If you operate a delivery fleet, the question is always the same: how do you deliver orders on time, undamaged, and at the lowest possible cost — with the fewest vehicles, miles driven, and drivers?
It’s a complex optimization problem with a lot of moving parts, which is why most leading fleet operators rely on route planning and optimization solutions. But as the operating environment continues to get more challenging, technology is also getting smarter. Advances in AI and machine learning are enabling companies to better understand the gap between what was planned and what actually happens in the field — unlocking hidden capacity and savings in the process.
So what does that look like in practice? I explored that question and its implications with Cyndi Brandt, VP, Fleet Solutions at Descartes, on a recent episode of Talking Logistics.
Current Challenges
We all know there is significant turmoil internationally and domestically right now, so I began our discussion by asking Cyndi how this is affecting fleet management. Rising fuel costs and volatility are an obvious challenge, but so are driver shortages. Simply put, not enough new drivers are choosing this as a career.
Against this backdrop, Cyndi said fleets are dealing with “this huge combination of rising operating costs,” while also facing “unprecedented delivery expectations” that accelerated during COVID with the growth of business-to-consumer delivery. At the same time, many fleets have underutilized or “locked” capacity they still struggle to access.
Cyndi also pointed to fragmented data and disconnected systems as a major obstacle. “It makes it really difficult to not only optimize your routes, monitor the execution, monitor safety and risk … and then make sure that you don’t have a lot of downtime there as well.”
Employing AI and Machine Learning
How does technology, especially AI and machine learning, help address these challenges?
Cyndi noted that transportation companies have been using machine learning for decades and are now entering the next generation of those capabilities. Yet despite that history, many fleets still rely on rough estimates of what will happen throughout the day.
As Cyndi put it, “All those little inaccuracies start to add up over time. It’s like compound interest.” Across an entire fleet, that becomes “a massive problem” that creates inefficient routes.
In contrast, machine learning can predict service times much more accurately based on real-world execution data and multiple inputs. “In the real world, every single stop is different because of the location, the delivery type, the time of day, even the product that’s being delivered,” Cyndi said. “And what happens is that plan that looks perfect on paper completely falls apart in execution because the modeling was bad.”
Using machine learning to analyze execution history and other variables enables fleets to create much more realistic route plans. “What it really truly means for the company,” Cyndi explained, “is more on-time deliveries, greater route density, better ETA information … less stress for the drivers … but ultimately lower operating costs.”
She also noted that the “secret sauce” lies in the algorithms that can parse and analyze multiple data inputs with far greater sophistication than in the past.
Increasing Route Density
I asked Cyndi to expand on the benefits of increasing route density, which research shows can improve by as much as 30%. In practice, she said, that means being able to serve more customers and make more stops without expanding the fleet or increasing driver hours.
In her words, “It’s growth. Growth with no extra drivers, no extra vehicles.”
She added that better route density also improves fleet utilization by reducing idle time, unnecessary miles, overtime, and fuel consumption. And ultimately, she said, “it’s about reducing the cost of delivery.” When labor, fuel, insurance, and maintenance costs are spread across more stops — without compromising safety — the cost per delivery comes down.
The Benefits
Beyond increased route density, what other benefits can fleets gain from better service-time predictions and more accurate routing? Cyndi highlighted one in particular: predictability.
“Predictability is one of those things that contributes to driver retention greatly,” she said. Drivers want to know how long a route is planned for and how long it is actually likely to take. When that gap is too large, frustration rises and retention suffers.
She also noted that by eliminating small inefficiencies and inconsistencies, companies can reduce overtime, missed deliveries, and re-dispatching while improving customer service. As she put it, “Accuracy doesn’t just improve efficiency — it takes chaos away from the complete process.”
How Will AI and Machine Learning Reshape Fleet Operations?
Although AI and machine learning have been used in transportation for decades, they seem to have reached an inflection point where more impactful outcomes are now within reach. Looking ahead, Cyndi highlighted four key shifts that point toward a future where AI and data-driven insights play a much more active and continuous role in fleet planning and execution. As Cyndi put it, “We’re just at the tip of the spear here.”
I encourage you to watch the full episode for her insights on what’s coming next. Then keep the conversation going with your own comments and questions.
