Geotab's Oliver Holt (Senior Sales Manager, EMEA) reveals how to max the benefits of Telematics

UK fleets face pressure from rising costs, tighter compliance and decarbonisation targets. Across Europe, which operational changes are delivering the fastest measurable gains for fleets?
I would start with getting more productive work from the fleet you already have. Reducing unnecessary engine idling, improving vehicle allocation and helping drivers adopt more efficient driving habits are changes operators can measure without waiting for a replacement programme.
Belgian postal operator bpost is a good example. Its fuel-efficiency monitoring and driver-training programme supported a 10% reduction in diesel consumption and associated exhaust emissions, with reported fuel savings of €1.8 million.
The first step is to compare vehicles doing similar work and identify avoidable fuel use and downtime. Then give somebody responsibility for acting on the findings. Track cost per mile and vehicle availability alongside service performance, so a saving in one area does not create a problem elsewhere.

Fleet managers have access to more vehicle and driver data than ever. Where are fleets failing to turn that information into meaningful cost savings or better operational decisions?
The gap is often between identifying a problem and acting on it. A report showing excessive fuel consumption has limited value unless somebody investigates the cause and follows through.
Go-Ahead London monitors low oil pressure and identifies eight to ten buses a month at risk of serious engine failure, with savings of up to £20,000 per engine. Monitoring coolant temperature and detecting faulty sensors also helps prevent breakdowns. The value comes from giving engineers information early enough to intervene.
It’s important to understand the context, too. Two vans can have different fuel consumption because of their loads, routes, condition or driving patterns. Managers need enough information to identify the cause, assign an action and check the result. That discipline is where the savings come from.

Geotab data suggests some fleet assets can sit idle for much of the year. How can operators use data to decide whether they have the right number and type of vehicles? And when should rightsizing come before replacing vehicles?
Before replacing a vehicle, establish that the business still needs it and that its specification matches the work. Otherwise, you risk carrying an existing inefficiency into the next purchasing cycle.
Also look beyond annual mileage. Examine days used, hours in service, peak demand, payload requirements and the availability of suitable vehicles elsewhere in the organisation.
Distinguish vehicles awaiting repairs from those held for seasonal demand or emergency cover. Low utilisation needs an explanation before disposing of vehicles. At bpost again, utilisation analysis supported a 2.1% fleet reduction, equivalent to 210 vans. That illustrates the value of reviewing the fleet before ordering replacements.
Rightsizing should come first when the evidence shows that work can be covered reliably through reassignment, shared vehicles or a different vehicle mix. The aim is enough capacity to deliver the service, including a justified operational reserve.

Driver safety technology is becoming increasingly predictive. How well can data and AI identify high-risk behaviour before it results in a collision? And how should fleets use those insights?
Data and AI can identify patterns associated with higher risk and help managers direct support where it is needed. They cannot tell us with certainty that a particular driver will have a collision.
We should distinguish an immediate warning about distraction or following distance from an assessment of risk developing over time. An urgent warning should be brief and relevant. Longer-term coaching should focus on a manageable number of behaviours, supported by evidence and practical advice.
Circet’s UK and Ireland operation recorded a nearly 16% improvement in its driver safety score during the first three months, with improvements in harsh acceleration, braking and cornering. That is a measured improvement in driving behaviour, rather than a claim about collision reduction. Drivers also need a voice. A braking event might reflect somebody avoiding a hazard. Coaching should recognise good decisions and improvement, while making clear how technology can protect drivers through evidence after an incident.

When deciding which vehicles to electrify, what does the data tell you fleet managers should analyse first? 
Start with the work each vehicle actually does: daily mileage, the longest regular journeys, payload, time parked and access to charging. Test those requirements against realistic range, including winter conditions and equipment that draws additional power. Then assess the full ownership cost, including charging infrastructure and the electricity the fleet expects to use.
Our work with United Utilities shows why parking and charging arrangements belong in the vehicle assessment from the outset. A suitable vehicle still needs a workable charging plan.
There are also applications that deserve a closer look before operators rule them out. The University of Lincoln and Trinsic cold-chain project used Geotab data to evaluate electric delivery vehicles. It found lower running costs on shorter routes, particularly with off-peak charging. Those advantages diminished or reversed on longer routes requiring frequent public charging.
Refrigerated delivery can therefore be a candidate, provided the route and charging evidence support it.

Where are you seeing AI produce genuinely measurable value today rather than simply making existing reports easier to access?
AI is a strategic priority for Geotab because it can help operators deliver more with the vehicles, people and resources they already have. That requires using it to improve operational decisions and how work gets done.
Route planning provides a measurable example. In Geotab’s 2026 UK connected-fleets report, 75% of UK respondents reported improved on-time delivery through AI-powered route optimisation, 71% reported shorter last-mile delivery times and 49% reported lower fuel costs. Those are survey-reported benefits, rather than the size of the improvements, but they demonstrate practical value in daily operations.
AI video analysis is another application. It can identify distraction or close following, support immediate feedback and help managers prioritise events for coaching. The benefit depends on acting on that information and tracking improvements in driving behaviour.
We are also applying AI throughout our own business. Our experience has shown that adoption needs to change how teams work. Leaders need to use the tools themselves, make experimentation straightforward and help colleagues share approaches that work.
For fleets, I would apply the same discipline. Pick a recurring operational problem, establish a baseline and give the team the tools and authority to improve it. Measure completed jobs, cost per mile, vehicle availability or time spent on administration.
Our longer-term ambition is to connect these capabilities across fleet operations, so AI can help identify a problem, assess the options and support action, with clear rules for when a manager must make the decision.

As fleets combine telematics, maintenance, charging, fuel, video and other operational systems, is integration becoming as important as the individual technologies themselves? What does a genuinely joined-up fleet data strategy look like?
Yes, because many fleet decisions depend on information held across several systems. Dispatch needs to know that a vehicle is available, maintained and sufficiently charged for its next job. Charging decisions need to reflect when vehicles must leave and the energy their routes require.
At bpost, combining route distances, remaining range and live battery charge data enabled an alternating charging schedule. Some vans charge every two, three or four days, reducing site power requirements while supporting their delivery schedules. 
A joined-up strategy starts with those operational decisions. It needs consistent vehicle records, reliable data exchange and clear responsibility for data quality and action. An open platform helps fleets connect different vehicles and equipment as their requirements change. The test is how well that information supports everyday decisions.

Looking three to five years ahead, what should fleets be preparing for now? What capability do you think will most change the fleet manager’s job – predictive AI, greater vehicle connectivity, automated compliance, smarter EV energy management or something else?
We expect AI to have the greatest effect on the fleet manager’s job, particularly as it becomes capable of coordinating work across connected systems. Connectivity supplies the information; AI can help interpret it and carry out authorised tasks.
Consider a vehicle developing a fault before a busy day. The opportunity is for AI to assess the likely operational impact, identify a suitable replacement, propose changes to the schedule and help organise maintenance. In an electric fleet, that assessment would also need to account for charge levels and charging availability. That is the direction we should prepare for, with people setting the objectives and retaining authority over consequential decisions.
I suggest starting with a defined workflow, such as reviewing maintenance exceptions or preparing the next day’s vehicle allocation. Give people secure access to the relevant information, set clear approval boundaries and assess the effect on output and service quality. Successful approaches can then be shared with colleagues and inspire them to take action. 
Over time, I expect fleet managers to spend less time assembling information and coordinating routine tasks, and more time managing exceptions, supporting drivers and making commercial decisions. Operators who build those skills now will be well placed to put each improvement in AI to work.