Decoding the real world - how AI powers smarter maps for commercial fleets

Nearly every time we get into a car, we interact with a dynamic digital twin of the world - a map. The first iteration of digital maps was static and two-dimensional. Just think back to the early days of satellite-operated, plug-in GPS devices. Today, we rely on consumer maps like Google Maps and Waze and take for granted how they surface real-world driving conditions like road closures or congested roads and reroute accordingly. This next wave of consumer maps overlaid real-time data onto maps to give us a holistic representation of the physical world on our phones.

Commercial fleets need commercial-grade mapping software

For commercial vehicles, digital transformation entails much more than navigating from point A to point B. It’s about reducing costs, protecting their frontline workers with proactive safety measures, and providing the confidence to handle the unexpected in an increasingly fragmented, unpredictable supply chain and transportation environment. From waste management to food services, construction, and retail, complex operations manage billions of dollars in assets and thousands of employees, introducing significant, inherent risk into nearly every facet of commercial navigation.

Imagine for a moment that you operate a commercial vehicle, like a 44-tonne tractor unit, around a congested major city like London or Paris. On a daily basis, you navigate narrow roads, cobblestone streets, distracted pedestrians, low bridges, and dozens of hazards. Perhaps you rely on consumer applications that lack an understanding of commercial vehicles to get around, leading to systemic risk. In the UK alone, five heavy goods vehicles crash into a rail bridge every day. Even more tragically, 14% of road fatalities in Europe involve a heavy goods vehicle. So how do we put an end to life-threatening incidents and protect our frontline workers who keep our communities running?

What physical AI means for commercial navigation

For digital maps to be effective for commercial fleets, they require real-world intelligence, or physical AI. In this context, physical AI is the convergence of real-time data and AI to create a multidimensional understanding and predictive engine of the world. This creates a unique data opportunity, because consumer maps lack visibility into commercial operations data—things like geometries of warehouses and docks, telemetry data, vehicle restrictions, and loading zones.

At Samsara, we process over 20 trillion data points and travel over 90 billion miles each year. This commercial operations data comes from the world’s largest connected network of cameras, sensors, and vehicle gateways across tens of thousands of fleets. It allows us to construct a multifaceted representation of the physical world, and apply intelligence on the edge to help commercial drivers operate more effectively. Unlike the homogenous data analyzed by consumer apps, commercial fleet data ranges from video to telematics, to sensor data and driver commentary, constructing a context-rich, dynamic understanding of the world, physical AI.

Commercial data creates real-world intelligence

For example, video enables us to identify a low bridge, telemetry indicates fuel efficiency of a 44-tonne tractor on rain-soaked cobbled streets, and AI generates situational awareness and real-time, in-cab audio alerts. With physical AI, we have built a map that understands the loading clearance of a bridge, identifies unmarked loading docks, calculates turning radiuses, and seamlessly integrates European compliance.

In the U.S., Melissa Dockery, Operations & Asset Manager, TP Trucking & Logistics, explains how the tool gives drivers:

...smarter routes that actually consider truck restrictions and real-time conditions. It helps us avoid delays, reduce driver stress on the road, and improve efficiency and driver satisfaction. It is intuitive and reliable—a great tool for any operation focused on safety and performance.

Similarly, global flooring manufacturer Mohawk Industries fully digitized their routing process. As a result, they reduced miles driven by 4.2 million, saving $7.75M across their entire fleet. These results illustrate that physical AI will fundamentally reshape the way fleets navigate, particularly for heavy-duty or mixed fleets with complex routes.

A roadmap to the future of intelligent operations

Our global economy, from the shoes on store shelves to the electricity powering our schools and hospitals, runs on a diverse array of physical operations that our digital tools are just beginning to understand. For logistics and transportation, the ambition isn’t visibility, it’s predictive agility. That’s why it’s important to evaluate commercial routing systems for height, weight, and length of vehicle configurations, integrations with regional compliance, and interoperability with live data like fuel levels to optimize routes.

We are just beginning to imagine the future of intelligent operations—an AI that understands the physical world, and, in turn, makes the world safer and more efficient. Ultimately, operators who embrace physical AI across their fleets will lead in the next era of transportation, leveraging commercial maps as fluid and intuitive as any smartphone application and empowering the critical frontline workers behind our global economy.