Drivers with higher scores will unlock broader coverage and more benefits.

Zurich Insurance Group Ltd.’s Hong Kong unit and YAS Digital Ltd. are using artificial intelligence (AI) to link insurance coverage directly to real-world driving behavior, which they say could reshape motor insurance for the city’s taxi industry.

The initiative addresses what both companies see as a persistent gap in traditional motor insurance, which relies heavily on static factors such as age, claims history and vehicle type, rather than how drivers actually behave on the road.

In Hong Kong’s dense urban environment, that has limited insurers’ ability to differentiate risk or reward safer driving.

Under the program, Zurich Insurance (Hong Kong) and YAS analyse driving data to generate daily performance scores for taxi drivers, which are reviewed quarterly.

Drivers are then placed into performance tiers, with higher scores unlocking broader personal accident coverage and additional benefits.

“Data is analysed to generate daily scores, which will be reviewed quarterly,” Eric Hui, CEO at Zurich Insurance for Greater China, said in an emailed reply to questions. “Drivers can receive corresponding insurance coverage and rewards based on their quarterly performance score.”

The model is built on YAS’s AI-assisted TAXY platform, which combines in-vehicle monitoring systems, smartphone sensors and real-time analytics. Over a six-month pilot period ending in September, the system processed about 39 million data points and delivered a 20% improvement in safety scores among participating taxi drivers.

Zurich said the early results suggest the approach could modernise underwriting in a segment where accident risk is high and pricing differentiation has historically been limited.

Drivers with stronger safety records are entitled to more comprehensive personal accident protection, including coverage that can extend to family members, according to Zurich.

The tiered structure reflects a broader industry shift toward usage-based insurance, where risk is assessed using actual driving patterns rather than demographic assumptions.

“Behavioural data is used to improve safety, not to punish drivers,” Hui said. “Driving behaviour helps us classify risk tiers so we can reward safer habits with better benefits, but we are mindful of keeping the system fair.”

Beyond insurance coverage, the data generated by the platform offers insights for drivers and fleet operators. The system analyses routes, working patterns and downtime, letting drivers cut wasted hours and improve daily earnings, whilst helping operators optimise fleet use, said William Lee, co-founder of YAS.

As insurers increasingly use behavioural data in pricing and benefits, concerns about algorithmic bias and transparency have grown across the industry. Zurich and YAS said the taxi program addresses those risks.

The scoring models rely on objective indicators such as harsh braking, speeding relative to legal limits, acceleration patterns, turning stability, distracted driving signals, accident black spots and environmental risks along each journey. These indicators are designed to reflect accident risk and are not influenced by personal traits such as age or gender.

To manage fairness, the program includes regular audits of model outcomes, human oversight before any tier changes are implemented, and transparency tools that let drivers understand scores.

Drivers can participate using government-mandated in-car cameras or YAS’s mobile app, which uses smartphone sensors and GPS data. The system integrates weather conditions, traffic patterns and identified danger zones using Hong Kong-specific traffic algorithms, with machine-learning models refined continuously as new data is collected.

Both companies view the taxi initiative as a testing ground for broader applications of telematic-driven underwriting. Hui said the data generated through the program could eventually form part of insurers’ underwriting considerations as usage-based products expand.

“Customers are increasingly looking for a more usage-based insurance cover,” Lee said, adding that the model could be extended to other fleet-based segments, particularly as new regulations shape Hong Kong’s ride-hailing industry.

Future phases may include links to digital commerce platforms and electric vehicle charging networks, aligning the program with longer-term trends in electrification and smart mobility.

For Zurich, the taxi sector carries broader economic significance. “Hong Kong’s taxis are the heartbeat of the city,” Hui said. “Using AI and smart telematics allows us to safeguard livelihoods while contributing to a more sustainable and resilient transport ecosystem.”