Robotaxi Leader Reduces Fleet Downtime and Support Costs

Scaling high-efficiency operations to eliminate stop-and-stall bottlenecks in autonomous mobility

RESULTS

50%
Reduction in labor costs with a decentralized workforce
25 minutes
Met this SLA goal for
vehicle recovery
24/7
Coverage for field operations

Minor issues like open doors stall robotaxis and drive up operating costs. A local, on-demand workforce prevents service delays and maintains fleet uptime.

The challenge
Optimizing real-time roadside assistance

Robotaxis are programmed for safety. When a system detects even a minor anomaly, such as a passenger leaving a door slightly ajar, its fail safe mechanism triggers an immediate stop. 

For one autonomous vehicle (AV) leader, nearly 600 monthly open-door incidents repeatedly halted a portion of its operations. Vehicles could not resume service until the door was closed, forcing the company to divert specialized roadside assistance (RA) agents — originally hired for complex technical malfunctions.

Using highly-trained technical experts for simple manual labor was inefficient and costly, while the operational bottleneck inhibited effective scale.

The solution
A high-recovery engine powered by gig workers

To eliminate downtime and address the growing backlog, the company partnered with TaskUs. We implemented a human-in-the-loop model powered by our crowdsourcing platform TaskVerse.

We vetted a localized gig workforce in the specific areas where the AV company operates. They completed non-disclosure agreements and mandatory safety certifications to ensure they met rigorous company standards. 

A retainer model replaced manual scheduling, allowing workers to claim 4-hour or 8-hour blocks to ensure predictable coverage in high-volume urban windows. 

To minimize the mean time to recover (MTTR), we implemented an automated triggering system that instantly detected vehicle anomalies. The system also sent car details, precise locations and safety protocols directly to the assigned worker’s phone.

A closed-loop verification process ensured that every vehicle was safe before it returned to the active fleet. To guarantee the vehicle was secure before returning to service, workers were required to submit digital proof of work, including photos and standardized workflow questionnaires, verified by geo-location tracking. 

Once validated, the vehicle was automatically tagged “active” in the fleet system and immediately placed back into service.

Results

Shifting field operations to this variable-cost model allowed the client to cut labor costs per task by 50% while maintaining high performance standards. The approach also proved the effectiveness of the decentralized workforce by consistently meeting a 25-minute service level agreement (SLA) for vehicle recovery.

Based on this success, the program was expanded to achieve 24/7 readiness and serve three additional locations.

Connect with a TaskUs Expert