AI in last-mile deliveries: new paradigm of sustainable Xmas logistics

Artificial Intelligence has shown unprecedented potential over time. An AI system not only replicates human activities on a large scale and at high speed but, thanks to its predictive capabilities, acts as a versatile and adaptable decision-support tool that can pave the way to more informed choices and more effective strategies.

Even in logistics, though the practical applications of AI in the Italian landscape are still in their infancy, this technology has already begun to permeate every link in the supply chain: from warehouse operations to resource planning, from demand forecasting to inventory management and distribution.

AI has a broad range of applications across several operational domains. Its machine learning algorithms are proving highly beneficial in predicting fluctuations in demand, offering timely information for improved resource planning. Additionally, computer vision systems are enhancing warehouse operations’ efficiency, and optimising critical processes.

AI is also a key ally for planning and optimising journeys, as well as for sorting loads, and its potential is unveiled in pick times, such as the Christmas one. It is not just a question of improving routine operations, but also of reducing costs and waste: avoiding, for example, empty trips means reducing vehicle kilometres travelled, with obvious savings, especially in FTL, where the large distances travelled make even the smallest tricks extremely effective.

One of the most challenging logistical issues for the industry in the run-up to the Christmas holidays is the last mile, which is an intricate area to navigate.

It is characterised by the variability of destinations, high frequency of deliveries, and significant heterogeneity of load sizes and types, making it a costly and problematic task. This logistics industry has significant impacts on the economy, cities and customers, causing traffic congestion, noise and air pollution, as well as consumer dissatisfaction in the event of delays or delivery problems.

The environmental cost of the last mile is no less significant. The rise in home deliveries, often executed with conventionally powered vehicles and frequently riddled with inefficiencies, is a significant contributor to greenhouse gas emissions. Therefore, it is essential to render the last mile more sustainable, not just to decrease operational expenses but also to achieve sustainability and emissions reduction goals.

AI could play a pivotal role in turning the last mile into a more efficient and sustainable process.

The question that remains unanswered, despite much talk about it, is how this technology can really add significant value. In general, AI has the ability to cross a high number of variables, learning from historical data and exploiting mathematical models to find the ideal solution for conditions even never imagined before. The predictive abilities of an AI-based system combined with its massive computing power enable it to take into account all the requirements of orchestrating deliveries: from the filling of vehicles, to the selection of the right vehicle for each type of goods, to the working hours of drivers, and much more. Therefore, it is more reliable than any proven traditional method.

For instance, AI-based forecasting systems can improve the accuracy of delivery time estimates, thereby increasing customer satisfaction. AI can also play a key role in integrating more sustainable mobility solutions, such as electric vehicles or drones, with traditional ones. Finally, AI can contribute to greater transparency and traceability throughout the supply chain, making it easier to comply with environmental regulations.

Yet integrating this approach into the business is a major paradigm shift: it requires an openness on the part of the company to the process of digitisation, which can often face barriers that are not so much economic as cultural. Although it may seem obvious that capturing every activity or delivery in an integrated database makes this data usable in multiple ways and at multiple levels, this small change affects every step of the process. Impacting every aspect, from the organisation to the sorting of deliveries to the departure of the drivers themselves, it requires a major adaptation of tools, infrastructure and resources. Digitising business is a key challenge with the potential to enable those who have chosen this approach to remain competitive and to successfully meet the growing expectations of sustainability that both the regulatory system and the market are placing on the logistics sector.