AI in logistics: only 27% have it in the TMS, the world is at 96% (and you're short 60,000 people)

In logistics and transport — a courier, a warehouse, a third-party haulage operator — the AI conversation starts from a different place than in any other sector. Not from cost, as in manufacturing. Not from a tool already bundled and switched off, as in professional practices. It starts from a question no other supply chain faces with the same urgency: who do you put on the work, if the people can't be found? And it's precisely from there that AI, in this sector, stops being a luxury and becomes the most concrete lever you have.

The figure that frames everything is about labour, not technology. Italy's transport and logistics sector is short of roughly 60,000 qualified professionals — drivers, warehouse and supply-chain staff: not one company's difficulty, but a sector-wide headcount crisis. And a shortage of that scale isn't something you fix by hiring — you've already tried. You fix it by getting more out of the people you have. That's exactly the role AI carves out for itself in logistics: not one worker fewer, but one worker's worth more of capacity, with the repetitive part of the job automated and scarce people freed for the coordination that matters.

The problem isn't the technology, it's the labour you can't find

It's worth pausing here, because it flips the usual way AI gets sold. In most sectors it's pitched as efficiency: do the same thing for less. In logistics the honest message is a different one — do the thing, full stop, because without automation you don't have the people to do it. The sector framing is clear: AI as a capacity multiplier with the human in the loop, not as replacement. By automating repetitive planning and tracking, output per worker can rise by up to 24% — which means fulfilling more shipments with the same headcount, not laying people off.

That's the difference that moves the decision. A logistics business owner who hears "AI saves you money" thinks of a project that can wait. The same owner who hears "AI lets you fulfil the orders you can't fulfil today because you're short-staffed" thinks of a problem for today. In logistics the second sentence is the true one — and it's why this sector has less room to postpone than it thinks.

Where you stand against the rest of the sector: 27 versus 96

Here's the second number that counts, and it captures an uncomfortable distance. Only 27% of Italian logistics companies have integrated AI or machine learning into their transport management system (the TMS) — "3 companies in 10," in the words of Roberto Vismara, Sales Director of Manhattan Associates Italia, commenting on a 2025 Manhattan Associates/Vanson Bourne study. The other 73% work with no AI in the TMS at all. Meanwhile, globally, 96% of transport leaders say they use AI somewhere in planning or operations, and 64% use it specifically to plan and execute transport (BCG, 2026).

The point isn't the lag itself, but against whom. In many sectors the Italian SME trails the large Italian firm, and you console yourself with "I have other priorities." In logistics the distance is against your own global competitors: whoever is fighting you for the same clients is already working with AI in the planning engine. And the intent, in Italy, isn't lacking — 87% consider transport management a strategic priority between now and 2030. What's lacking is execution: between saying "it's strategic" and actually having it in the system sits the whole gap between 27% and 96%.

Where AI pays off in logistics

Four workflows, in transport and logistics, are high-volume, repetitive and rule-driven — AI's natural ground, and the place where the staff shortage bites hardest:

• Dynamic route and network planning — instead of hand-built rounds and static routes. It's the sector's most mature use case: 64% of logistics operators already apply AI here, to plan and execute transport. It's also the workflow that most directly offsets the shortage of experienced planners.
• Shipment visibility and tracking — instead of status checks done by hand, on the phone or by email. Around 50% of the companies surveyed use AI for shipment visibility and quality control, and 60% consider supply-chain visibility crucial to cutting transport costs: fewer "where's my truck" calls, more time on what people do better than the machine.
• Forecasting and optimization inside the TMS — instead of capacity and demand planning kept on spreadsheets. It's the most "green-field" of the four use cases: in Italy today only 27% have it, unlike the verticals where AI is already bundled into the software. Whoever integrates it now moves where almost no one, in Italy, has moved yet.
• Predictive maintenance of fleet and warehouse — instead of calendar-scheduled maintenance or the fix-it-when-it-breaks kind. It's the same mechanism we document in manufacturing, with the same reduction in downtime of around 25%, applied to vehicles and warehouse equipment: an unplanned vehicle down, in a short-staffed sector, is capacity you don't get back.

The criterion for starting isn't "which is the most impressive," but "which gives me back the most hours of people I don't have today": for most operators the answer is route planning — high volume, clear rules, and the most direct effect on the capacity you're missing.

The 27-versus-96 gap is an advantage still on the table

Here is the reading that flips the lag. That only 3 Italian companies in 10 have AI in the TMS isn't just a bad report card: it's the snapshot of an advantage still on the table. In a sector where global competitors are at 96% but the Italian market is stuck at 27%, whoever brings even just route planning into the system pulls ahead of seven companies in ten in their own country — not by hiring more (they can't), but by getting more out of who they already have.

That's why, in logistics, the right question isn't "can I afford AI," but "can I afford to stay among the 73% who don't have it, while whoever is fighting me for clients is already among the 96% who use it, and on top of that I have fewer people than before." Read as a cost, AI in this sector looks postponable; read as the only lever left when hiring no longer suffices, it becomes the least postponable project you have.

The capacity lever runs through people: training and governance

There's a reason why, in this sector more than elsewhere, AI doesn't work like a switch. 63% of logistics companies already recognize they need to invest in reskilling their staff for AI adoption to hold up over time. It makes sense: if AI is a capacity multiplier with the human in the loop, that multiplier is worth exactly as much as the person governing it. A planning tool nobody knows how to read stays switched on to no effect, exactly as in retail.

This is where our approach differs from those who sell the licence alone: in logistics honest implementation is tool plus training plus workflow redesign, not the tool by itself. And it hooks naturally into governance: the human-oversight and training obligations set out by the EU AI Act and gathered in our compliance overlay aren't extra weight: they're the very thing that makes the capacity multiplier reliable instead of fragile.

And compliance? Within limits, but with method

In logistics, AI touches operational decisions that carry weight — on routes, on loads, on timings — and often on data about people (drivers, staff, sometimes recipients). The transparency and documented human oversight duties set out by the EU AI Act, together with the already-familiar data-protection rules, fall on whoever automates planning and tracking. "The human in the loop" isn't just a good way to sell AI to a short-staffed sector: it's also the compliance premise that keeps the automated decision inside a defensible process.

It's exactly what our compliance overlay attaches to every workflow we design: the person who stays in the loop isn't just recovered capacity, it's the guarantee that the AI-generated plan is verifiable and traceable to whoever answers for it. For a logistics operator this isn't extra bureaucracy — it's the way to scale capacity without losing control.

Where to start, in practice

If you run a transport or logistics business and AI is on your radar, the sensible path starts from the capacity problem, not from the software:

• Start from the headcount bottleneck — where the shortage of people slows you most. The first question isn't "which AI am I missing," but "which high-volume workflow gives me back the most hours of people I don't have today."
• Pick a single process — route planning or shipment visibility give the sharpest effect on capacity. Not a rollout across the whole company, one workflow.
• Keep the human in the loop — AI prepares and flags, the person decides on the cases that matter. That's how the capacity multiplier works and how it stays compliant: no operational decision that carries weight left to an automatism without control.
• Train whoever will use it — a planning engine nobody knows how to read recovers no capacity. 63% of the sector has already grasped it: the tool without the training stays switched on to no effect.

Before even choosing the workflow, though, it helps to know where you stand: our AI-readiness assessment helps you understand which process to start from for the most return and least friction, and which controls — and which training — to put around the first workflow. If the issue is the compliance of what automates operational decisions, our compliance overlay explains how we attach the controls to every design.

We've turned the first step into a free, self-serve assessment: a few questions and a pointer on where to start, with which controls around it. Take the AI-readiness assessment — then, if it makes sense, let's talk.

This article is for guidance only. The figures cited (27% of Italian logistics companies with AI/ML integrated in the TMS and 73% without, from a 2025 Manhattan Associates/Vanson Bourne study; 96% of global transport leaders using AI somewhere and 64% in transport planning and execution, BCG 2026; 87% considering transport management a strategic priority by 2030; roughly 60,000 professionals missing from the sector in Italy; up to 24% higher output per worker; around 50% using AI for shipment visibility and 60% considering supply-chain visibility crucial to transport costs; around 25% less downtime with predictive maintenance; 63% recognizing they need to invest in staff reskilling) come from sector sources and should be read as indications of direction, not guarantees of results. Any automation that touches operational decisions or personal data must be assessed on the data, the controls and the context of the individual business.