The benefits of AI adoption in logistics are manifold. Applications range from real-time route optimization and predictive demand forecasting to robotic warehouse automation and intelligent customer service systems. These solutions offer measurable improvements in cost efficiency, service reliability, and speed of delivery. The current surge in AI adoption is not only being driven by technological improvement, but also by broader strategic recognition. According to a recent PwC survey, 56% of executives report more efficient use of working time 1. Introduction Artificial intelligence (AI) has emerged as one of the most transformative technologies of the 21st century. Its rapid development has brought AI to the forefront of industrial and service applications. Logistics, as the backbone of global supply chains, is no exception. In times of high volatility on the global markets and increasing demand for efficiency, resilience, and transparency, AI is now considered not an optional add-on, but an essential driver of competitiveness and innovation in logistics. due to generative AI, while 32% have experienced revenue growth and 34% have seen increased profitability (PwC, 2025). This paper aims to provide a structured overview of how AI technologies are being applied across core logistics processes along three established planning horizons: strategic, tactical, and operational. ƒ Strategic planning (years) encompasses long-term supply chain design, network structuring, and investment decisions. ƒ Tactical planning (months) addresses medium-term tasks such as aggregate Intelligent Logistics: Analyzing the Dissemination of AI Applications Artificial Intelligence, Supply Chain Management, Planning Horizons, Predictive Analytics Analyzing the dissemination of AI in logistics, this paper structures applications along strategic, tactical, and operational horizons. It concludes that while AI offers decision support for long-term strategy, its adoption is most mature in operational areas like route optimization and warehousing, delivering immediate efficiency gains. Eugen Truschkin, Emin Huseynov, Remmon Sarka DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 3 demand forecasting, capacity allocation, and inventory management. ƒ Operational planning (days to weeks) focuses on the day-to-day execution of logistics activities, including shift scheduling, vehicle dispatching, and load optimization. Differentiating the use of AI based on these planning horizons enables a clearer understanding of the current adoption patterns. Strategic applications generally offer substantial benefits, but require longer lead times. Operational AI solutions, on the other hand, quickly demonstrate value through immediate process improvements. This paper is based on desk research and the synthesis of evidence from academic research, industry reports, and real-world business cases. It leverages these findings to map out current AI use along the three planning horizons described above, identify where tangible benefits are already being realized, and explore future opportunities. This paper is structured as follows. Section 2 discusses AI applications identified at the strategic logistics level, and Section 3 explores AI applications at the tactical level. In Section 4, we provide an overview of AI applications at the operational logistics level. Section 5 discusses observed trends, and Section 6 concludes with our view on the prospects of AI applications in logistics. 2. AI applications at the strategic planning level Artificial intelligence is evolving into a major instrument supporting strategic decisions in logistics—choices that shape assets, partnerships, and capital over multi-year horizons. At the strategic level, supply chain network design and its three sub-tasks (transport corridor/ route configuration, prioritization of capital allocation, supplier and partner relationship management) are of specific practical interest. Across these topics, AI (which is often embedded in digital twins) does two main things: It puts messy internal and external data on a common footing, and it generates comparable scenarios that make trade-offs visible in terms of services, costs, risks, and emissions. The goal is not to replace engineering and feasibility work, but to identify and narrow down options, improve data quality, streamline routine activities, and generally support human experts in decision-making. Supply chain network design (SCND) and optimization Supply chain network design (SCND) involves establishing the footprint and roles of suppliers, plants, warehouses, and distribution centers to configure material and information flows. Traditionally, this has been treated as an occasional, static exercise. Today’s volatility in geopolitics, trade frictions, and shifting customer expectations, however, are imposing new requirements and demanding a more dynamic approach. AI is shifting SCND from episodic, spreadsheet-driven studies toward a continuous, evidence-based cycle that links strategic targets with day-to-day decisions. This shift can be described as an iterative “align-create-implement-validate” loop (Tengler Consulting, 2025): align business objectives and KPIs; create scenarios with AI optimization; implement short-term structural changes; and validate performance with continuous monitoring. The argument is that volatile demand, supply shocks, and service expectations now require SCND to operate as an ongoing capability rather than a one-time project. At the same time, tasks like capacity allocation through the placement of warehouses, terminals, cross-docks, and even new railway lines and roads still remain long-term strategic decisions. The described design and optimization cycle (“align-create-implement-validate”) can be efficiently executed inside a supply-chain digital twin that mirrors end-toend behavior and supports ongoing scenario testing and updates. In this context, shortterm actions (lane rerouting, inventory re-allocation) coexist with medium-/ longterm structural moves (opening/ closing facilities, dual sourcing), making network design a living, data-driven process rather than a one-time plan and providing a possibility for rapid course corrections as conditions change. In other words, digital twins are virtual replicas that ingest operational and financial data to simulate and optimize decisions from forecasting and sourcing to distribution and returns. Paired with predictive and prescriptive AI, they enable “self-healing” supply chains and various benefits, such as up to a 20% improvement in fulfilling delivery promises, a ~10% reduction in labor costs, and a ~5% rise in revenue in representative use cases (McKinsey, 2024a). In real-life scenarios, there are companies that specialize in network design and optimization and also actively integrate AI elements into their solutions to streamline processes. For instance, River Logic focuses on prescriptive, finance-aware decisioning through its Digital Planning Twin - a constraint-based digital representation of a value chain. At the same time, the platform’s Network Design solution addresses fundamental decisions (e.g. number and location of facilities) and scenario planning with a focus on formulating feasible plans under capacity, contractual, and regulatory constraints (River Logic, 2025). Another real-life example of a supply-chain design platform with a slightly different focus is anyLogistix, which combines analytical network optimization with dynamic simulation to evaluate end-to-end logistics networks, inventory, and production capacity. Its major functions include Greenfield Analysis (to identify possible facility locations based on demand and geography) and Network Optimization. While ALX is not an “AI-first” tool, it integrates machine-learning models so users can directly embed external ML - for example, demand forecasts - in simulations/ digital experiments (anyLogistix, 2025). The sections that follow examine three key SCND sub-tasks in greater detail: transport corridor/ route configuration, prioritization of capital allocation, and supplier and partner relationship management. Transport corridor/ route configuration As a part of broader SCND development, transport corridor/ route configuration and development is a long-term strategic process (from the point of view of an investor, producer or/ and infrastructure owner wishing to establish an intermodal logistics route or corridor, for instance). It often involves large design spaces (e.g. multi-national logistics corridors) and implies complex planning processes. Here, studies once again emphasize how AI-supported digital twins can offer benefits in terms of process streamlining, reduction of complexity, and improvement of efficiency by consolidating demand models and finance logic in order to evaluate long-term options (Wu et al., 2025; Nag et al., 2025). What AI adds at the strategic scale is the improvement of fundamental elements like data inputs (e.g. entity matching, data cleaning) and models (OD-demand and transport mode share forecasting). Guizzardi et al. (2025) show how digital twin frameworks for transport and logistics can combine simulation with machine learning to test corridor policies and facility allocations, as well as how resilience KPIs can be set in advance so that various options can be ranked under disruption scenarios. Recent studies also apply deep reinforcement learning to corridor capacity planning, indicating that AI can search wide policy/ design spaces, formulate phasing plans (with human governance), and convert simulated outcomes into KPIs and metrics for financial and operational decision-making (Farahani et al., 2025). In practice, a digital twin encodes various options for the infrastructure at hand (railway lines, terminal locations) and formulates several corridor development options; AI then tests each plan against numerous what-ifs - demand shifts, envi- LOGISTICS  Intelligent Logistics DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 4 ternative-sourcing analysis (AlMahri et al., 2024). Complementary research focuses on trustworthy link prediction in supplier KGs (e.g. by demonstrating possible company relationships and dependencies) to improve transparency, which is important for governance and escalation decisions in SRM (Kosasih & Brintrup, 2024/ 2025). Together, these techniques turn scattered, incomplete inputs into a continuously updated “network view” that enables selection and risk monitoring. In real life, numerous companies provide supplier-data platforms with AI integration. For instance, TealBook uses AI to unify and enrich fragmented supplier records and refresh profiles across systems. Knowledge-graph approaches are also already deployed at scale: Companies like Interos and Sourcemap use KGs to map millions of companies and complex multi-tier networks. AI’s second contribution is to support the design of the commercial architecture that underpins long-term partnerships: term structures, risk-sharing and price indexation, service targets, penalties/ bonuses, and governance of partnerships across multi-year frameworks. Models trained on contract portfolios and market data help benchmark clauses, simulate total cost under various price adjustments or service terms, and highlight the most important points to negotiate before going to market; the results are aligned with enterprise KPIs and rolled up to financial and service metrics for decision-making (GEP, 2024). CLM vendors such as Icertis and Agiloft now ship generative-AI copilots that specialize in drafting/ amending contract clauses. 3. AI applications at the tactical planning level Artificial intelligence is quietly transforming the tactical pillar of logistics, altering how service providers adopt a mid-term strategy and make choices. AI is also increasingly gaining traction in the critical middle ground, guiding demand forecasting, inventory management, supplier evaluation, workforce and shift planning, and asset management. A more detailed overview of these AI-driven tasks is provided below. Demand forecasting and capacity planning Demand forecasting is a staple of logistics, but with the emergence of artificial intelligence, it is changing its roots. No longer dependent on stagnant historical data and seasonality, forecasting is now informed by sophisticated algorithms, live analytics, and AI-based services in order to create much more precise and effective forecasts. Using AI to monitor customer trends, price movement, delivery schedules, and market ronmental instabilities, new policies, and other external factors. It then highlights the options that still perform well under those shocks. This approach is also relevant for separate logistics chain/ corridor elements like transport infrastructure and logistics facilities in general. For instance, a real-life business case in the Port of Corpus Christi (TX, USA) followed a similar logic, using a digital twin of the port to test various development options before actually committing to capex (Business Insider, 2025). Prioritization of capital allocation The same engine that steers daily flows and short-term decisions can also rank long-term investments and plan the timing of their rollout, strengthening the link between short-term performance and strategic capital allocation. In effect, AI-enabled digital twins can serve as a central decision layer for ranking and phasing logistics-infrastructure investments. A McKinsey study of public mega-projects shows that twins that bring together engineering, demand, and cost data can boost capitaland operating efficiency by 20-30% (McKinsey, 2025). A recent review of digital twin technology for infrastructure finds that “digital twin plus AI” toolchains cut facility-allocation modelling time by ~40% and raise net-present value through rapid scenario iteration (Moshood et al., 2024). Supplier and partner relationship management (SRM) SRM addresses how firms discover, evaluate, contract with, and develop suppliers to meet cost, quality, service, resilience, and sustainability objectives. Historically, SRM has been less structured and more fragmented; recent literature positions it as an ongoing capability that combines internal master data with external signals (news, disclosures, operational events). AI supports the SRM process by turning unstructured or inconsistent inputs into analyzable supplier profiles and aligning selection and development with enterprise KPIs (Mohammed, 2023). At a strategic scale, the first contribution of AI is to industrialize SRM data: entity matching across systems, automated entity classification (e.g. category, region, risk flags), and extraction of supplier facts from text. Recent work proposes knowledge-graph (KG) pipelines (a map of companies and how they are linked) that use large language models (LLMs) to research public sources, fill in the KG, and infer relationships beyond tier-1/ -2 suppliers (showing the supplier and the supplier’s supplier), thereby expanding visibility. This KG+LLM approach supports discovery, mapping of critical dependencies, and alindicators in real-time is helping logistics companies predict demand more accurately. This change involves much more than enhanced precision; more intelligent predictions result in more intelligent inventory control, more efficient transportation scheduling, and improved warehouse control. The eventual outcomes include reduced costs, less waste, and supply chains that keep up with (and even anticipate) the latest developments. AI-driven demand forecasting is now a strategic requirement in this new landscape. Organizations such as Blue Yonder are applying machine learning to forecast demand and optimize entire supply chains. Relied upon by the largest companies in the world (like Walgreens and Bayer), their planning solutions use real-time data to predict demand fluctuations, reduce waste levels, and drive increased turnover. They enable organizations to remain agile, efficient, and responsive to customer demands through their sustained capability to react to variability in the market (Blue Yonder, 2025). In addition, Amazon has recently introduced an AI-powered demand forecasting model to support the customer experience. This new foundational model is designed to predict what customers will want, as well as where and when, for hundreds of millions of products per day (Amazon, 2025). While the previous systems Amazon used relied on sales histories to guide inventory planning decisions, its AI-powered forecasting model adds time-bound data like weather patterns and holiday schedules to place the right products in the right locations more accurately (Amazon, 2025). Amazon has reported that these forecasts have contributed to a 10% improvement in long-term national forecasts for deal events and a 20% improvement in the regional forecasts for millions of popular items (Amazon, 2025). The forecasting method has already been implemented in the U.S., Canada, Mexico, and Brazil, and the benefits can already be seen. Packages arrive faster than they previously did and delivery partners travel fewer miles, which in turn reduces traffic and carbon emissions (Amazon, 2025). AI-powered data forecasting tools used by Amazon and Blue Yonder are among the many solutions that are currently reshaping the logistics landscape. They help companies become more efficient, reduce waste, and improve their adaptability. Inventory management and stocking policies As AI handles more and more aspects of demand forecasting, the advantages automatically carry over to inventory management and stocking policy. By analyzing large DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 5 Intelligent Logistics  LOGISTICS amounts of data in real time, AI applications calculate the most effective reorder points and quantities in order to replenish inventory in the most efficient manner, thereby minimizing carrying costs and releasing capital that is held up in excess stock. This degree of precision allows companies to maintain their service levels as they minimize waste and make their supply chains more agile. The adoption of AI-based inventory optimization by small and medium-sized businesses across different markets is demonstrated by platforms like Netstock. These systems use smart product classification to rank products based on sales velocity and performance and customize replenishment cycles, fill rates, and safety stock levels. They also help in identifying obsolete or stock-out items, enabling companies to adopt a leaner stocking policy that will respond directly to demand signals. Inventory policies can be fixed on a national level or can vary by product, supplier, or location, with manual overrides that facilitate faster decision-making under turbulent conditions (Netstock, 2025). Similarly, ToolsGroup uses AI and probabilistic forecasting to control inventory with complex, multi-level supply chains. Its methodology combines sets of demand scenarios and probabilities to help deal with uncertainty and volatility, which traditional techniques fail to handle. Stock levels can be aligned with real demands by using multi-echelon optimization and service-oriented planning capabilities. This adaptability is taken to an even greater extent with machine learning, which uses internal and external signals to enhance predictions and inventory targets in a continuous process (ToolsGroup, 2025). Collectively, these technologies represent a move toward more adaptive, data-driven inventory management that balances financial, operational, and customer service goals with constant, smart adaptation. Procurement and supplier evaluation Artificial intelligence is also transforming the procurement process in logistics by facilitating smarter sourcing decisions, the ongoing evaluation of suppliers, and more responsive contract management. Current AI technology makes it possible to keep supplier performance on track in real time by converting performance factors such as delivery reliability, quality, and contract compliance into numerical scores. Machine learning can be used to detect risks in a more efficient way by indicating potential warning signs such as financial strain, employee conflicts, or ESG infringement based on sources in news, social media, and financial platforms. Sustainability is also considered, with AI assessing suppliers on the basis of emissions, ethical sourcing, and rule adherence. In contract management, AI can scan large sets of contract portfolios to find pricing anomalies, opportunities to exploit contract loopholes, and vulnerabilities to legal issues. Automated negotiation systems are also able to process routine negotiations, simplifying the procedures involved and freeing up human resources for more strategic activities. The use of artificial intelligence in end-to-end procurement and supplier management is evident on AI-based platforms like JAGGAER and Scoutbee. JAGGAER, which is used at companies such as Lenovo and Rolls-Royce, uses machine learning for autonomous sourcing, supplier risk and performance analysis, and predictive demand and spend forecasting (JAG- GAER, 2025). In a similar fashion, Scoutbee uses AI to improve supplier discovery and analytics on a platform that includes capabilities, ESG compliance, intelligent supplier scouting, automated supplier matching, and data enrichment. Siemens, Unilever, and Audi are some of the corporations that utilize these tools to digitalize procurement and get a better view of suppliers in the international market (Scoutbee, 2025). Asset management Asset management can be seen as one of the AI applications that has been adopted more broadly in logistics. One of the fields in this domain that can be associated with the tactical level is predictive maintenance. AI-driven predictive maintenance platforms analyze real-time sensor data such as engine temperature, vibration, and pressure to anticipate vehicle failures and schedule maintenance proactively. This reduces unplanned downtime, lowers maintenance costs, and extends asset lifespans. Here, the case of Hitachi can be cited as an example of AI-driven fleet management solutions that enable fleet operators to streamline their operations, maintenance, and repair processes, leading to reduced vehicle downtime and better management of both fleets and risks (Hitachi, 2025). 4. AI applications at the operational planning level Nowadays, established supply chain structures compete in a highly volatile environment with increased customer expectations with regard to service level. AI tools support companies in coping with this increased complexity by optimizing ongoing operations, saving costs, improving their service level, and contributing to their sustainability agenda. Our research reveals a wide dissemination of AI applications, specifically at the operational level of logistics (fulfillment). As an example, respondents in a recent survey of 130 companies (mostly from Germany) consider optimized routing, improved resource utilization, reduction of transport costs, and an improved CO2 footprint among the major gains afforded by AI (Inform, 2024). The following section provides an overview of AI applications at the operational logistics level. Warehousing (picking and sorting) AI-powered tools can unlock 7-15% more capacity in warehouse networks by identifying daily spare capacity, understanding variability in resource availability, and evaluating opportunities to improve efficiency (McKinsey, 2024). The examples below demonstrate how current AI applications are used in warehouse operations. Amazon has invested in new robotic inventory management systems like Sequoia. Sequoia makes it possible to identify and store inventory at Amazon fulfillment centers up to 75% faster, which results in items being listed for sale on Amazon.com more quickly - a benefit for both sellers and customers. It also reduces the time it takes to process an order through a fulfillment center by up to 25%, which improves shipping predictability and increases the number of goods Amazon can offer for same-day or next-day shipping (Amazon, 2023). Digit (see Figure 1), a mobile robot from Agility Robotics, is being tested by Amazon to help employees with tote recycling - a highly repetitive process of picking up and moving empty totes once inventory has been completely picked out of them (Amazon, 2023). Meanwhile, a new agentic AI team within Amazon Robotics is working on the next step in the robotics evolution, adding the ability for robots to hear and understand natural language, process it in a rational manner, and act autonomously (Amazon, 2025). At DHL, AI-powered sorting robots dubbed DHLBots are increasing sorting capacity by 40% or more. These robots are capable of sorting over 1,000 small parcels per hour with 99% accuracy (DHL, 2025a). DHL expects the industry to make use of warehouse management assistants in optimizing inventory placement, automating picking and packing processes, monitoring equipment maintenance schedules, and enhancing overall operational efficiency within warehouses and distribution centers (Dohrmann et al., 2024). Vision picking can be considered one of the further applications of AI use cases in warehousing. With TeamViewer Frontline Pick technology, DHL warehouse employees can utilize smart glasses equipped with vision picking software that provides visual LOGISTICS  Intelligent Logistics DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 6 pacity and improve its network capacity management, resulting in a 7-9% total cost reduction (Transmetrics, 2022). ƒ Vehicle inspection and condition monitoring: Focalx uses computer vision AI to automate vehicle inspections and damage detection. Its platform reduces manual inspection time by up to 80% while minimizing disputes and accelerating fleet readiness and turnover (Focalx, 2025). Customs clearance Customs clearance is often cited as a slow, complicated, bureaucratic affair due to complex regulations and the volume of data to be processed. AI-powered customs clearance solutions support customers in streamlining the process, resulting in time savings, increased accuracy, and less human error. KlearNow and eClear are examples of digital AI-powered customs brokers that offer customers the benefits of process automation. KlearNow, a U.S.-based company founded in 2018, and its KlearCustoms platform automatically converts and organizes all customs documents from customers’ email chains, ensuring the accuracy of customs declarations along with their compliance with the local regulations. In addition, real-time information on shipments is provided. An integrated ROI calculator for importers reports cost savings of 28% when using the default settings of KlearNow.ai (Klear- Now.ai, 2025). eClear, a Germany-based company founded in 2016, offers CustomsAI, which focuses on customs clearance in the EU market. It automates tasks like document classification, tariff code determination, and risk assessment (where it identifies shipments that may warrant closer inspection). Through AI-assisted analysis of customers’ product information (e.g. EAN, GTIN, ASIN, descripareas such as warehouses, delivery centers, and transportation terminals. To estimate the required number of workers, AI analyses previous workload information, seasonal patterns, real-time order volumes, and forecast models. This helps reduce costs, as it can identify where to anticipate overtime and offer alternatives, such as changing schedules to available personnel. AI also increases employee satisfaction by taking into account each individual’s preferences and free time, which enhances work-life balance while reducing absenteeism and churn. In addition, AI makes sure that labor laws and company regulations are followed, thereby ensuring high productivity and compliance (ROCKCREST, 2024). AI-based tools such as UKG Pro Workforce Management streamline scheduling, attendance tracking, and compliance monitoring. Its machine learning component forecasts the number of workers required and creates the most optimal schedules based on previous performance and emerging trends (UKG, 2021a). This reduces overstaffing and overtime costs and aligns staffing with current requirements. A built-in rules engine monitors break time, overtime restrictions, and local labor laws, alerting managers of potential infractions and enforcing compliance with rules (UKG, 2021b). Logistics staffing decisions can be improved even further with the rise of AI and corresponding tools like this (UKG, 2021b). Asset management The following application fields can be identified in the asset management domain at the operational level: ƒ Truck utilization and capacity assessment: AI improves smart truck utilization by improving cargo loading efficiency and reducing underutilized capacity. Using Transmetrics’ AI, DPD was able to obtain better data about its loading cacues throughout the picking process. For real-time data access, the Frontline solution integrates seamlessly with DHL’s warehouse management system, establishing a process that is fully connected end-to-end. This has resulted in a 15% productivity increase, a reduction of error rates to 0.1%, and time savings of 50-70% during onboarding (TeamViewer, 2025). Overall, AI adoption in postal and parcel systems is increasing significantly, particularly in connection with sorting, routing, and customer operations (Dobrodolac et al., 2025). Some of the application cases for AI-powered routing are demonstrated below. Route optimization Route optimization systems driven by AI assist businesses in removing congestion and bottlenecks by optimizing transportation routes, schedules, traffic flows, and capacity distribution while reducing travel time according to real-time demand fluctuations. This entails traffic flow forecasting, route adaptation for weather-related delays, lastmile delivery utilizing geospatial intelligence, and multi-stop optimization to identify the optimal delivery sequence based on vehicle capacities, travel distances, and time windows. Consequently, these systems enhance delivery speed, optimize fuel usage, and lower harmful emissions. Route optimization can be identified as one of the most widely applied AI use cases in logistics. Various companies actively involved in the organization of transport benefit from AI in transportation routing, where one can differentiate between inhouse solutions (e.g. Wellspring from Amazon, ORION from UPS) and vendor solutions (e.g. DispatchTrack or Wise Systems, the latter of which is in use at DHL). The quantifiable benefits of AI-driven route optimization are significant. The introduction of ORION resulted in $320 million in savings and reduced UPS’s fuel consumption by 10 million gallons (RoboticsTomorrow, 2025). In 2024, Maersk achieved a reduction in fuel consumption of up to 15% while also reducing shipping times for critical routes by approximately 20% (Codex, 2025). This was made possible by Maersk’s AI-driven predictive analytics solution, which considers weather patterns, ocean currents, and other influencing factors such as port congestion to determine the optimal route for each of its shipments (Reddy, 2023; Raj, 2024). Workforce and shift planning Artificial intelligence is becoming crucial in assisting logistics firms with planning shifts and staffing. It aligns employees with evolving demands, particularly in high-traffic Figure 1: Digit robot in its testing phase at Amazon (source: Amazon, 2023) Intelligent Logistics  LOGISTICS DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 7 Figure 2: myDHLi - an AI-powered virtual assistant from DHL Global Forwarding (DHL, 2025b) human-steered pipeline with the following logic: objective setting and constraints (human), large-scale scenario generation and financial/ service/ resilience scoring (AI), decision governance (human), and the incorporation of post-implementation learning back into the digital twin (human + AI). This allows for long-term logistics chain commitments while making the respective network adaptable in the short and medium term. The processes of demand forecasting, inventory management, supplier evaluation, and asset management are overseen at the tactical level and are being redefined by artificial intelligence at the medium-term logistics planning level. AI provides support in making more strategic procurement decisions, optimizing inventory policy, forecasting demand, and continuously evaluating suppliers and contracts with real-time analytics and predictive modelling. All these choices feed into each other, making the corresponding system anticipatory rather than reactive. These insights result in logistical functions that are lean, adaptive, and strategically aligned with the long-term vision of the organization. As these moving pieces become tighter and more synchronized at the tactical level, it reduces waste while streamlining costs and responsiveness. AI therefore enhances the tactical level by establishing an essential connection between strategic intent and operational implementation. The operational logistics level is characterized by a wide dissemination of AI-driven applications, with use cases in warehousing (picking and sorting), route optimization, asset management (truck load optimization, 5. Discussion Based on desk research, this paper attempts to reveal the current dissemination level of AI applications in logistics. Figure 3 summarizes the outcomes of our study, presenting the respective AI applications in association with both the logistics planning levels and various elements of the supply chain (procurement, manufacturing, and distribution). The creation of physical infrastructure (e.g. terminals, cross-docks) and the design of complex multi-national logistics routes/ corridors remain strategic, longterm, largely irreversible investments that must be evaluated via feasibility studies and multi-criteria analysis. The related literature underscores that facility allocation choices are costly to reverse and span long time horizons (Snyder et al., 2005), which is why they are considered at the strategic level even as AI improves the surrounding analytics. Despite the increasing maturity of AI tools and AI integration into existing solutions for strategy-level tasks in general, the current consensus is that today’s systems still mainly offer decision support and cannot serve as fully autonomous greenfield designers. AI already excels at components of the entire pipeline (demand forecasting, risk analysis, routing, financial optimization inside digital twins), but fully autonomous, end-to-end steering of all the steps within high-complexity tasks (from greenfield site screening to feasibility analysis, engineering design, financing, and staged rollout) without active human governance is not yet standard practice. One pragmatic short-term target involves an AI-assisted, tion texts, or brand names), CustomsAI instantly identifies the matching TARIC code (the integrated tariff code of the European Union). Existing customs classifications are also checked and corrected if necessary, and product-specific VAT rates are assigned automatically (eClear, 2025). Customer service AI-powered chatbots are used in the transport and logistics industry to better understand customer intent, answer queries in an efficient manner, and provide shipping updates. While there has still not been a broad dissemination of a standalone chatbot product in customer service among the industry giants, some examples from the world’s transport and logistics leaders are presented below. MyDHLi (see Figure 2), an AI-powered virtual assistant used by DHL Global Forwarding, was launched in 2020. Using this 24/ 7 chatbot, customers receive real-time access to quotes, transport modes, carbon emissions, and other key shipment data. The tool is accessible from any device and allows customers to analyze their logistics activities using a range of filters and customizable views. A set of further features, like daily proactive updates listing key milestones on demand, is also available (DHL, 2025b). Meanwhile, enterprises seem to favor vendor partnerships and tend to publicize outcomes as vendor case studies rather than in-house deployments. As an example, Live- Person technology and LivePerson Conversational Cloud was used to develop and integrate a chatbot into GLS Slovenia’s webpage, SMS channel, and Viber and WhatsApp messaging platforms (2Mobile, 2025). LOGISTICS  Intelligent Logistics DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 8 codex.team/ blog/ ai-and-supply-chain-optimization-leading-to-a-25-reduction-in-delivery-times D.E.M.M.C.S, 2025: Risk Monitoring in AI-Enabled Supply Chains: Why Human Oversight Remains Critical. https: / / www.demmcs.com/ archives/ risk-monitoring in ai enable d supplychain s -whyhu man-oversight-remains-critical DHL, 2025a: AI in logistics: lending a helping … arm. https: / / www.dhl.com/ global-en/ delivered/ innovation/ ai-in-logistics.html DHL, 2025b: A New Generative AI Assistant to Boost Your Logistics Efficiency. https: / / www.dhl.com/ usen/ home/ global-forwarding/ latest-news-and-webinars/ mydhli-generative-ai-logistics.html Dohrmann, K. et al., 2024: Logistics Trend Radar 7.0. DHL Group. https: / / www.dhl.com/ de-en/ home/ innovation-in-logistics/ logistics-trend-radar.html Dobrodolac, M. et al., 2025: Exploring the Potential Applications of Artificial Intelligence in Parcel. Management Science Advances. Volume 2, Issue 1 (2025) 107-11. eClear, 2025: CustomsAI. https: / / eclear.com/ product/ customsai/ Farahani et al., 2025: Capacity planning in logistics corridors: Deep reinforcement learning for the dynamic stochastic temporal bin packing problem. https: / / www.sciencedirect.com/ science/ article/ pii/ S1366554524003338 Focalx, 2025: Fleet Inspections: Monitor your fleet condition with AI. https: / / focalx.ai/ fleet-management/ GEP, 2024: How AI-Powered Supplier Relationship Management Can Unlock the Hidden Value of Procurement. https: / / www.gep.com/ blog/ strategy/ ai-powered-supplier-relationship-management-for-procurement Guizzardi et al., 2025: Requirements Analysis for a Digital Twin to Increase the Resilience of Multimodal Corridors: A Case Study in the Twente Region. https: / / link.springer.com/ chapter/ 10.1007/ 978-3-031-94931-9_18 Hitachi, 2025: Fleet management is easy with Hitachi. https: / / social-innovation.hitachi/ en-us/ thinkahead/ transportation/ fleet-management/ Inform, 2024: Trendreport (Generative) KI in der Logistik - Hype oder Gamechanger? Accessed 25.06.2025. Available at: https: / / www.inform-software.com/ de/ landingpages/ trendreport-generative-ki-in-der-logistik#Get-results different logistics and supply chain operations. Incomplete, biased, or quickly evolving data can lead AI models to a conclusion that seems correct on the surface but is, in fact, flawed. In a field as complicated and expensive as this, such mistakes can rapidly turn into a financial loss, a tarnished reputation, or a regulatory risk (D.E.M.M.C.S, 2025). Ultimately, the way forward should be a balanced combination of innovation and responsibility. AI should become a tool that not only drives efficiency and adaptivity in logistics, but also serves as a trusted foundation for sustainable and resilient logistics processes. ▪ SOURCES AlMahri et al., 2024: Enhancing Supply Chain Visibility with Knowledge Graphs and Large Language Models. https: / / arxiv.org/ abs/ 2408.07705 Amazon, 2025: Amazon announces 3 AI-powered innovations to get packages to customers faster. https: / / www.aboutamazon.com/ news/ operations/ amazon-ai-innovations-delivery-forecasting-robotics Amazon, 2023: Amazon announces 2 new ways it’s using robots to assist employees and deliver for customers. https: / / www.aboutamazon.com/ news/ operations/ amazon-introduces-new-robotics-solutions anyLogistix, 2025: Supply Chain Network Optimization. https: / / www.anylogistix.com/ features/ supply-chain-network-optimization/ Blue Yonder, 2025: AI & Machine learning. https: / / blueyonder.com/ why-blue-yonder/ ai-and-machine-learning#trust Business Insider, 2025 : A massive seaport in Texas is using an AI-powered digital replica to track ships and prepare for emergencies. https: / / www. businessinsider.com/ corpus-christi-port-ai-shiptracking-emergency-training-2025-5 Codex, 2025: AI and Supply Chain Optimization Leading to a 25% Reduction in Delivery Times. https: / / vehicle inspection), workforce planning, customs clearance, and customer service. Overall, it was identified that both vendor-owned solutions and in-house AI deployments are present among the companies. While comparing the dissemination of AI at all the planning levels in logistics, operation-level use cases in particular (optimized routing, improved resource utilization) appear to play a “mainstream” role in the current logistics AI landscape. These in turn are driven by logistics giants like Amazon, DHL, or Maersk, which see optimizing the productivity of their existing logistics infrastructure as a target function. Along with cost reductions, a significant improvement in service level is being achieved here, as the examples covered demonstrate. 6. Conclusion AI is becoming a structural driver that is transforming the way supply chains are designed, run, and implemented at all levels. At the strategic level, it increases the adaptability of network design, investment prioritization, and the management of supplier relationships, which results in more shock-resilient supply chains. At the tactical level, it transforms demand forecasting, inventory management, procurement, and asset management into anticipatory, data-driven functions. At the operational level, AI has already started becoming the norm, with measurable efficiency improvements in warehousing, routing, workforce planning, and customer engagement. Combined, the above-mentioned developments indicate that AI in logistics is no longer about isolated applications, but connected ecosystems that bridge the gap between long-term vision and daily execution. The purpose of the technology is still predominantly decision support (with human governance still necessary), but the trend is toward more autonomous and self-optimizing supply chains. The main challenge facing logistics companies is not whether to implement AI, but how to integrate it in a manner that strikes a balance between efficiency with resilience and innovation with responsibility. The application of AI in logistics offers many benefits, and yet the implementation process should be approached carefully because it also presents some challenges. The two main concerns are data security and privacy, and the expenses involved in installing and maintaining the required infrastructure. To ensure the safety of sensitive customer data, companies need to invest in the secure storage of such information and comply with the regulations concerning data maintenance (the ILS Company, 2025). Moreover, human supervision is very much needed when applying AI in Operational planning level Strategic planning level Tactical planning level Distribution Procurement Manufacturing Supply Chain Network Design and Optimization Demand forecasting and capacity planning Customer Service Supplier Evaluation Workforce & shift planning Warehouse picking & sorting Route optimization Predictive maintenance Customs clearance Customs clearance Inventory Management Truck utilization & capacity assessment Vehicle inspection and condition monitoring Figure 3: Overview of AI applications at the different logistics planning levels (authors’ own research) Intelligent Logistics  LOGISTICS DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 9 UKG, 2021a: Flexible scheduling solutions to support your people and their unique life-work experiences. https: / / www.ukg.com/ sites/ default/ files/ legacy/ kronos/ resources/ pdf/ en/ UKG -Dimensions-Scheduling-Profile.pdf UKG, 2021b: How technology helps restaurants keep up with evolving labor laws and regulations. https: / / www.ukg.com/ sites/ default/ files/ legacy/ kronos/ resources/ pdf/ en/ rt0256-USv3-restaurant-compliance-ebook.pdf Wu et al., 2025: Digital Twin Technology in Transportation Infrastructure: A Comprehensive Survey of Current Applications, Challenges, and Future Directions. https: / / www.mdpi.com/ 2076- 3417/ 15/ 4/ 1911 2Mobile, 2025: GLS Slovenia - Improving Customer Service. https: / / www.2mobile.si/ en-gb/ case-studies#MGLS Cover image: © iStock.com/ imaginima Raj, Anita, 2024: AI in Transportation and Logistics: Steering Towards Smarter Operations. https: / / throughput.world/ blog/ ai-in-transportation-and-logistics/ Reddy, Rahul Kumar, 2023: Case Studies Of How AI Is Being Used To Improve Sustainable Supply Chains. https: / / ruralhandmade.com/ blog/ case-studiesof-how-ai-is-being-used-to-improve-sustainable River Logic, 2025: Network Design Optimization for Supply Chain. https: / / riverlogic.com/ solutions/ network-design-optimization-for-supply-chain/ River Logic RoboticsTomorrow, 2025: AI Route Optimization Saves Money, Cuts Fuel Consumption and Provides Faster Delivery Times. https: / / www.roboticstomorrow.com/ stor y/ 2025/ 01/ ai-route optimi z ati o n s ave s m o n eycu t s f u e l co n s um p tion-and-provides-faster-delivery-times/ 23800/ ROCKCREST, 2024: The role of AI and automation in workforce management for smarter scheduling. https: / / www.rockcrest.com/ post/ the-role-of-aiand-automation-in-workforce-management-forsmarter-scheduling Scoutbee, 2025: Scoutbee. https: / / www.scoutbee.com/ Snyder et al., 2005: Facility Location under Uncertainty: A Review (long-lived, hard-to-reverse strategic decisions). https: / / coral.ise.lehigh.edu/ larry/ files/ pubs/ stochloc.pdf TeamViewer, 2025: DHL supply chain uses Frontline Pick to help increase productivity and reduce error rates. TeamViewer Customer Success Story: DHL Tengler Consulting, 2025: From one-time planning to continuous optimization: The AI revolution in supply chain network design. https: / / tenglerconsulting.com/ en/ ai-driven-supply-chain-networkdesign/ The ILS Company, 2025: Unlocking the Potential of AI in Logistics: Benefits, Challenges & Best Practices. https: / / www.ilscompany.com/ ai-in-logistics/ ToolsGroup, 2025: Responsive Inventory Management and Rebalancing. https: / / www.toolsgroup.com/ solutions/ rebalance-ai/ Transmetrics, 2022: Improving Logistics Network Capacity Management with AI & Demand Forecasting. https: / / www.transmetrics.ai/ blog/ logistics-network-capacity-management/ Jaggaer, 2025: Complexity simplified. https: / / www. jaggaer.com/ complexity-simplified KlearNow.ai, 2025: Importer ROI calculator. https: / / www.klearnow.ai/ roi-calculator-usa/ Kosasih, E. and Brintrup, A., 2024/ 2025: Towards knowledge graph reasoning for supply chain risk management using graph neural networks. h t t p s : / / i d e a s . r e p e c . o r g / a / t a f / t p r s x x / v 6 2 y - 2024i15p5596-5612.html McKinsey, 2025: Digital twins: Boosting ROI of government infrastructure investments. https: / / www.mckinsey.com/ industries/ public-sector/ our-insights/ digital-twins-boosting-roi-of-government-infrastructure-investments McKinsey, 2024a: Digital twins: The key to unlocking end-to-end supply chain growth. https: / / www. mckinsey.com/ capabilities/ quantumblack/ our-insights/ digital-twins-the-key-to-unlocking-end-toend-supply-chain-growth McKinsey, 2024b: Harnessing the power of AI in distribution operations. https: / / www.mckinsey.com/ industries/ industrials-and-electronics/ our-insights/ distribution-blog/ harnessing-the-power-of-ai-in-distribution-operations Moshood et al., 2024: Infrastructure digital twin technology: A new paradigm for future construction industry. https: / / www.sciencedirect.com/ science/ article/ pii/ S0160791X24000678 Mohammed, I. A., 2023: Artificial Intelligence in Supplier Selection and Performance Monitoring: A Framework for Supply Chain Managers. https: / / kuey.net/ index.php/ kuey/ article/ view/ 8650/ 6502 Nag et al., 2025: Exploring digital twins for transport planning: a review. https: / / etrr.springeropen.com/ articles/ 10.1186/ s12544-025-00713-0 Netstock, 2025: Netstock Announces the Release of its AI-Powered Predictive Planning Suite. https: / / www.netstock.com/ blog/ netstock-announcesthe-release-of-its-ai-powered-predictive-planning-suite/ PwC, 2025: PwC’s 28th CEO Survey. Accessed 21.08.2025. Available at: https: / / www.pwc.nl/ nl/ actueel-publicaties/ assets/ pdfs/ 28th-ceo-survey.pdf Eugen Truschkin, Dr., Director Rail and Intermodal Logistics Consulting (I.TAC 211), DB Engineering & Consulting GmbH (Deutsche Bahn AG) Emin Huseynov, Expert Rail and Intermodal Logistics Consulting, DB Engineering & Consulting GmbH (Deutsche Bahn AG) Remmon Sarka, Logistics Consultant, Rail and Intermodal Logistics Consulting, DB Engineering & Consulting GmbH (Deutsche Bahn AG) LOGISTICS  Intelligent Logistics DOI: 10.24053/ IV-2025-0073 International Transportation (77) Collection ǀ 2025 10