Why Fleet AI Works Best When You Start with One Problem, not a Transformation. Insights from Aptean’s 2026 Artificial Intelligence Research
Walk onto the operations floor of many UK fleet businesses today and the scene is instantly recognisable. A planner with a clipboard, a whiteboard covered in routes, and a team working hard to stay on top of whatever the day brings.
As Ben Glossop, VP of Sales at Aptean, puts it, none of that is a sign of a business falling behind, it’s simply the reality of an industry defined by tight margins, unpredictable disruption, and very little time to experiment with unproven technology.
“This is how a lot of companies still work, and it’s exactly why the most successful AI programmes in transportation don’t begin with sweeping transformation. They begin with one well-chosen business challenge.” Ben Glossop, VP of Sales, Aptean
Strong Interest, Patchy Adoption
Aptean’s 2026 Artificial Intelligence Research, conducted by UK, independent research agency Vanson Bourne, surveyed 325 transport and distribution executives across six markets: the UK, France, Germany, the Netherlands, the US, and Canada.
The pattern it found is an interesting one for transport operators. Leaders in the European markets surveyed are more likely than their North American counterparts to say industry-specific AI, models trained on sector data rather than general-purpose tools, matter to their future, and less likely to have adopted it.
- 35% of leaders across the European markets surveyed say industry-specific AI is critically important, against 22% in the US and Canada
- Yet 41% of European respondents have adopted industry-specific AI, against 57% in the US and Canada
Confidence in the safeguards around AI runs the same way. Just over half of European respondents in the markets surveyed (53%) report clear procedures for handling AI errors, against 65% in the US and Canada. On ongoing monitoring of AI accuracy, it is 59% against 73%.
Belief is not the constraint. Trust in the frameworks around AI is still catching up.
That shapes how a fleet should start. Confidence in a system is not built by argument; it is built by watching it work on something small enough to check. A narrow deployment with a result you can verify is how the operating trust those numbers say is missing gets made.
The Gap Between What AI Could Do and What It’s Doing
Glossop sees the same pattern across UK operations.
“Operators want earlier warnings when something’s going off plan, not after the fact. Knowing there’s a delay on the M6 before it snowballs into three missed deliveries allows a planner to act, not just react.” Ben Glossop, VP of Sales, Aptean
Across the six markets surveyed, 77% of respondents said AI’s value is clear but not yet fully realised. The most reported gains were better decision-making and lower operating costs. Yet transport and distribution recorded one of the lowest improvements in on-time, in-full (OTIF) delivery of any sector in the study.
Across the six markets surveyed, 77% of respondents said AI’s value is clear but not yet fully realised. Transport and distribution beat the cross-sector average on the gains AI is delivering – 47% report better decision-making against a 45% average, 44% report lower operating costs against 40%. Yet on on-time delivery and order fulfilment, the sector came last of the five surveyed, with 12% reporting a significant improvement.
That combination is the whole story. The value is visible, but it isn’t reaching the number that matters most to a transport team. For most operations, the problem isn’t appetite. It’s that AI hasn’t been pointed at the right problems yet.
One Problem, One Win
A US transport operator interviewed for Aptean’s research offers a clear example of what a narrow focus and starting point can achieve. Rather than attempting a broad programme, the business went after a single recurring cost: dispatch and route errors caused by vehicle breakdowns.
It deployed AI that flagged equipment likely to fail within the next 500 to 2,000 miles, giving planners time to reassign loads before a vehicle went off the road. The operator reported error rates falling by around 90%.
That is the pattern worth copying – one well-scoped deployment, aimed at a cost the business already knows it carries.
The Data You Already Have
There is an asset in most fleet operations that nobody thinks of as an AI asset. Telematics feeds, electronic proof of delivery, tachograph records, dwell and detention logs, carrier performance history – this is data you already capture because a customer, an auditor, or a regulator might ask for it.
It is timestamped and attributable because compliance demands both. It is complete because gaps create problems of their own. Those are precisely the qualities an AI system needs from its inputs, and most fleets are closer to ready than they assume. The work is connecting what is already there rather than collecting something new.
Where Should UK Fleets Start
Glossop’s advice is straightforward.
“Don’t start by asking how to use AI. Start by identifying the one recurring problem that costs you the most, in terms of time, money, or predictability,” says Glossop.
From there:
- Evaluate the tool internally first, the same way you would onboard a new employee.
- Prove it on one well-scoped use case.
- Let that result make the case for the next step.
“The first thing people ask is: does it work?” Glossop says. “Once companies have something in place they can count on, that’s when you hear, ‘I wish we’d done this sooner.'”
The clipboard and the whiteboard aren’t the enemy. They’re a perfectly effective way to run a fleet, until a well-scoped AI tool proves it can do one part of that job better.
The full study breaks these findings down by market and by sector, including where transport and distribution sits against the other four industries surveyed.
Read the full report:
AI in Transport & Distribution: 2026 AI Report | Aptean
Transport & Logistics Magazine – Driving The Industry Forward
