Any discussion on the uses of AI in most domains is dominated by a very technology-centric perspective, with the benefits seen as overwhelming and indisputable, and the embrace assumed to be inevitable. This narrative, naturally, leans heavily on the viewpoint of the creators and designers of data analytics and AI-based solutions. For most enterprises used to legacy solutions and a playbook around marketing and selling the same, the sheer scale of the change in front of them can be overwhelming, even though they might not dispute the value such technologies can eventually accrue to their businesses. In most situations, save outlier early-adopters, it becomes necessary to look at the adoption of analytics as a continuum of change and value creation, across multiple stages of the adoption curve. For fleet management, where AI-driven advances have been coming fast and thick, this is especially relevant.
Traditional fleet management has typically been driven by data around the location, speed, and other parameters exposed by the vehicle bus (ignition status, fuel consumption, etc.). Tracking an asset in real-time, keeping an eye on vehicle utilization and fuel use, and regulatory compliance, have been the primary use-cases around which fleet management solutions were built. The advent of AI has uncovered use-cases not possible before, through a combination of hardware and software innovations. Fleets, however, are slow in making leaps of faith into technologies whose value to their bottom lines is unproven. Conversely, they do live with real everyday problems that need fixing, and in the bridging of this gap is where the most effective path towards AI adoption lies.
Let us look at some examples around how AI is making a material difference to fleets today, and priming them for a continuum of innovations down the road:
From ELD applications to a variety of other use cases that need to identify the person behind the wheel, driver identification is becoming a critical part of any TSPs workflow. Traditional telematics has been asset-centric, with CAN devices being a reliable source of unique vehicle/asset identifiers. With driver behavior now becoming a central focus of telematics workflows, knowing who is driving a vehicle at any point of time reliably is becoming unavoidable. Interim solutions have looked at RFID based sign-in, driver log-in through mobile applications and a variety of convoluted ways to map drivers to assets. This is an instance where AI is uniquely positioned to deliver huge benefits today. Using face recognition on video or images captured by either driver-facing cameras or mobile devices helps create seamless driver registration workflows, removing a lot of overhead in time and effort for fleet managers.
Driver fatigue and distraction
While HOS and other similar mandates worldwide have helped (to an extent) in addressing issues around over-work, fatigue remains a major cause for accidents on our roads and highways. To make things worse, with each passing year the epidemic around distraction caused by cell phones continues unabated. This has severe consequences for fleets, who are for the most part grappling unsuccessfully with this problem. Piecemeal solutions that monitor or lock mobile devices when the driver is driving, however useful, attempt to solve a small part of the problem. DMS, or essentially the ability for AI on driver-facing cameras to actively warn drivers and protect fleets, covers an umbrella of use cases around improving in-cab safety – cell phone use, drowsiness, seat belts not worn, and many more potentially dangerous behaviors. Though challenges remain around privacy, that we have talked about before, in many cases fleets are finding that with the proper messaging around benefits, drivers can be brought on board.
The entire process around insurance claims and settlements starts with the First Notice of Loss (FNOL). Till now this has been a predominantly manual process, made especially challenging around critical accidents involving death or serious injuries. The modern commercial vehicle, however, is now a moving IoT device, laden with CAN devices reading off the vehicle bus, GPS, g-sensors and increasingly cameras looking at both road and driver. AI on these synchronized multi-modal data streams can now very accurately detect impacts and other critical incidents, automatically triggering a process of notifying first responders, law enforcement and insurance, along with a rich and informative data set including velocity and g-sensor profiles (and optionally, video), to make the claims management process significantly more efficient. In such critical situations where every second matters, AI can meaningfully and materially affect fleet operations today.
This is by no means an exhaustive list, and from vehicle inspection reports to trailer monitoring, and more, there are opportunities to use AI to remove existing fleet pain points at every turn. For creators and sellers of such technologies, looking at existing workflows and finding natural efficiencies that AI can bring are a great place to start. Once the on-boarded and incomplete realization of their benefits, fleets will be much more amenable to walking along the innovation journey with you.