BTOB NEWS No. 54 – March 2025
Will artificial intelligence revolutionise wholesale trade? In any case, it offers staggering prospects for productivity gains in most professions, from logistics to sales and human resources. Provided that its limitations, risks and hidden costs are also taken into account. By Anne Denis
Excerpt from the article […]
Logistics: significant potential for transformation
In the transport and logistics sectors, the potential of AI is promising. Stéphane Antiglio, also chairman of the PHE-Autodistribution group, confirms this: “Well-processed data already enables us to optimise deliveries and the allocation of space in our warehouses.”
This technology is also accessible to SMEs, according to Boris Richeux, managing director of fruit and vegetable wholesaler Bratigny in Rungis: “I think the phenomenal computing power being developed in France will be useful to our sector. Take the case of a customer who places orders across all our departments: AI can help the order picker optimise the “Lego-style” assembly of their pallet, taking into account the sizes of the different packages. This task, which is almost impossible for a human, will be very easy for AI.” Guillaume Desveaux is president of the AI Cargo Foundation, an association created in 2020 to help players in the transport and logistics sector integrate ‘neutral and sovereign’ AI tools into the supply chain and, above all, use them to decarbonise their activities. He distinguishes between two main categories of artificial intelligence: ‘Deterministic AI has been around for 40 years and allows predictions to be made using historical data; generative AI, which appeared in 2022, is a real anthropological breakthrough because it can create original content from data.’ According to him, concrete use cases generally combine the two categories.
Which applications are relevant to the B2B sector? Guillaume Desveaux recommends taking inspiration from Amazon, even though the American giant is the benchmark in B2C, ‘because it is the world’s leading logistics provider and is data native’. The American giant’s greatest strength, he explains, is that it has a huge volume of data at its disposal, enabling its predictive AI to achieve a very high level of accuracy. This allows for extreme optimisation of supplies, which seems impossible to compete with. ‘Why not pool data from several wholesalers to train AI models and improve their predictability?’ suggests Guillaume Desveaux. The confidentiality, or even anonymisation, of each player’s data can be guaranteed, he points out, thanks to the intervention of a trusted third party – one of the missions of the AI Cargo Foundation.
This mass of data also makes it possible to optimise warehouse management by predicting orders before 
customers even place them. ‘Today, Amazon keeps trucks loaded and ready to go because, statistically, within 15 minutes, there will be orders matching a prediction.’ Here too, there are potential gains in time and productivity that B2B can learn from. As for delivery routes, Guillaume Desveaux believes that wholesalers can still improve efficiency through pooling: “What is optimised at the unit level is not necessarily optimised at the overall system level. AI makes it possible to change scale.” He cites the case of the Rungis-based start-up Califrais, which consolidates deliveries and with which AI Cargo collaborates. ‘If three restaurants on the same street in Paris place orders every day with several wholesalers in Rungis, Califrais can coordinate these flows as they leave Rungis. And a single van, instead of a dozen, will deliver to this street.’ This has both economic and environmental benefits.
According to him, AI also opens up new prospects for the pharmaceutical distribution sector: “Its model requires deliveries to be made to pharmacies twice a day so that patients can receive their medication within the same day. If we re-examine the real needs, we can change the architecture of the system and dramatically reduce environmental costs by decoupling restocking flows from stock orders. This would be impossible without AI, given the complexity of the data models.” In line with its decarbonisation objective, the AI Cargo Foundation has also designed the ‘Appel d’aiR’ platform for shippers to encourage a modal shift from road (89% of freight) to rail and river transport. “Even though their environmental awareness is stronger than before, professionals will only choose this modal shift if it is competitive. But given that a train or barge is equivalent to 40 lorries, it is only by pooling the needs of several players that we will be able to saturate these modes of transport,” explains Guillaume Desveaux. After modelling the entire rail and river networks and collecting millions of transport orders to map freight flows, Appel d’aiR runs AI algorithms to simulate the number of trains or barges needed for optimal capacity utilisation.
Read the full issue of BtoB News No. 54, the magazine of the Confédération des Grossistes de France (CGF)
☑️ Browse: https://lnkd.in/eCvfr2iA
✅ Download: https://lnkd.in/ef3WKA3w
Editors: CGF, Anne Denis, external contributors
Design: Agence Cithéa