcases how online retail businesses leverage AI technologies to streamline their supply chain and distribution networks. It involves utilising AI for optimizing networks to improve shipping, lower transportation expenses accelerate delivery speeds and introduce real-time monitoring systems for logistical choices. Collectively these aspects assess the impact of AI, in evolving network management into cost-efficient operations | Pournader et al. (2021), Leung et al. (2020) |
| Purchasing Spend Analytics | This item is pivotal in optimising procurement processes and enhancing operational decision-making. This component of AIU is defined by the extent to which e-commerce businesses leverage AI tools to enable real-time tracking and comprehensive monitoring of logistics operations. It encompasses: Real-time Shipment and Delivery Tracking, Comprehensive Supply Chain Visibility, Data-Driven Decision Making | Dash et al. (2019), Shao et al. (2019) |
| Forecasting/Demand Management | This item is crucial for enhancing operational efficiency and accuracy. This component of AIU is defined by the extent to which e-commerce businesses employ AI tools to refine and optimise their forecasting and demand management practices. More specifically it includes: Integration of Forecasting/Demand Management; Data-Driven Problem Solving and enhancement of Inventory Management | Jauhar et al. (2023), Wang et al. (2022) |
| Warehouse Operations Improvement | Warehouse Operations Improvement is critical for maximising efficiency and accuracy in storage and distribution processes. This aspect of AIU is characterised by the extent to which e-commerce businesses utilise AI tools to enhance various facets of their warehouse operations. It includes: Enhanced Inventory Management and Storage; Improved Order Fulfillment Accuracy and Speed; Reduction in Manual Errors and Operational Costs | Dash et al. (2019), Belhadi et al. (2021) |
| Supply Chain Consistency | Supply Chain Consistency is a critical construct that assesses the reliability and uniformity of operations across the entire supply chain within an e-commerce context. This construct is defined by the extent to which an e-commerce business can maintain: Adherence to Delivery Schedules; Uniformity in Quality and Inventory Management; Coordination Across Supply Chain Components; Operational Dependability Amidst Changes | Hazen et al. (2014), Modgil et al. (2022), Hu et al. (2024) |
| Last-Mile Logistics | Last-Mile Logistics is a crucial construct in assessing the impact of AI on the terminal phase of the delivery process. This construct is defined by the extent to which an e-commerce business leverages AI technologies to enhance efficiency and service quality in last-mile delivery. It encompasses: Increased Delivery Efficiency; Reduction in Delivery Times and Operational Expenses; Improvement in Customer Satisfaction and Service Quality; Overall Efficiency in Logistics Operations | Grover and Ashraf (2023), Simoni et al. (2020) |
| Coordination and Collaboration | Coordination and Collaboration are crucial for leveraging AI to enhance supply chain effectiveness and operational synergy. This construct is defined by the extent to which AI-enabled logistics practices facilitate cooperative efforts among various elements of the supply chain. It includes: Improvement in Supply Chain Consistency; Enhancement of Decision-Making and Operational Synergy; Impact on Last-Mile Logistics Solution | Dubey et al. (2020), Kawa and Zdrenka (2023) |
| Logistics Efficiency | Logistic Efficiency is a critical construct that assesses the effectiveness and performance of logistics operations within an e-commerce context. This construct is defined by the extent to which an e-commerce business’s logistics operations achieve: Delivery Speed and Reliability; Responsiveness to Demand and Supply Changes; Resource Utilization; Cost-effectiveness Relative to Industry Benchmarks | Dmuchowski (2021), Zhou et al. (2023) |
Table 2
Demographic profile of the respondents
| Demographic variable | Categories | Frequency (n) | Percentage (%) |
|---|---|---|---|
| Gender | Male | 137 | 55.92 |
| Female | 108 | 44.08 | |
| Age groups | 18–35 | 102 | 41.63 |
| 36–50 | 84 | 34.29 | |
| 50 year+ | 59 | 24.08 | |
| Year of experience | <10 years | 47 | 19.18 |
| 11–20 years | 155 | 63.27 | |
| 21 years+ | 43 | 17.55 | |
| Functional area of respondents | |||
| Last-Mile Delivery Services | 64 | 26.12 | |
| Warehouse Management and Operations | 37 | 15.10 | |
| Road Transportation and Trucking Services | 29 | 11.84 | |
| Distribution Center Operations | 26 | 10.61 | |
| Reverse Logistics and Returns Management | 19 | 7.76 | |
| Supply Chain Analytics and Technology Solutions | 17 | 6.94 | |
| E-commerce Fulfilment Services | 14 | 5.71 | |
| Others | 39 | 15.92 | |
Source(s): Authors’ work
Table 3
Description of constructs and items of the study and measurement assessment
| Construct and items | SFL |
|---|---|
| Artificial Intelligence Usage (AIU) | |
| Network design/optimisation (ND/O) (AVE = 0.688; CR = 0.868; α = 0.870) | |
| AIU1: Our company makes extensive use of AI-driven network optimisation tools to improve delivery | 0.794 |
| AIU2:Network optimisation has cut our fleet’s costs and speed up delivery | 0.796 |
| AIU3: Real-time network/route management systems help our company make logistics decisions | 0.894 |
| Purchasing spend analytics (PSA) (AVE = 0.653; CR = 0.848; α = 0.84) | |
| AIU4: Real-time shipment and delivery tracking improves our logistics operations | 0.872 |
| AIU5: We can see our entire supply chain thanks to real-time tracking and monitoring | 0.863 |
| AIU6: Our logistics operations often make critical decisions using real-time tracking system data | 0.674 |
| Forecasting/Demand management (F/DM) (AVE = 0.819; CR = 0.932; α = 0.935) | |
| AIU7: Our company heavily integrates forecasting/demand management practices to improve logistics efficiency and accuracy | 0.913 |
| AIU8: We are much better able to solve logistical problems and make decisions with the data analysis from forecasting/demand management practices | 0.902 |
| AIU9: We have significantly enhanced our inventory management and decreased operational downtime through the integration of forecasting/demand management practices | 0.901 |
| Warehouse Operations Improvement (WOI) (AVE = 0.632; CR = 0.837; α = 0.841) | |
| AIU10: Warehouse operations has greatly improved inventory management and storage at our company | 0.78 |
| AIU11: Warehouse operations has greatly improved our order fulfilment accuracy and speed | 0.805 |
| AIU12: Warehouse operations has greatly reduced manual errors and operational costs in warehouse management | 0.799 |
| Coordination and Collaboration (C&C) (AVE = 0.603; CR = 0.818; α = 0.816) | |
| CC1: How much does logistics network coordination and collaboration improve AI-driven supply chain consistency for logistics efficiency? | 0.629 |
| CC2: AI-enabled logistics coordination and collaboration has improved logistics decision-making, problem-solving and operational synergy in our firm | 0.826 |
| CC3: How does logistics network coordination and collaboration affect AI-enabled last-mile logistics solutions' adaptability and efficiency? | 0.856 |
| Supply Chain Consistency (SCC) (AVE = 0.613; CR = 0.860; α = 0.807) | |
| SCC1: To what extent does our supply chain maintain a consistent record of adhering to delivery schedules without substantial disruptions or delays? | 0.59 |
| SCC2: To what degree do we provide a uniform level of quality and precision in inventory management across every phase of our supply chain? | 0.937 |
| SCC3: How well do our supply chain components—procurement, manufacturing, logistics, etc.—coordinate to ensure smooth operations? | 0.851 |
| SCC4: To what extent does our supply chain maintain operational consistency and dependability while adapting to changes (such as supply disruptions and fluctuations in demand)? | 0.708 |
| Last Mile Logistics (LML) (AVE = 0.57; CR = 0.838; α = 0.838) | |
| LML1: Our use of AI technology in last-mile logistics has increased delivery efficiency significantly | 0.815 |
| LML2: As a result of integrating AI into our last-mile logistics, delivery times and operational expenses have been reduced significantly | 0.856 |
| LML3: AI-driven last-mile logistics innovations have dramatically improved customer satisfaction and service quality | 0.75 |
| LML4: We use AI in our last-mile logistics operations, which has made our logistics system much more efficient overall | 0.565 |
| Logistics Efficiency (LE) (AVE = 0.709; CR = 0.905; α = 0.891) | |
| LE1: Our logistics operations constantly meet their objectives in terms of delivery speed and reliability | 0.935 |
| LE2: Our logistics system responds quickly to demand and supply changes, ensuring efficient operations | 0.911 |
| LE3: Our logistics operations effectively utilise resources (like manpower, vehicles and technology) to maximise operational efficiency | 0.867 |
| LE4: In comparison to industry benchmarks, our logistics operations exhibit a commendable level of cost-effectiveness | 0.617 |
Note(s): SFL: Standardised factor loading; AVE: Average variance extracted; α: Cronbach’s alpha coefficient; CR: Composite reliability
Source(s): Authors’ work
Table 4
Correlation matrix, descriptive statistics and discriminant validity
| AIU | CC | SCC | LML | LE | MVS | MaxR(H) | |
|---|---|---|---|---|---|---|---|
| Artificial Intelligence Usage (AIU) | 0.810 | 0.442* | 0.594* | 0.738* | 0.767* | 0.595 | 0.893 |
| Coordination and Collaboration (CC) | 0.413 | 0.777 | 0.222* | 0.365* | 0.287* | 0.170 | 0.847 |
| Supply Chain Consistency (SCC) | 0.589 | 0.286 | 0.783 | 0.594* | 0.588* | 0.346 | 0.919 |
| Last Mile Logistics (LML) | 0.736 | 0.312 | 0.506 | 0.755 | 0.698* | 0.542 | 0.866 |
| Logistics Efficiency (LE) | 0.771 | 0.2 | 0.533 | 0.64 | 0.842 | 0.595 | 0.939 |
Note(s): Diagonal values italic and underlined indicate the square roots of AVEs (Fornell–Larcker criterion)
* values in italics represent HTMT (Henseler et al. criterion)
Source(s): Authors’ work
Table 5
Direct and mediation path analysis
| Hypothesised path | Effect | t-value | SE/BootSE | Bootstrap 95% CI | Hypothesis | |
|---|---|---|---|---|---|---|
| LLCI | ULCI | |||||
| H1: AIU → LE | 0.744 | 15.265 | 0.049 | 0.648 | 0.84 | Supported |
| H2: AIU → SCC → LE | ||||||
| Direct effect | 0.643 | 11.63 | 0.055 | 0.534 | 0.752 | Supported, Partial mediation |
| Indirect effect | 0.101 | 0.027 | 0.053 | 0.156 | ||
| H3: AIU → LML → LE | ||||||
| Direct effect | 0.562 | 9.103 | 0.062 | 0.44 | 0.684 | Supported, Partial mediation |
| Indirect effect | 0.182 | 0.044 | 0.099 | 0.273 | ||
| H4: AIU → SCC →LML → LE | ||||||
| Direct effect | 0.513 | 8.056 | 0.064 | 0.388 | 0.639 | Supported, sequential mediation |
| Indirect effect | 0.026 | 0.01 | 0.009 | 0.049 | ||
Note(s): AIU - Artificial Intelligence usage; LE - Logistics Efficiency; SCC = Supply chain consistency; LML = Last mile logistics performance
Source(s): Authors’ work
Table 6
Moderation analysis
| Moderating effect of CC | Decision | |||||
|---|---|---|---|---|---|---|
| Interaction effects | B | t | p-value | LLCI | ULCI | |
| H5a: Interaction effect of CC on path AIU → SCC→ LE | 0.651 | 11 | <0.05 | 0.534 | 0.767 | H5a Significant, Supported |
| Conditional indirect effect of AIU on LE (via SCC) at different levels | Effect | BootSE | LLCI | ULCI | ||
| +1 SD (CC) | 0.121* | 0.034 | 0.056 | 0.194 | ||
| Mean (CC) | 0.099* | 0.027 | 0.049 | 0.157 | ||
| −1 SD (CC) | 0.078* | 0.033 | 0.014 | 0.147 | ||
| H5b: Interaction effect of CC on path AIU → LML → LE | B | t | p-value | LLCI | ULCI | |
| 0.578 | 9.025 | <0.05 | 0.452 | 0.704 | H5b Significant, Supported | |
| Conditional indirect effect of AIU on LE (via LML) at different levels | Effect | BootSE | LLCI | ULCI | ||
| +1 SD (CC) | 0.173* | 0.05 | 0.079 | 0.275 | ||
| Mean (CC) | 0.186* | 0.044 | 0.104 | 0.277 | ||
| −1 SD (CC) | 0.199* | 0.054 | 0.095 | 0.309 | ||
Note(s): AIU – Artificial Intelligence usage; LE – Logistics Efficiency; SCC = Supply chain consistency; LML = Last mile logistics performance; CC = Coordination and collaboration)
* indices significant value at 0.05 level
Source(s): Authors’ work
© Emerald Publishing Limited.