Wednesday, May 6th, 2026
On 30 April 2026, ALICE, together with the EU-funded IKIGAI project and KEDGE Business School, hosted the third and final webinar in the IPIC 2026 series, focusing on how artificial intelligence (AI) can support the development of the Physical Internet (PI). The session brought together research perspectives on how data-driven approaches can enable more efficient, collaborative and adaptive logistics systems, particularly in complex urban environments.
Opening the webinar, Pablo Segura (ALICE) highlighted the role of the Physical Internet in achieving climate-neutral and efficient logistics. He emphasised that interconnected systems, shared resources and coordinated operations increasingly depend on data and intelligent decision-making. In this context, AI is not an add-on, but a key enabler of scalable and responsive Physical Internet solutions.
A central theme of the session was the need for a shared operational foundation in urban logistics. Tianyuan Zhang (KEDGE Business School) presented a data-driven framework that transforms open data into multi-tier logistics districts, creating a common map of the city. Today, logistics zones are typically defined independently by operators, cities or researchers, making collaboration difficult and limiting comparability. By contrast, a shared and reproducible districting approach can support coordination, benchmarking and more transparent decision-making across stakeholders.
The framework connects directly to Physical Internet principles by translating conceptual ideas, such as urban “pixelisation” and multi-level logistics networks, into real-world applications based on actual city data. Using open sources such as road networks and building footprints, the approach enables scalable and adaptable territorial design. It also allows operators to integrate their own demand data, ensuring practical relevance while maintaining flexibility.
The second presentation, delivered by Nafe Moradkhani (KEDGE Business School), focused on how AI can support collaboration decisions within Physical Internet systems. While the PI promotes openness and resource sharing, the session highlighted that collaboration is not always equally beneficial. Instead, its value depends on context, including demand patterns, capacity constraints and operational conditions.
To address this, the proposed framework combines machine learning and optimisation to determine when and how collaboration should be activated. Rather than assuming full openness, the model identifies the appropriate level of cooperation, ranging from fully private operations to partial or fully shared systems. This enables more targeted and efficient collaboration, avoiding unnecessary coordination costs while still capturing the benefits of resource pooling.
Key factors influencing these decisions include workload pressure, interaction between actors, service performance and how conditions evolve over time. By analysing these signals, AI models can predict the most suitable level of openness, while optimisation methods determine how resources and operations should be configured in practice.
Results from a real-world case study showed that selective collaboration can significantly improve performance, particularly by reducing overload in urban delivery systems. Importantly, the findings suggest that full openness is not always required. Instead, the Physical Internet can be implemented progressively, with collaboration activated where and when it creates value.
Across both presentations, a clear message emerged: the Physical Internet requires both structure and adaptability. Shared spatial frameworks enable coordination, while AI-driven decision-making allows systems to respond dynamically to changing conditions. Together, these elements support the transition from fragmented logistics towards more integrated and resilient networks.
The webinar also highlighted strong links with ongoing European initiatives, particularly the IKIGAI project, which focuses on scaling up innovation in Physical Internet systems. These approaches are complementary, combining technical frameworks, operational tools and data-driven methods to support real-world deployment.
The session concluded with an invitation to continue the discussion at the International Physical Internet Conference (IPIC 2026) in Bordeaux. As logistics systems become increasingly complex and data-driven, artificial intelligence is expected to play a central role in enabling the next generation of collaborative, efficient and sustainable logistics networks.
Read the full report for detailed insights and stakeholder perspectives.