In this episode we talk to the ops directors at two mid-cap terminals about the AI bets they made last year — which landed, which didn't, and why.
A 28-minute conversation with terminal operators on where AI has shifted the needle — and where it hasn't.
Both operators saw real uplift on gate turn-time (12–18% faster) and on predictive maintenance (40%+ fewer unplanned crane downtimes). Both say the biggest win wasn't the AI itself — it was the data discipline the AI forced on them.
Berth-allocation AI underwhelmed for one of them. The operator's service agreements were too constrained — the AI kept recommending allocations that violated contract terms they couldn't surface to the model. Honest take: feature-complete on paper, not ready for the real contract graph.
The other terminal had a better experience, mostly because their contract graph was simpler and more of it was already in structured data.
Start with the boring wins. Gate and crane predictive maintenance have short payback and small blast radius. Leave berth-allocation AI until your contract graph is in structured data. Don't let a vendor convince you otherwise.
One email a fortnight. Field notes + case studies + early access to webinars.