AI is becoming part of logistics execution

AI is often discussed as a future-facing concept, but its value in ecommerce logistics depends on whether it can improve real operational decisions.

The session from Maersk and Mo Fashion focused on how AI can be embedded into parcel network design, routing logic, carrier selection and customer service workflows. In a fragmented European logistics environment, parcels often move through multiple hubs, carriers, trade lanes and service options before reaching the customer.

That complexity creates cost, variability and operational pressure. It also creates an opportunity for better decision-making.

Delivery networks need dynamic optimisation

Traditional logistics networks can be relatively static. A parcel may follow a fixed route or carrier rule based on destination, service level or commercial agreement.

The session explored a more dynamic approach. AI can support network design by analysing demand patterns, carrier performance, capacity, transport cost and service-level requirements. It can also help optimise parcel-level routing by considering disruption risk, trade lane performance and available carrier options at the moment a label is generated.

For ecommerce brands, this matters because two parcels with similar characteristics may not always need the same route.

Data quality is the foundation

A recurring theme was that AI is only useful when the data foundation is strong.

Track-and-trace events, carrier performance data, address quality, postal codes, service levels and customer support signals all need to be clean enough to use. Without that foundation, AI pilots may look promising but become difficult to scale.

Mo Fashion’s perspective highlighted this clearly. AI can assist with data cleaning, address correction, operational simulations and proactive customer service, but the underlying data work remains essential.

AI can support customer experience

The session also showed how AI can help teams move from reactive to proactive customer support.

When parcel events are monitored more intelligently, teams can identify potential issues before the customer raises a ticket. Delays, missing scans or failed deliveries can trigger internal actions, carrier follow-up or customer communication.

This is where AI becomes less about automation for its own sake and more about improving the reliability of the delivery promise.

What this means for the DELIVER community

For retailers and brands, AI in logistics should start with practical use cases: carrier selection, routing, address quality, delivery monitoring and customer service. For logistics providers, the opportunity is to use AI to improve network design, reduce cost variance and create more reliable delivery outcomes.

The strongest AI strategies will be built on clean operational data, measurable use cases and a clear understanding of where logistics complexity creates customer or cost impact.