Green Last-Mile Delivery Using AI-Enabled Electric Fleets: Aligning with Millennial Values

The dynamism of e-commerce has drastically changed consumer purchasing habits and increased the requirements of efficient, reliable, and sustainable last-mile delivery (LMD) services. The LMD is generally regarded as the most costly and complex part of a logistics chain moving products from a distribution center or local logistics hub to the end consumer [1, 2]. Although LMD represents the shortest geographical distance in the supply chain, it can still contribute to a significant portion of the total logistics costs because of traffic jams in cities, non-regular delivery routes, delivery failures and growing customer demands in terms of delivery velocity and convenience [1-3]. Additionally, the rise of online shopping has led to more freight transport activity in the city which has led to more energy use, GHG emissions and strain on existing transport systems [4, 5].

Artificial Intelligence (AI) has become an important technology solution to these operational and environmental challenges in modern logistics systems. AI represents one of the core enabling technologies of Industry 4.0, and it relies on technologies such as Machine Learning (ML), Deep Neural Networks (DNNs), predictive analytics, and Natural Language Processing (NLP) to enhance the efficiency and effectiveness of logistics networks [1, 6]. By leveraging real-time and historical data from various sources such as traffic conditions, weather forecasts, customer orders, and delivery history, AI-driven systems can make more efficient routing decisions, predict customer demand, and enhance fleet scheduling [1, 4]. Current research indicates that by optimizing routes through AI, companies can achieve substantial improvements in delivery delays, operational cost savings, minimisation of empty vehicle runs, and enhanced quality of customer services via predictive scheduling and communication technologies [1, 6]. However, many of the findings to support such benefits are based on simulation studies and empirical studies in realistic logistics environments are desired [6, 7].

In addition to technological innovation, sustainability is a strategic focus in urban logistics. The traditional delivery service models (LMDs) mainly use internal combustion engine (ICE) vehicles, which cause more air pollution, traffic congestion and carbon emissions in urban areas [3-5]. With ambitious climate targets and net-zero emission targets by governments and organisations, Electric Vehicles (EVs) have become more and more of a consideration for the transport of goods in urban areas [5, 8]. EVs have multiple environmental benefits such as emission reduction of greenhouse gases (GHG), minimizing noise pollution, and the conservation of fossil fuels [3, 5]. Their use in last mile logistics is still controlled by difficulties faced with them like battery ranges, charging infrastructure, charging time, and too high of capital investments [5, 9]. Recent studies indicate that intelligent route planning, charging optimization, energy-aware fleet scheduling, and battery management systems are potential solutions that can be addressed using AI technologies, which would improve the operational feasibility and sustainability of EV-based logistics operations [4, 5].

Consumer behavior has become a major factor influencing the adoption of sustainable logistics, in addition to operational and environmental factors. Millennials, those born between 1981 and 1996, are one of the biggest and most significant consumer groups in the world [10, 11]. Millennials are digital natives who use digital technologies extensively, are environmentally aware and have an increasing preference for socially responsible products and services [12, 13]. Sustainability issues have been found to be more and more important factors in consumer buying decisions such as delivery methods and logistics companies' preferences [14, 15]. Online social media also play a role in strengthening consumption behavior in terms of sustainability via electronic word-of-mouth (eWOM) and influencers, as well as establishing subjective social norms that inspire environmentally responsible purchasing [15-17]. However, millennials should not be considered a reactive group, but an active part of the demand side whose values, expectations and purchasing patterns impact upon the decisions made by organizations in relation to the implementation of sustainable delivery systems.

There are various theoretical frameworks that offer insights into the relationship between consumer behavior and the adoption of sustainable technology. According to the Theory of Planned Behavior (TPB), attitudes, subjective norms, and perceived behavioral control environmentally responsible behaviors [11, 12]. The Technology Acceptance Model (TAM) has been identified as a key theoretical framework within the research on technology acceptance, especially in logistics services, which are increasingly supporting customer operations with AI features such as real-time tracking and predictive notifications to help reduce inefficiencies [18]. In the same way, Diffusion of Innovation (DOI) theory helps understand the process in which new technologies diffuse through social groups or systems, and Stakeholder Theory focuses on the economic, environmental, and social aspects of sustainable logistics projects and their value for society [19-21]. In total, these theories imply that consumer values on sustainability and technology acceptance can be key mechanisms that impact the adoption and diffusion of AI-powered EV fleets.

Although substantial research has examined AI applications in logistics, EV adoption, and sustainable consumer behavior, the literature remains fragmented. Existing reviews have generally focused on individual dimensions of the problem, such as green vehicle adoption [5], logistics optimization technologies [6], or the environmental implications of EV deployment [22]. More recent studies have explored the operational synergy between AI and EV technologies in urban logistics systems [3]. However, limited attention has been devoted to understanding how consumer sustainability preferences, particularly those of millennials, influence the adoption of AI-enabled sustainable logistics solutions. Furthermore, previous studies frequently report operational and environmental performance improvements without clearly distinguishing between empirical evidence, simulation-based findings, and industry projections, thereby limiting the transparency and interpretability of reported outcomes [6].

In order to overcome these, this study carries out a Systematic Literature Review (SLR) of the intersection between the three domains: AI, EV fleets and millennial sustainability values, in the context of last mile delivery. Specifically, the review will aim to: (1) understand the effects of AI technologies on delivery efficiency and operational performance; (2) consider the environmental and operational implications of AI-supported EV fleet management; (3) explore how millennial sustainability values contribute to the acceptance and diffusion of sustainable logistics solutions; and (4) create an integrated framework that combines technological innovation, sustainability outcomes and consumer behaviour.

This review is a significant addition to the literature in 3 ways. First, it brings together disarticulated research from multiple streams of studies in logistics, transportation, sustainability, and consumer behaviour. Second, it lays the groundwork for an evidence-classification process, which classifies studies by the type of evidence given (e.g., empirical peer-reviewed, peer-reviewed simulation, conference paper, thesis, industry report, or preprint), thus improving methodological transparency. Third, it provides an integrated conceptual framework that places values among millennials as a demand-side driver of the adoption of AI technologies among EV fleets in LMD systems. The study integrates technological, environmental, and behavioural angles, offering a complete view of the potential of intelligent and sustainable logistics systems to shape the future of urban freight distribution.

3.1 Introduction

This systematic review was based on the analysis of literature related to the influence of AI, EVs, sustainability and consumer behaviour in LMD from 47 studies. The literature review shows that during the last few years, the academic and industrial interest in sustainable solutions for urban logistics has been growing steadily, with some countries dedicated to this cause to a greater extent than others. The majority of the studies were published between 2022 and 2025, signifying that these areas of logistics research and EV-based delivery are still nascent and developing fast.

Geographically, the evidence base was limited to developed economies, and mainly Europe, North America and China. Empirical studies, on the other hand, are comparatively limited in emerging economy countries like Africa and Latin America, despite the growing urbanization and logistical difficulties in these countries. The imbalance reveals a critical need for evidence specific to developing markets.

As in the aim of methodological transparency, the included studies were classified into types of publications as it is presented in Table 6.

Table 6. Distribution of included studies by publication type

Table 6 presents the number and proportion of studies in each publication category.

Most of the sources were peer-reviewed journal articles, enhancing the credibility of the review conclusions, while the use of grey literature helped to get coverage on new technologies and the recent trends of the industry that could not be found much in the academic literature.

3.2 Artificial Intelligence applications in last-mile delivery

The review identified AI as a central enabler of operational efficiency in last-mile logistics. Three dominant application areas emerged from the literature: route optimization, demand forecasting, and customer communication [1, 23, 24].

Within the literature reviewed, the most widely studied application of AI was found to be route optimization. Typically, the performance of ML, Reinforcement Learning (RL) and other AI based optimization algorithms is compared to traditional routing heuristics and human-planning methods. In the reviewed simulation-based studies, travel distance was consistently lowered by 12-22% and operating costs by 15-40% with the assistance of AI-based routing, but this is the projected improvement for the simulated operation and does not necessarily mean that the same reduction would occur in a large-scale deployment in the field [1, 6, 23]. Furthermore, predictive demand forecasting with ML models also helped to improve the accuracy of the forecasts by up to 50% compared to the traditional forecasting method, which showed a better use of the fleet and fewer empty movements [1, 6, 23]. AI-powered customer communication systems that use NLP and predictive scheduling further increased delivery reliability by minimizing failed deliveries and delivering more satisfying customer communications compared to traditional notification systems [2, 22]. The evidence suggests that AI has a significant positive effect on operational performance, but the majority of evidence comes from simulation studies, suggesting that empirical evidence is needed to validate the AI's impact on operational performance.

The quantified improvements in performance shown in Table 7 have been collated from previous studies and thus should be treated as secondary evidence and not as findings of the present review. The improvements reported most frequently were from simulation-based experiments or modelling investigations involving comparisons between the AI enabled approach and traditional routing/scheduling, forecasting, or customer communication practices. While these studies consistently show the promise of AI to increase the efficiency of operations, the reported performance gains can differ due to assumption made in modelling, data used, logistics situations, and optimization algorithms. Results therefore require further validation from large-scale empirical studies in realistic operating conditions.

Table 7. Quantified impact of AI applications in last-mile delivery

Overall, the reviewed studies consistently indicate that AI improves operational efficiency in LMD through enhanced routing, forecasting, scheduling, and customer communication. Nevertheless, because most available evidence is simulation-based, further empirical investigations are required to confirm these findings in real-world logistics environments.

3.3 Artificial Intelligence-enabled Electric Vehicle adoption in last-mile delivery

EVs were consistently stated as one of the important means to curb the effects on the environment of urban freight transport [3, 8, 19]. Literature has identified several advantages of EV deployment, such as reduction in GHG emissions, less noise pollution, and higher energy efficiency over ICE vehicles [5, 9, 22]. However, a lot of challenges exist, especially operational and infrastructural difficulties, which hinder the widespread adoption of these benefits. The top three problems identified in the reviewed studies [6, 18, 19] were range anxiety, charging infrastructure and high purchase prices. A constant theme in identifying the barriers was the need for AI to be a key enabler to address them. Energy-aware routing systems minimize battery consumption and charging management algorithms maximize vehicle availability and reduce disruption during operation [6, 9, 23]. However, a few studies highlighted the fact that EV fleet environmental performance is still influenced by the CO₂ intensity of electricity generation and battery manufacturing [3, 5, 8, 9].

3.4 Sustainability outcomes of Artificial Intelligence-enabled Electric Vehicle fleets

The reviewed studies consistently linked AI-enabled EV fleets with improvements across environmental, economic, and social dimensions of sustainability. Environmental outcomes were the most extensively reported, particularly reductions in GHG emissions and energy consumption [3, 5, 6, 22, 25].

Economic benefits included lower operating costs, improved vehicle utilization, and greater delivery efficiency resulting from AI-assisted fleet management [1, 2, 23]. Social benefits included reduced traffic congestion, lower urban noise levels, and enhanced customer satisfaction through improved service reliability (Table 8) [16, 17, 21].

Although the overall sustainability impact of AI-enabled EV fleets was positive, several studies emphasised that these environmental benefits are influenced by the carbon intensity of local electricity generation, battery manufacturing processes, and lifecycle battery emissions. Consequently, the long-term sustainability of EV fleets depends not only on vehicle electrification but also on cleaner energy systems and sustainable battery lifecycle management [8, 25]

Table 8. Sustainability outcomes associated with AI-enabled EV fleets

3.5 Millennial consumer perspectives on sustainable delivery

Millennials emerged as a key stakeholder influencing the adoption of sustainable delivery solutions. A positive correlation between EA and attitudes towards sustainability and support for green logistics initiatives was consistently reported in the literature [10, 11, 26]. Research based on the TPB showed that environmental attitudes and subjective norms have an important effect on green purchase intentions [12, 26-28]. Likewise, studies using the TAM demonstrated the positive relationships between perceived usefulness and ease of use, and their positive impacts on acceptance of technology-oriented logistics services [18, 29].

Social media emerged as a key factor in influencing attitudes and behavioral intentions for sustainability. Millennials' pro-environmental purchasing increased through exposure to influencer messages, eWOM, and peer-generated pro-environmental messages [14-17]. But there was still an intention-behaviour gap. While the attitudes of millennials in relation to sustainable delivery methods were generally positive, there were a number of issues relating to delivery speed, convenience and extra costs which posed barriers to actual use of these delivery methods [26-28, 30].

3.6 Integrated evidence on Artificial Intelligence, Electric Vehicles, and millennial sustainability values

The final synthesis shows that the adoption of technology and consumer demand complement each other and work best together for sustainable LMD. AI enhances operational efficiency and transparency of delivery, EV fleets reduce GHG emissions and improve energy efficiency, and millennials are driving the market for sustainable logistic services [1, 3, 22, 23].

There is emerging evidence that an integrated system of AI-EV delivery provides better results than the technological interventions alone. Research indicated AI optimization, along with electrified delivery fleets, improved operational efficiency, environmental performance, and customer satisfaction [2, 3, 5, 23]. The results corroborate the proposed framework in which technological capabilities and consumers' sustainability values are complementary drivers for the adoption of green last mile delivery.

Results of the synthesis show that technological innovation is not enough to realize sustainable logistics transformation. Instead, the combination of AI features and EV adoption, along with consumer preferences for sustainable delivery, will be the main driver behind strategy for EV adoption in the last-mile sector (Table 9).

Table 9. Mapping of key findings to research objectives

5.1 Conclusion

This study systematically reviewed and analyzed the literature related to the influence of AI integrated EV fleet and millennial sustainability values on the sustainable LMD concept. A total of 47 studies were identified, screened, evaluated, and synthesized in this study following the PRISMA framework, between 2017 and 2026. The review included empirical studies, modelling and simulation research, systematic reviews, conference papers, theses and industry reports to gain a comprehensive understanding of sustainable LMD [37, 43].

The study reveals how AI has emerged as an essential tool for achieving sustainable logistics operations. Scale logistics solutions such as AI systems noted for route optimisation, predicting demand, and intelligent communication systems have had a distinct and positive impact on operational efficiency, all while cutting delivery durations, mileage, and logistics expenses [1, 2, 22, 23]. The review also highlighted how AI systems in decision-making processes contribute to the reliability and satisfaction of the services by allowing real-time responses to the ever-changing dynamics of urban logistics [4, 23, 24].

The study also validates the potential of EVs to decarbonise urban freight transport. In the literature surveyed, EV use was linked to decreases in GHG emissions, increases in energy efficiency and decreases in the long-term operating costs of EVs [3-6, 22]. Charging infrastructure, electricity generation sources, managing battery life, and supportive policy environments, however, significantly impact the sustainability performance of the EV fleet [19, 28, 29].

The review confirmed that millennial consumers are an influential force that contributes to the adoption of sustainable logistics solutions from a behavioural perspective. The main factors that influence green delivery preferences were environmental attitudes, social influence, sustainability awareness, and digital involvement [16, 21, 26, 27]. The results lend support to the TPB, TAM and DOI theory, in that they indicate that not only is technology performance important, but also perceived usefulness, compatibility with lifestyle, and social norms affect the adoption of sustainable delivery [18, 34, 35].

More than anything, the review shows there is no one-size-fits-all solution for sustainable last-mile deliveries in the form of technological fixes. Instead, the successful implementation relies on interaction between the AI technologies, EV fleet management systems, supportive infrastructure, stakeholder cooperation and consumer acceptance [4, 8, 22, 31, 37]. The data indicates that AI acts as the operational intelligence that optimizes EV performance and millennial sustainability values are driving the market demand needed for EV uptake. Sustainable LMD should therefore be looked at as a socio-technical ecosystem, and not just a mere technological approach [40, 41, 44].

Overall, this review adds to the sustainable logistics body of knowledge by collating the sparse knowledge from technological, environmental and behavioural perspectives. It suggests the increased alignment of digital transformation with sustainability goals and offers a solid base to build on for future research and application in the dynamic domain of urban logistics.

The study makes several theoretical and practical contributions to sustainable logistics research.

5.2 Theoretical contributions

First, the review adds to existing theory by combining three theories TPB, TAM, and DOI themselves combined into a single framework for exploring sustainable LMD. It integrates multiple theories, such as TPB, TAM, and DOI, into a single framework when investigating sustainable LMD. Research on technology adoption, on environmental sustainability, on consumer behaviour are generally investigated separately from each other. This review shows that these dimensions give rise to sustainable logistics outcomes as a result of their interaction, not directly from any one dimension alone [11, 29, 36].

Secondly, the study demonstrates that sustainability concepts, such as carbon transparency, green delivery and environmental accountability are becoming more prevalent in perceived usefulness [18, 29, 30, 34].

Third, the review extends TPB by identifying the significant role played by social media and digital communities on influencing sustainable consumption related subjective norms among millennials [14-16, 21].

5.3 Practical contributions

The results offer insightful recommendations for logistics providers, retailers, policy makers and technology makers. The review shows how investments in AI-powered logistics systems can have a beneficial effect on operations, while also contributing to environmental goals. It also emphasizes the role of the engagement strategy for consumers, the development of charging infrastructure, and collaboration among stakeholders in the context of sustainable logistics transformation [1, 3, 5, 22].

5.4 Recommendations

Based on the results of this review, the following recommendations are made.

Recommendation 1: Increase empirical validation of AI-enabled logistics systems. Although simulation studies consistently report operational improvements, more large-scale field experiments and longitudinal implementation studies are needed to validate the scalability and effectiveness of AI-powered logistics systems under real-world operating conditions [3, 31].

Recommendation 2: Improve EV infrastructure development.

Investment in charging infrastructure, integration of renewable energy sources, and smart-grid technologies will be key to ensuring the sustainability of EV fleets [3, 19, 28].

Recommendation 3: Raise consumer awareness and encouraging the uptake of green delivery.

Retailers and logistics companies should establish clear sustainability communication strategies, such as carbon tracking systems, green delivery labels, and digital consumer awareness promotion campaigns to inspire consumers to make sustainable choices [14, 15, 25, 30].

Recommendation 4: Promote cooperation between various stakeholders.

Sustainable logistics programmes should be a joint effort between governments, logistics operators, technology suppliers, energy suppliers and consumers. Collaborations between those parties are crucial to overcome infrastructure limitations and to implement green logistics solutions [8, 36].

Close cooperation of the government with NGOs, the public, and the private sector is essential for creating integrated sustainability frameworks.

Research and practice in the future should shift from technology-based solutions towards an integrated solution that takes into account both operational efficiency and environmental and consumer acceptance as well. These strategies more readily create possible scalable and sustainable solutions to urban logistics problems [1, 3, 22, 35, 42].

5.5 Final remark

One of the greatest challenges and opportunities in today's logistics landscape is moving toward sustainable last-mile deliveries. Overall, the findings of this review indicate that integrating AI into EV fleets, along with aligning with millennial sustainability values, offers a promising strategy to realize operational efficiency, environmental sustainability, and customer satisfaction. This, however, will depend on concerted effort from all industry stakeholders, policy makers, technology developers and the consumers. This trend towards the convergence of digital transformation and sustainability means that integrated AI-EV logistics ecosystems are poised to emerge as a decisive factor in the future of urban freight logistics.