Dnata has identified the pace of technology deployment — not the technology itself — as the central failure mode in the troubled July 2025 opening of its highly automated cargo facility at Amsterdam Schiphol, with chief cargo officer Guillaume Crozier telling the Aviation Connect event in Athens that operators should phase automation rather than deploy it end-to-end in a single move.

The 61,000 sq m facility combines automated storage and retrieval, Automated Guided Vehicles (AGVs), Elevated Transfer Vehicles (ETV), smart gates, weight and dimension systems, forklift guidance systems and truck management systems. When operations launched in July 2025, the site faced data issues across its various systems, and dnata had to clear a backlog of volumes before completing the transfer of all traffic from its previous facility.

Crozier was direct about what went wrong. The key challenges of moving into the new building were the simultaneous ramp-up of volumes, systems and people, combined with limited data visibility and performance insight. Without dashboards and reliable metrics from day one, management lacked the instruments to diagnose problems as they emerged.

The recovery effort followed a consistent pattern. Dnata stabilised operations, transferred remaining volumes from the old facility, used data and dashboards to drive decision-making, improved end-to-end collaboration with stakeholders and established key performance indicators. Today, Crozier said, the facility is "ticking nicely" and receiving good feedback from the local market.

The phasing argument

His most pointed lesson concerns sequencing. On paper, launching an integrated automation suite in one step looks rational — because cargo handling is an end-to-end process, automating one element in isolation appears to leave gaps downstream.

"It sounds right in terms of planning," Crozier said. "It will be a big leap, so to have all of that together, it makes sense. As it is an end-to-end process, you think end to end — if you start automating one thing, you want to cover it throughout the process."

"Reflecting on the situation, maybe not always," he continued. "[Management of the] pace of innovation you are delivering, especially when it is something that has not been seen in the market, is potentially better."

The AGV deployment illustrated the point. Crozier stressed that operators must fully understand the material flow around the vehicles before putting them into service — knowledge that in practice only emerges once the surrounding processes are running predictably.

Readiness, buffers and local knowledge

Beyond phasing, Crozier listed the conditions he now considers prerequisites for large automation projects. Early operational readiness, validated through simulations, tops the list. So does having usable data from day one, clear governance structures and close working relationships with key stakeholders — a category that at a congested hub like Schiphol inevitably includes the airport, airlines and regulator-facing counterparts.

Timeline discipline also mattered. The project experience showed that schedules need enough buffer built in to absorb unexpected setbacks, an admission that the original plan left insufficient margin around a facility whose combination of systems had no direct precedent in the market.

Local market knowledge proved indispensable. Crozier noted there is a lot of local specificity that cannot be managed from overseas, a caution for global handlers attempting to standardise automated operations across regions. Contingency planning, change management and training rounded out his list of success factors.

The labour driver

The Schiphol experience has not dimmed dnata's commitment to automation. Crozier argued that automation of handling facilities is becoming increasingly important as recruitment becomes more challenging — a structural constraint facing ground handlers and cargo operators across Europe's major hubs.

That tension defines the industry's current position: labour scarcity pushes handlers toward automation faster than their operational readiness may warrant, while the Schiphol case demonstrates the cost of compressing that transition into a single cutover. The facility's stabilisation shows the problems were solvable; the backlog and disruption in the opening months show they were also expensive.

For handlers planning comparable automated facilities elsewhere, dnata's account offers a working template: simulate before launch, secure data visibility from the first day, phase the technology, and budget time for failure. Whether that template shortens ramp-up timelines at dnata's next automated site will be the test of the lessons learned.