Why fleet safety is moving beyond compliance – and into real-time intervention
- 23 hours ago
- 5 min read
By Charles Dawson, CEO, AutoSense Australia
For a long time, fleet safety has been built around compliance.
Logbooks, policies and procedures all have their place. But they’re designed to show rules are being followed, not whether risk is actually being managed in real time. That gap is becoming harder to ignore.
A driver can be fully compliant on paper and still be operating in a state of significant impairment. Poor sleep, disrupted routines, or underlying fatigue can all impact performance well before a shift begins. Yet most traditional safety systems are not designed to detect that.
That’s the challenge the industry is now starting to confront. In Australia alone, the latest data from the Guardian Insights Report by Seeing Machines – distributed in Australia and New Zealand by AutoSense – identified more than 100,000 confirmed fatigue events and over 380,000 distraction events across monitored fleets. Most of these don’t result in a reportable incident. They happen quietly, in the background, and often go unseen.
The reality is that compliance has always been a proxy for safety. It gives organisations a framework, but it doesn’t provide visibility of what is actually going on behind the wheel, in the moments that matter most.
The limits of hindsight
For years, safety technology has played an important role in helping fleets understand risk. Dashcams, telematics and in-cab monitoring systems have steadily improved visibility, giving operators a clearer picture of what’s happening on the road and where risks emerge. That progress has been significant.
But much of that visibility has still been retrospective in nature. Even when events are detected in-cab, the focus has often been on clear, observable indicators, such as prolonged eye closure or sustained distraction.
By the time those signals appear, the risk has already escalated. The next challenge has been detecting impairment earlier, at the point where a driver may still appear alert but is no longer fully processing what’s happening around them.
That’s not a simple problem to solve. Fatigue develops gradually, and drivers are not reliable judges of their own state. There’s no clear moment where someone recognises they are no longer safe to drive. Which means that relying on self-assessment, or compliance alone, has always left a gap.
From detection to intervention
As more providers enter the market, the distinction between generic camera-based systems and purpose-built fleet safety technology is becoming increasingly important, particularly in complex fleet environments.
Next-generation systems such as Guardian Generation 3 are redefining what’s possible. They can identify early signs of cognitive impairment before a driver reaches the point of microsleep. Some of the highest-risk moments happen when a driver’s eyes are open, so detecting that earlier is a meaningful shift.
Rather than relying on basic indicators like eye closure or head position, these systems use advanced gaze tracking and machine learning to understand where a driver’s attention is actually directed. Drivers can be looking straight ahead and still be distracted, a situation which can now be detected.
This level of precision is already uncovering risks that were previously invisible. In a single month, AutoSense recorded approximately 9,000 traditional distraction events across its ~6,300 units in Australia and New Zealand; however, within a small subset of just ~380 Gen 3-enabled vehicles, over 21,000 attention-sharing events were detected in the same period — more than twice as many despite accounting for only around 6% of installations. The majority of this behaviour would not have been captured by earlier systems.
At the same time, early impairment detection allows fleet managers to intervene sooner. Instead of waiting for a driver to fall asleep, the system identifies declining alertness and intervenes immediately.
That intervention happens in real time – through audio, visual and physical feedback, including seat vibration – giving the driver a chance to course-correct before a lapse becomes an incident.
Underpinning all of this is scale. Specialised systems like Guardian by Seeing Machines have been trained over more than 30 billion kilometres of real-world driving, continuously refining how risk is identified.
Changing driver behaviour in real time
One of the most significant benefits of real-time intervention is its impact on behaviour. When drivers receive immediate feedback, they become more aware of their own state and how it changes over time. Over time, this creates a feedback loop. Drivers begin to recognise early signs of fatigue or distraction and take action sooner, rather than pushing through.
Independent research commissioned by the National Heavy Vehicle Regulator — spanning multiple fleets and technologies across Australia — found that drivers using real-time fatigue monitoring systems began recognising their own early warning signs of fatigue, which they had previously overlooked or ignored. Over time, many reported acting on these cues proactively, taking breaks before any alert was triggered.[1]
This is where the real value sits. Not in the alert itself, but in what it prevents.
Beyond compliance, towards accountability
This shift also has implications at an organisational level. Australia’s Chain of Responsibility framework has already expanded accountability across the supply chain. Increasingly, it’s not enough for organisations to show that systems are in place; they need to demonstrate that those systems are effective in managing real-world risk.
One of the biggest challenges for leadership teams is the gap between what is assumed to be happening operationally, and what’s actually happening on the ground.
Real-time data begins to close that gap. It provides visibility not just of incidents, but of near misses, early warning signs, and patterns of behaviour that would otherwise go unnoticed.
Research reinforces the impact of this visibility in action. A 2017 study[2] found that fatigue events were reduced by 66% when drivers received real-time in-cab alerts, with incidents becoming shorter and occurring later in trips—clear signs of improved driver responsiveness.
In practice, the impact is even greater when real-time insight is paired with human intervention. Models like the Guardian Centre, where fatigue events are verified in real time and fleet managers are immediately notified, enable rapid follow-up with drivers. In these environments, reductions in fatigue events have been shown to reach as high as 95%, demonstrating how coordinated response drives materially better safety outcomes.
Where this is heading
We are now moving towards a model where safety is not a periodic review process, but a continuous one.
Advances in monitoring, data and machine learning are making it possible to detect risk earlier and intervene faster, building a far more accurate picture of what’s happening across a fleet.
In practical terms, that means fewer surprises, better-informed decisions, and ultimately safer outcomes for drivers and other road users.
Compliance will always have a role to play. But it is no longer enough on its own. The organisations making real progress are the ones treating safety as something dynamic – something that can be measured, understood and influenced in real time.
Comments