India’s road-safety reality is unambiguous and, for commercial fleets, it’s impossible to ignore. In the calendar year 2023, the country recorded 4,80,583 crashes, 1,72,890 deaths, and 4,62,825 injuries, roughly 55 crashes and 20 lives lost every hour. Viewed over time, fatalities have hovered above 1.5 lakh annually since 2018. These are not just public safety figures; they translate into daily operating risk for logistics networks that keep the economy moving. The human toll is immense, and the business fallout, ranging from insurance claims and operational downtime to unpredictable delivery performance, is just as significant.
Commercial fleets sit at the sharpest end of this exposure. Long routes, tight service windows and congested hubs compress decision time, raise fatigue and increase the chances of a momentary lapse turning into an incident. The traditional response to risk in fleet technology was centered around efficiency and security. Mainly, telematics were used for tracking, monitoring, and surveillance, which was a reactive and often punitive approach focused on the asset.
That is now shifting as AI and edge computing are moving safety from “after the fact” to “in the moment” and, equally important, transitioning from surveillance to support. The aim is clear and straightforward. It is to reduce incidents, enable better decisions and cultivate higher trust.
Policy is pushing in the same direction. The government issued a draft amendment on March 2025, to the Central Motor Vehicles Rules stating that, from April 1, 2026, for new models and October 1, 2026, for existing models, heavy passenger and goods vehicles should be equipped with advanced driver assistance, including Driver Drowsiness & Attention Warning, Lane Departure Warning, Advanced Emergency Braking and Vehicle Stability Function. This draft signals that proactive driver assistance will become the baseline on Indian Roads.
It also sits alongside earlier safety mandates, such as the 2021 notification that made the co-passenger airbag mandatory, cumulatively highlighting a “safety-first” posture in regulation.
The “Edge” Advantage: The Real-Time Enabler
The critical breakthrough enabling this new era is moving AI processing from the distant cloud to the device itself. This is edge computing. In the past, a risky event or a good one would have to be recorded, uploaded, and reviewed hours or days later. On the road, latency makes proactive intervention impossible. Edge computing changes the game by running deep-learning models directly on a multi-camera device; the system functions like a real-time “virtual co-pilot.” It has “eyes” in the form of cameras and “brains” in an onboard AI processor. This allows it to analyze 100% of the driving time, and this “on-device” intelligence means alerts are instantaneous.
The moment the system detects a high-risk behavior, it provides an immediate in-cab audio cue. This isn’t surveillance; it’s intervention. It’s the digital equivalent of a co-passenger saying, “Watch out!” or “Put your phone down”, an alert that can correct dangerous behavior, the very second it occurs. AI-based Advanced Driver Monitoring Systems use in-cab sensors that focus specifically on the driver’s eyes. Using sleep science, these systems apply validated biometric measurements, like the Percentage of Eyelid Closure over the Pupil over Time (PERCLOS), to detect the earliest signs of fatigue, distraction, and over-speeding. The AI intervenes with a real-time alert before a lapse becomes a catastrophe.
Simultaneously, AI-powered outward-facing cameras continuously scan the environment and calculate the probability of a collision. This is crucial because even when a driver is driving well, other vehicles on the road may not be. The AI alerts the driver to these impending risks, giving them the precious seconds needed to react. What makes this tech powerful is that they are highly contextual. The AI doesn’t just give blind alerts; it understands the driver’s normal patterns and, critically, the surrounding situation. Drivers manage numerous complex situations, and a necessary action in one context, such as accelerating to avoid a collision, might appear to be a violation in another. The AI understands this distinction. Cutting alert noise while preserving sensitivity is what keeps the “ AI co-pilot” credible to the person who matters most: the driver.
Beyond Prevention: AI as a Catalyst for Cultural Change
Technology alone does not change culture, but it can make culture change possible. For years, safety programmes leaned on “gotcha” moments: catch the mistake, flag the driver, repeat. That model delivers compliance on paper and cynicism in the cab. The alternative is positive reinforcement, recognising smart decisions in real time, from creating space to avoid a hazard to maintaining buffers in heavy traffic, and reinforcing “safe streaks” over hours and days. Vision-based, edge-AI platforms that analyse full drive time and turn those moments into coaching and recognition have reframed the relationship between drivers and safety teams. Evidence from fleets shows that when drivers feel supported (and not surveilled), safety sticks and retention improve.
Netradyne’s own work with GreenZone® scoring, which is the industry’s first driver score that publishes both safe and risky driving behaviors, shows that as driver scores rise, collision rates fall. The GreenZone Score is calculated by analyzing 100% of driving time using AI and edge computing to detect both positive and risky behaviors. Safe actions like smooth braking and attentive driving improve the score, while violations like speeding or distraction reduce it. The score updates in real time, offering a live, accurate view of driver performance. Internal analyses showing that a 50-point increase in GreenZone correlates with roughly a 13-15% reduction in collision frequency; crucially, the scoring arises from both safe and risky behaviours, so drivers see credit for what they do right.
It is also worth connecting safety to the business math because that is where scale happens. Every avoided incident reduces volatility in cost-per-kilometre: fewer claims, fewer unplanned repairs, steadier asset availability and more predictable delivery performance. This is why many leadership teams now view driver assistance and coaching not as a compliance overhead but as an operations lever that stabilises service levels and improves customer experience.
All of this only works if trust is a metric and not a slogan. Drivers must know that the system exists to keep them safe, protect them from false narratives and credit their skill. That is the difference between a technology you install and a technology people will use. The arc of policy is supportive, the technology is mature, and the Indian operating context (diverse traffic, improving highways, uneven connectivity) makes real-time, on-device assistance a “must,” not a “nice to have.” Put differently, we can now turn seconds into outcomes at scale.
The Road Ahead
The results from early adopters are already clear. Fleets leveraging this technology are reporting accident reductions of up to 50% or more. This is just the beginning. As more fleets come online, the collective data will feed the AI models, making them progressively smarter, more predictive, and more effective at preventing incidents before they happen. AI and edge computing might be the most significant leap in road safety since the seatbelt. By predicting risk, they provide an essential safety net. But their true legacy will be in changing behavior and building trust.
This technology creates a new ecosystem where fairness and quality are valued. It also protects the drivers from false claims, rewarding them for their skill, and ensuring a smoother, safer ride for passengers. Ultimately, this superior driver’s performance translates directly into superior fleet performance: greater reliability, on-time deliveries, and a higher, more dependable level of service for end customers.
India is at an inflection point, and the coming years will be a watershed for safety technology in heavy vehicles. The numbers demand urgency, but the opportunity is bigger than compliance. AI at the edge has the potential to become the most meaningful leap in road safety since the seatbelt, less because of any single feature and more because it changes behaviour in the moment.
Durgadutt Nedungadi is Senior Vice President of EMEA & APAC Business. Views expressed are the author’s personal.