Mobility / Luxury Car Rental
Fengyue Rent-a-Car — Full-Stack AI Operations Platform for China's Largest Luxury Fleet
The Challenge
At 5,000+ luxury vehicles across 50+ cities and 3,000+ outlets, Fengyue Rent-a-Car faces the revenue optimization problem that every large fleet operator knows intimately: manual dispatching cannot process enough variables fast enough to minimize idle vehicles, match supply to demand peaks, or price dynamically. Every hour of suboptimal allocation is measurable revenue loss. NoBounds built a unified AI cockpit platform with a real-time data lake at its foundation: rental orders, vehicle GPS positions, customer credit profiles, and historical incident records all ingest at zero latency via a standardized log collection agent deployed across all outlets. A time-series forecasting engine generates 7–14 day rolling demand predictions per vehicle model and region. These predictions feed directly into a reinforcement learning dispatch engine that writes optimal scheduling recommendations and dynamic pricing back to each outlet's staffing system via REST API — enabling human-AI collaborative dispatch at enterprise scale.
Our Approach
On the risk side, NoBounds built a binary classification model trained on 300,000 labeled records — combining user identity data, credit bureau signals, order history, and historical incident flags. The model returns a real-time risk probability score per incoming order; high-risk cases automatically trigger SNS/Slack alerts and route to an operations review queue, achieving >95% high-risk order interception with minimal false-positive friction on legitimate customers. The management panorama screen unifies historical, live, and predicted data in a single view: pickup/return heat maps, fleet utilization rates, and revenue forecasts rendered at a glance, with a one-click dispatch recommendation that compresses management decision time from hours to under 10 seconds. For fleet operators, logistics companies, or any asset-intensive business where demand volatility and risk management are daily operational realities, this platform represents the AI operations layer that makes scale manageable.