SynCab, a Hong Kong taxi fleet, has partnered with the University of Hong Kong to launch the city’s first AI-powered taxi dispatch system.

The system, developed by SynCab and the university’s Smart Mobility Lab, uses real-time data such as traffic conditions, weather, and electric cab battery levels to assign rides.

SynCab, which operates 160 vehicles, said the system had a soft launch in January, with a full rollout planned for July 14.

Drivers who use the system receive ride assignments without the option to reject them, which the company said could help shorten match times and improve order acceptance.

SynCab plans to expand the system to up to 425 taxis.

Professor Ke Jintao from the university said the AI uses deep learning to estimate remaining battery power and optimize order allocation.

Five premium taxi fleets currently operate in Hong Kong, but a recent test by South China Morning Post found SynCab was the only fleet to confirm a ride within 30 minutes, though its fares were higher than traditional taxis and Uber.

SynCab plans to release a new app version with improved payment tracking features.

🔗 Source: South China Morning Post

🧠 Food for thought

1️⃣ AI dispatch systems show proven results when implemented properly

SynCab’s success reflects a broader pattern of AI dispatch systems delivering measurable improvements in taxi operations worldwide.

A collaboration between Toyota, JapanTaxi, KDDI, and Accenture in Tokyo achieved 94.1% accuracy in predicting taxi demand using real-time data including weather, traffic, and public transport availability1.

Drivers using that AI system saw an average sales increase of 20.4% in February 2018, demonstrating that algorithmic dispatch can significantly boost driver earnings1.

The Tokyo system operates on a 500-meter grid with demand predictions updated every 30 minutes, similar to how SynCab processes real-time data including battery levels of electric cabs and traffic conditions2.

SynCab’s July results, where orders placed and taken both doubled compared to June, align with this global pattern of AI systems improving operational efficiency once fully deployed.