Unilever reduced detention and demurrage costs by 20% by treating transportation data as a trigger for decisions rather than a dashboard to monitor. That result came from a specific operational shift: moving from knowing where shipments are to deciding what to do because of what the data shows.
Decision iIntelligence is the practice of turning real-time transportation data into specific, recommended or automated actions before a disruption reaches the customer. It builds on visibility. But the two aren’t the same thing, and most supply chain teams are still running on visibility alone.
The cost of relying solely on visibility shows up clearly. When an alert gets generated and nobody acts on it in time, the delay still hits the dock, the warehouse, and the customer. Visibility without action is information without an outcome.
At Reuters Events: Supply Chain Europe 2026, Carlo Andrea Carmenati, Primary & Inbound OpEx Lead Europe at Unilever, joined project44’s Tom Lintsen to discuss how Unilever is moving from shipment tracking toward decision intelligence.
Here are five lessons from their conversation.
1. Stop chasing shipment status
Before real-time visibility, answering a simple question about a shipment would require calls, emails, and messages across teams and carriers. For Unilever, that has changed dramatically.
“My team was used to receiving 100 calls every day, emails, and team chats asking, ‘Where is my truck? Where is my order?’ Nowadays, we have real-time visibility, and everything is in just one click.” Carlo Andrea Carmenati, Primary & Inbound OpEx Lead Europe, Unilever
When teams stop spending their time chasing basic status information, they can shift their focus to the exceptions that need intervention.
2. Turn changing ETAs into decisions
Imagine a truck is scheduled to arrive at a warehouse at 10:00. Its latest ETA now shows 11:30.
Visibility tells you it will be late. Decision intelligence asks: what should change because it will be late?
For Unilever, one answer is dynamic dock scheduling. A significant ETA change can trigger a proposed adjustment to the truck’s dock slot, helping the warehouse respond before the delayed arrival creates further disruption downstream.
In practice, this looks like once an ETA changes, a new dock slot is proposed, then the warehouse operator confirms, and finally the schedule adjusts to reflect the change.
“The ideal scenario is to have dynamic dock booking and rebooking,” Carmenati explained. “If the truck is arriving earlier or later, the idea is to reshuffle the slot.”
Unilever is currently testing this approach with human confirmation still part of the process. That combination matters. AI can identify changes and recommend actions, while people provide the operational context that data alone cannot always capture. The goal isn’t to remove people from the decision,; it’s to reduce the time between a data signal and a well-formed recommendation so the person confirming it can act faster.
3. Let AI do more of the exception chasing
A transportation exception often creates another problem: someone has to figure out why it happened. Information may be missing, a carrier may need to be contacted, and data may need to be collected and entered into a TMS. Each step adds manual work.
Agentic AI creates an opportunity to take some of that work off transportation teams.
“There is an agentic AI sending emails or making phone calls asking for the missing data. The agentic AI then gets the data and puts it back into our TMS system,” said Carmenati.
Rather than generating another alert, an intelligent workflow flags the exception, gathers missing context, and identifies where human attention is required. That shifts the dynamic from “here is the problem” to “here is the problem, here is the context, and here is what needs to happen next.”
4. Connect visibility to measurable outcomes
Decision intelligence only matters if it improves real-world outcomes. Detention and demurrage is one of the clearest places to see that change.
Geofenced arrival and departure timestamps give Unilever an objective record of when shipments reach and leave locations. Carmenati cited a 20% reduction in detention and demurrage costs as a direct result of that improved data quality and accountability.
Reliable transit data also supports inventory decisions. Observed transit times give planners greater confidence when setting buffer stock levels. A delay in a lane that typically runs on time changes the inventory calculation, and when planners can see that delay in real time, they can act on it before a stock-out reaches the customer.
Real-time shipment data helps teams protect on-time, in-full (OTIF) performance by identifying at-risk deliveries earlier. For higher-risk lanes, route deviations and extended stops can trigger alerts, giving teams a window to investigate before a delay reaches the customer.
5. Connect the whole journey
Unilever works with roughly 150 carriers, and there is no single way to connect all of them. One carrier might connect through an API. Another may require license-plate and subcontractor information. Another might rely on a driver app. Visibility at that scale, is a continuous improvement effort across the entire carrier network.
But connectivity is only part of the challenge. The bigger opportunity is connecting decisions across modes.
For example, an ocean delay doesn’t end when a vessel reaches the port. It can affect the road leg, warehouse schedules, inventory availability, and customer commitments.
Carmenati put the challenge this way: “If we tie together the ocean and the road, and give context about the dwell time of an ocean vessel, what does that mean for the impact on my road leg and for the end customer?”
That’s where visibility from vessel departure to final delivery starts becoming decision intelligence across the full journey.
What Unilever’s approach means for your supply chain
Some supply chain teams will hear Unilever’s results and assume the lesson only applies to companies with 150 carriers and a dedicated operations technology team. It doesn’t.
Unilever is automating decisions like dock rebooking, exception investigation, and proactive customer communication. These exist in some form in every transportation operation. What changes with scale is the volume of exceptions and the number of carriers to connect. The underlying logic, turning a data signal into a timely action, applies whether a team is managing 15 carriers or 150.
project44 helps supply chain teams connect real-time transportation data with the decisions and workflows that keep goods moving, customers informed, and operations on track.
To learn more about project44 and see our platform in action, request a demo and join us at signal44.
