Next-Gen Logistics: How Generative AI is Remapping the Supply Chain

The logistics industry is no stranger to the use of artificial intelligence (AI), having deployed it to positive effect for years to optimise routes, manage inventory and track the movement of goods. However, the advent of generative AI (Gen AI) has given these applications a massive shot in the arm in the form of vastly improved computing power to analyse data and perform complex tasks, improving existing use cases and creating new ones. 

The industry recognises the transformative power of this new technology, which is expected to cut costs by 15%, generate service level enhancements of up to 65% and provide a competitive edge in a challenging and crowded industry. Research shows nearly half (40%) of supply chain organisations are investing in Gen AI and the market for Gen AI in logistics is projected to expand from around US$800 million in 2024 to over US$13 billion by 2032.

According to Axelle Ratte, Director of Process, Methods and Industrialisation for FM Logistic, the industry has embraced Gen AI technology to optimise operations and enhance productivity while being watchful about its current limitations and related challenges.

“As an industry, we are both enthusiastic and cautious about Gen AI. While it’s early days and we’re yet to concretely measure the gains from the technology, the technology’s potential for optimisation, especially in administrative tasks, is widely recognised and it has been very well received by our employees because it saves time, improves productivity and facilitates decision-making,” Axelle Ratte says, noting the company has adopted Gen AI applications through the Google Workspace platform to improve productivity and facilitate decisions. “Yet, we need to carefully address challenges, such as ensuring data quality and integration, tackling cybersecurity risks and beware of AI’s tendency to hallucinate.”

While the use cases highlighted below are by no means exhaustive, they address core logistics challenges by enabling smarter and faster data-driven decision-making. They also provide measurable returns on investment (ROI) through cost reductions, improved service levels and operational efficiency gains – and, in doing so, are reshaping the industry.

Route and fleet optimisation. Traditional AI has long suggested optimal routes from A to B based on historical data. Generative AI takes this to the next level, analysing metrics ranging from delivery deadlines, fleet capacity, traffic and weather patterns to factors including geopolitical events and their impact on trade routes, fuel costs and transport infrastructure. It can even work with user-generated hypothetical scenarios to offer contingency route plans. Additionally, Gen AI algorithms maximise fleet usage by ensuring vehicles carry an optimal load at all times, including on return journeys to the warehouse.

Warehouse layout and operations design. Generative AI can optimise goods storage inside a warehouse so as to aid the fulfilment of orders containing different product types (say chips and alcohol). It can also be used to track the frequency of incoming orders, to enhance accessibility and to minimise congestion – all of which can reduce the time and effort needed to fulfil orders and even enhance worker ergonomics. 

Predictive maintenance and asset management. Gen AI applications can analyse usage patterns and equipment failure rates to plan maintenance protocols that optimise replacement times and ensure equipment is efficiently maintained. For instance, detailed analysis of telemetry data from a cargo truck fleet can provide actionable suggestions about the intervention needed for each vehicle. That is projected to reduce downtime by as much as 30% and cut maintenance and repair costs up to 10%.

Supply chain visibility, quality control and compliance. Gen AI-powered applications can significantly enhance logistics service providers’ product traceability capabilities by improving visibility across the supply chain through real-time data synthesis and predictive analytics. It can also elevate quality control and assurance by automatically analysing patterns and anomalies in shipment and handling processes, facilitating faster identification of defects and compliance issues. Gen AI further supports regulatory compliance, including complex requirements such as Scope 3 emissions reporting, by aggregating and verifying data from diverse supply chain activities, generating accurate emissions metrics and automating reporting processes.  

Demand forecasting and inventory management. Generative AI’s capacity to analyse prodigious amounts of information means it can crunch a range of data points, including historical sales patterns, economic indicators, weather forecasts and even social media sentiment. It can also assess synthetic data sets based on simulations to enhance predictive capabilities. This can help third-party logistic providers (3PL) plan and optimise resources and aid their customers efficiently manage inventory systems, which have tended to whiplash between “just in time” and “just in case” models, and  potentially lower operational costs by up to 20%.

Customer communications and documentation. From hyper-personalised, proactive updates on shipments to expediting customer complaints and claims, Gen AI can raise the game when it comes to customer experience. This boosts the bottom line as research shows nearly 80% of customers are more likely to choose a brand that provides personalised service. Gen AI applications also cut admin overheads and minimise errors by automating documentation like delivery notifications, contracts, invoices, shipping manifests and customs paperwork.

At FM Logistic, for instance, the company’s warehouse management systems use Generative AI to read thousands of classification documents to ensure dangerous goods are appropriately handled in line with regulatory requirements, according to Axelle Ratte. 

The advent of Agentic AI. Billed as the next frontier, Agentic AI takes Generative AI to the next level by giving AI tools the agency and autonomy to execute routine tasks – from managing warehouse inventory and optimising shelf space to monitoring shipments in real time. This can improve productivity and free up the human workforce for more strategic responsibilities. Estimates suggest using Agentic AI can improve operational efficiency by 30% and generate significant savings all while enhancing customer service and reducing carbon emissions.

FM Logistic, Axelle Ratte says, is testing Agentic AI to analyse and address IT system tickets as well as retrieve data from carrier portals to provide customers with proof of delivery (PoD), and check for confirmation of the PoD status based on the delivery date. “In this way, AI agents, which can work with each other, and understand and respond to complex requests, can fully automate a variety of administrative processes.”

While Gen AI’s potential to revolutionise the logistics industry is clear, the path to harnessing its power is anything but. Consequently, companies must be mindful of the challenges they will face in moving from proof-of-concept to deploying the technology at scale.

High upfront investment and ongoing maintenance costs. Adopting Generative AI requires substantial capital expenditure to fund the licensing of sophisticated models and pay for the computing power needed to run those models. Additionally, companies must invest in integrating siloed information into unified data sets, which are critical to successfully harnessing the technology. For instance, FM Logistic has migrated its entire IT system to the Google Cloud Platform, which now aids the company’s warehouse and transport management systems.

AI models also require continuous monitoring and retraining with new data to prevent so-called “model drift” and keep them from becoming obsolete, which can represent a significant operational expense. Companies will also need to hire AI and data science experts, a costly endeavour given the scarcity of such talent.

Workforce training and change management. A related challenge stemming from the lack of talent is the need to invest in upskilling the existing workforce – here, companies could face pushback from employees fearful of losing their jobs. Investing time and resources to sensitively manage change and build trust in the new technology is essential.

Legacy systems and infrastructure limitations. The logistics industry, like most early adopters of technology, runs on often decades-old technology, which is not typically compatible with, or even capable of, running new applications. Integrating them with AI platforms is not just inadvisable; it can prove impossible.

Data availability, quality and security. Data is the fuel that powers AI applications so companies must ensure there is enough of it. It is also crucial that the information AI applications process is of high quality to ensure the output is reliable. This includes implementing systems that can compile data from various, often siloed, sources and formats into unified datasets that can be fed into AI applications. Companies must also consider factors like data governance, security and the privacy of sensitive information, especially if they are using third-party, cloud-based AI applications. 

The green transition. Yet another challenge for logistics companies is to reconcile the use of a power-hungry resource like Gen AI with their sustainability goals. However, as Axelle Ratte notes, the energy-intensive technology also helps companies shrink their carbon footprint by optimising transport routes and admin tasks, and enhancing workforce productivity, all of which can help reduce overall emissions and generate savings, both in terms of cash and resources, which can then be devoted to aiding the green transition. 

As Axelle Ratte puts it: “If we use Gen AI as a tool in the service of sustainability, to improve our inventory management, our energy efficiency, to reduce emissions, reduce waste, and to responsibly manage resources, it can become an essential tool in aiding the transition.”

The key takeaway is clear: Deploying Gen AI in logistics holds enormous promise to help address the many challenges faced by the industry but this is an evolving space whose transformative potential must overcome the significant hurdles that implementation brings. To navigate this landscape successfully, companies must develop a clear and comprehensive next-gen AI strategy that is anchored by well-defined goals and measurable objectives to ensure they fully maximise their returns.