Translated with DeepL
At the start of a logistics process, AI analyses large volumes of data, allocates goods, identifies deviations and thus creates a reliable basis for the next steps. This early structuring then influences the entire system, from warehouse management to transport planning.
AI systems are increasingly replacing manual checks in logistics centres. They analyse images and text information, count parcels, detect damage and automatically assign goods to orders. Autonomous warehouse robots also optimise transport routes and shelf space. This makes internal processes more precise, faster and less error-prone.
Early data analysis also has an impact on the transport chain. Modern optimisation processes select suitable means of transport, plan routes and distribute orders in such a way that adherence to delivery dates, capacity utilisation and sustainability are taken into account. These systems react immediately to changes and generate solutions that are significantly more accurate and faster than manual planning.
The use of AI is also changing the tasks of logistics experts. Once AI systems have allocated goods, evaluated data and optimised processes, specialists take over the next steps. This includes checking the AI proposals, setting priorities, adjusting parameters and deciding on implementation during ongoing operations. They therefore control those process phases that build on the automated analyses and are central to a smooth overall process.
Read the article to find out how logistics is developing with AI and which steps make sense for SMEs.
This article in KMU-Magazin is part of a series that has emerged as a follow-up activity to the publication Orientation AI: Challenges and Opportunities for Swiss SMEs. This SATW publication supports small and medium-sized enterprises in recognising the potential of artificial intelligence and planning concrete next steps.
The content was created in close collaboration with stakeholders from the SAIROP (Swiss AI Research Overview Platform) network. SAIROP promotes the exchange between science, business and society, makes Swiss AI expertise visible and provides orientation in the dynamic AI ecosystem.
Intra-logistics encompasses all processes within a company, such as incoming goods, storage, order picking and outgoing goods. Intra-logistics is increasingly benefiting from AI applications.
Extra logistics refers to the transport of goods between companies and end customers. Extra logistics is increasingly benefiting from AI applications.
AI analyses images and texts, counts parcels, detects damage and automatically assigns goods to orders. Autonomous robots optimise shelf space and transport routes. Companies such as Amazon are already making widespread use of such systems.
The focus is on selecting suitable means of transport, distributing orders across the fleet and optimising route planning. Modern algorithms can load pallets efficiently, predict transport volumes and reliably solve complex route problems such as the travelling salesman problem.
Yes, although these problems are considered difficult to solve in practice, AI algorithms now offer optimal or near-optimal solutions, even for variants with many locations. Large tech companies provide specialised libraries such as Google OR tools for this purpose.
Machine learning (ML) is used to forecast transport volumes more precisely. However, ML is only suitable for complex optimisation problems to a limited extent, as similar logistics situations often require different solutions.
AI systems are controlled via target functions that are based on key performance indicators such as adherence to deadlines, capacity utilisation, sustainability, accuracy and employee safety. The selection of KPIs directly influences the optimisation result.
Logistics experts are developing fewer of their own solutions and are working more closely with AI. They define requirements, check algorithmic proposals and control digital processes. This shifts the focus from manual planning to the overarching design and evaluation of systems.
Research teams are working on deep reinforcement learning and quantum-based methods. In addition, large language models are already providing support in the modelling of complex optimisation problems, for example in the Gurobi AI Modelling Assistant.
