Initiative details
Road accidents kill 1.35 million people globally every year, with 94% caused by human factors — distraction and fatigue being the leading contributors. Professional drivers, particularly in ride-hailing and commercial fleets, face heightened risk due to long shifts, irregular hours, and constant exposure to urban traffic conditions. Existing driver monitoring technology relies on dedicated hardware — infrared cameras, sensors, integrated systems — costing hundreds or thousands of euros per vehicle. This makes meaningful deployment economically unviable at scale, particularly across the diverse, large fleets operated by ride-hailing platforms in emerging and developing markets, where road fatality rates are highest but resources are most constrained.
WingDriver addresses this gap directly: distraction and fatigue are the largest preventable causes of crashes, yet no scalable, affordable technology existed to detect and intervene against them in real time, on the vehicles and drivers who need it most. Our project with inDrive — one of the world's largest ride-hailing platforms, with 240 million downloads across 48 countries — set out to prove that AI-powered driver monitoring could be deployed entirely through drivers' existing smartphones, eliminating the hardware barrier that has limited adoption of this life-saving technology for over a decade.
WingDriver addresses this gap directly: distraction and fatigue are the largest preventable causes of crashes, yet no scalable, affordable technology existed to detect and intervene against them in real time, on the vehicles and drivers who need it most. Our project with inDrive — one of the world's largest ride-hailing platforms, with 240 million downloads across 48 countries — set out to prove that AI-powered driver monitoring could be deployed entirely through drivers' existing smartphones, eliminating the hardware barrier that has limited adoption of this life-saving technology for over a decade.
Initiative date
to
Who was/is your target audience?
Car drivers
Topic
Training
Organisation details
WingDriver
Enterprise
Portugal
Porto
Contact name
André Gonçalo Azevedo
Telephone number
914253574
andre@wingdriver.com
Website link
Project activities
If you work together with external partners, list the most important partners and briefly describe their role.
inDrive — One of the world's largest ride-hailing and mobility platforms, with over 240 million downloads across 48 countries. inDrive provided the driver fleet, operational coordination, and platform integration necessary to deploy WingDriver at scale, and co-published the pilot's results publicly, reflecting their commitment to driver safety as a strategic priority.
Please describe the project activities you carried/are carrying out and the time period over which these were implemented.
WingDriver deployed its AI-powered driver safety platform across inDrive's commercial driver fleet in Kazakhstan over a six-month pilot period, running from December 2025 to May 2026.
The platform operates entirely on the driver's own smartphone, requiring no additional hardware or vehicle modification. Using on-device edge AI and computer vision, WingDriver continuously analyses the driver's facial state and gaze patterns to detect early signs of fatigue, microsleep events, and visual or cognitive distraction. When risk indicators are detected, the system delivers real-time audio and visual alerts directly to the driver, prompting immediate corrective action before a dangerous situation can escalate into a collision.
Throughout the pilot, drivers were onboarded with simple setup guidance, and the system began generating behavioural data from their very first trips. inDrive's operations team monitored fleet-wide safety metrics through WingDriver's cloud dashboard, while individual drivers received personalised feedback and gamified incentives — including a coin-based reward system — to reinforce safer driving habits over time. The pilot scaled from an initial cohort to over 1,000 active drivers, generating more than 100,000 kilometres of monitored driving in March 2026 alone, with results validated through structured data analysis and direct driver feedback surveys.
The platform operates entirely on the driver's own smartphone, requiring no additional hardware or vehicle modification. Using on-device edge AI and computer vision, WingDriver continuously analyses the driver's facial state and gaze patterns to detect early signs of fatigue, microsleep events, and visual or cognitive distraction. When risk indicators are detected, the system delivers real-time audio and visual alerts directly to the driver, prompting immediate corrective action before a dangerous situation can escalate into a collision.
Throughout the pilot, drivers were onboarded with simple setup guidance, and the system began generating behavioural data from their very first trips. inDrive's operations team monitored fleet-wide safety metrics through WingDriver's cloud dashboard, while individual drivers received personalised feedback and gamified incentives — including a coin-based reward system — to reinforce safer driving habits over time. The pilot scaled from an initial cohort to over 1,000 active drivers, generating more than 100,000 kilometres of monitored driving in March 2026 alone, with results validated through structured data analysis and direct driver feedback surveys.
In terms of implementation, what worked well and what challenges did you need to overcome?
The smartphone-based, hardware-free model proved to be the project's greatest strength. Deployment required no vehicle downtime, no procurement process, and no technician installation — drivers simply downloaded the app and began driving. This eliminated the single largest barrier that has historically limited driver monitoring technology adoption at scale.
Driver acceptance was a notable success: rather than perceiving the technology as surveillance, drivers responded positively to its framing as a personal safety companion. Gamification elements — including safety scores, coin rewards, and leaderboards — were instrumental in encouraging voluntary, sustained engagement rather than passive compliance.
The principal challenge was building driver trust in a technology that processes facial data. This was addressed through a privacy-first architecture: all video processing occurs entirely on-device, with no raw footage transmitted or stored, and only anonymised behavioural metadata shared with fleet operators. Communicating this clearly to drivers, in their own language and context, was essential to adoption.
A second challenge was calibrating detection accuracy across diverse lighting conditions, vehicle interiors, and driving environments specific to Kazakhstan's geography and climate — addressed through iterative on-device model tuning during the pilot's early weeks.
Driver acceptance was a notable success: rather than perceiving the technology as surveillance, drivers responded positively to its framing as a personal safety companion. Gamification elements — including safety scores, coin rewards, and leaderboards — were instrumental in encouraging voluntary, sustained engagement rather than passive compliance.
The principal challenge was building driver trust in a technology that processes facial data. This was addressed through a privacy-first architecture: all video processing occurs entirely on-device, with no raw footage transmitted or stored, and only anonymised behavioural metadata shared with fleet operators. Communicating this clearly to drivers, in their own language and context, was essential to adoption.
A second challenge was calibrating detection accuracy across diverse lighting conditions, vehicle interiors, and driving environments specific to Kazakhstan's geography and climate — addressed through iterative on-device model tuning during the pilot's early weeks.
Evaluation
Please summarise how you have evaluated the initiative’s impact (e.g. social media reach, survey, feedback forms, statistics).
The pilot's impact was evaluated through three complementary methods. First, quantitative behavioural data was collected continuously throughout the pilot via the WingDriver platform itself, measuring the frequency of distraction events, dangerous driving behaviours, and cellphone-related incidents per driver and per kilometre driven, compared against baseline measurements taken during an initial monitoring-only phase.
Second, structured driver feedback was gathered through in-app surveys following the pilot period, capturing drivers' subjective perception of the platform's safety value and any behavioural changes they personally observed.
Third, aggregate fleet-level statistics were analysed jointly with inDrive's operations team to assess overall trends across the full driver cohort, validating that improvements observed were consistent and sustained rather than isolated or short-term.
This combined approach — objective behavioural data alongside direct driver testimony — allowed the evaluation to capture both the measurable safety impact and the qualitative experience of drivers using the technology daily, providing a comprehensive picture of the intervention's real-world effectiveness across a genuinely large-scale, commercial deployment rather than a small controlled study.
Second, structured driver feedback was gathered through in-app surveys following the pilot period, capturing drivers' subjective perception of the platform's safety value and any behavioural changes they personally observed.
Third, aggregate fleet-level statistics were analysed jointly with inDrive's operations team to assess overall trends across the full driver cohort, validating that improvements observed were consistent and sustained rather than isolated or short-term.
This combined approach — objective behavioural data alongside direct driver testimony — allowed the evaluation to capture both the measurable safety impact and the qualitative experience of drivers using the technology daily, providing a comprehensive picture of the intervention's real-world effectiveness across a genuinely large-scale, commercial deployment rather than a small controlled study.
What has been the effect of the activities?
The results were significant and consistent across the driver cohort. WingDriver delivered a 71% reduction in overall driver distraction events and a 79% reduction in dangerous driving behaviours over the course of the pilot. Cellphone-related distraction specifically fell by 30%.
Beyond the quantitative metrics, driver feedback confirmed the platform's real-world value: 77% of drivers reported that WingDriver had highly contributed to their safety, and 33% stated that the platform's alerts had directly helped them avoid what they believed could have been an accident.
One driver's account illustrates this impact directly: after an exhausting eight-hour shift, she nearly fell asleep at the wheel just 500 metres from home — WingDriver's fatigue alert woke her in time to prevent a potential accident.
Over 1,000 professional drivers were actively monitored during the pilot, generating more than 100,000 kilometres of safer driving in March 2026 alone. Based on these behavioural improvements, we estimate the pilot contributed to preventing approximately 330 potential road incidents across the driver cohort — a substantial safety outcome achieved entirely through software, without any hardware investment by the platform or its drivers.
Beyond the quantitative metrics, driver feedback confirmed the platform's real-world value: 77% of drivers reported that WingDriver had highly contributed to their safety, and 33% stated that the platform's alerts had directly helped them avoid what they believed could have been an accident.
One driver's account illustrates this impact directly: after an exhausting eight-hour shift, she nearly fell asleep at the wheel just 500 metres from home — WingDriver's fatigue alert woke her in time to prevent a potential accident.
Over 1,000 professional drivers were actively monitored during the pilot, generating more than 100,000 kilometres of safer driving in March 2026 alone. Based on these behavioural improvements, we estimate the pilot contributed to preventing approximately 330 potential road incidents across the driver cohort — a substantial safety outcome achieved entirely through software, without any hardware investment by the platform or its drivers.
Please briefly explain why your initiative is a good example of improving road safety.
The pilot's impact was felt at multiple levels simultaneously. At the local and national level, it directly improved safety outcomes for over 1,000 professional drivers operating across Kazakhstan, a market with significant road safety challenges and limited prior access to advanced driver safety technology.
At the international level, the pilot's success — validated through inDrive's public co-published results — has established a replicable model now being prepared for expansion to a 10,000-driver rollout within Kazakhstan and active deployment planning across Brazil, Asia, and Latin America. Because inDrive operates in 48 countries, the validated pilot model creates a direct pathway to scaling driver safety improvements across multiple national road safety contexts simultaneously.
This initiative demonstrates that meaningful, measurable road safety improvement does not require expensive infrastructure or hardware investment — it can be achieved through software deployed on devices people already own. By eliminating the cost and complexity barriers that have historically restricted driver monitoring technology to wealthy fleets and developed markets, WingDriver's approach with inDrive proves that life-saving safety technology can be made accessible to any commercial driver fleet, anywhere in the world.
The model is straightforward for other organisations to replicate: any ride-hailing platform, fleet operator, or transport authority can deploy a smartphone-based driver monitoring solution without procurement cycles, hardware investment, or lengthy installation processes. The privacy-first, on-device processing architecture also provides a template for addressing the data protection concerns that often delay safety technology adoption. Because the approach integrates with drivers' existing devices and daily routines, it can be scaled rapidly across diverse vehicle fleets, road conditions, and regulatory environments — making it one of the most accessible and immediately implementable road safety interventions currently available to transport operators globally.
At the international level, the pilot's success — validated through inDrive's public co-published results — has established a replicable model now being prepared for expansion to a 10,000-driver rollout within Kazakhstan and active deployment planning across Brazil, Asia, and Latin America. Because inDrive operates in 48 countries, the validated pilot model creates a direct pathway to scaling driver safety improvements across multiple national road safety contexts simultaneously.
This initiative demonstrates that meaningful, measurable road safety improvement does not require expensive infrastructure or hardware investment — it can be achieved through software deployed on devices people already own. By eliminating the cost and complexity barriers that have historically restricted driver monitoring technology to wealthy fleets and developed markets, WingDriver's approach with inDrive proves that life-saving safety technology can be made accessible to any commercial driver fleet, anywhere in the world.
The model is straightforward for other organisations to replicate: any ride-hailing platform, fleet operator, or transport authority can deploy a smartphone-based driver monitoring solution without procurement cycles, hardware investment, or lengthy installation processes. The privacy-first, on-device processing architecture also provides a template for addressing the data protection concerns that often delay safety technology adoption. Because the approach integrates with drivers' existing devices and daily routines, it can be scaled rapidly across diverse vehicle fleets, road conditions, and regulatory environments — making it one of the most accessible and immediately implementable road safety interventions currently available to transport operators globally.
How have you shared information about your project and its results?
Results from the pilot were shared through a joint press release co-published with inDrive, reaching media outlets across the technology, mobility, and insurance sectors. The findings have also been presented to academic and industry audiences, including road safety researchers at the AAA Foundation for Traffic Safety and the Insurance Institute for Highway Safety. WingDriver further shares ongoing learnings through its dedicated road safety podcast, "Get Home Safely," featuring conversations with road safety advocates, researchers, and mobility industry leaders, and through direct engagement with insurance, fleet, and ride-hailing industry partners evaluating similar deployments.