Role
UX Researcher
Timeline
- ~7 weeks
Team
- Product Managers
- Designers
- Developer Engineers
Method
Survey Analysis
Interviews
Usability Testing
Pilot Evaluation
UX Researcher
Survey Analysis
Interviews
Usability Testing
Pilot Evaluation
The Challenge
Amazon delivery drivers had only 20 minutes to complete vehicle inspections, load packages, and begin routes.
Research revealed that the manual vehicle inspection process consumed nearly half of that time, creating operational delays and frustration for drivers and station teams.
Goal
Identify the biggest source of friction within the delivery load-out process and determine whether vehicle inspection automation could improve driver efficiency and station compliance.
Delivery drivers were struggling to complete required vehicle inspections, package loading, and departure activities within a 20-minute operational window.
Initial stakeholder assumptions focused broadly on improving the Under-the-Roof (UTR) process, but it was unclear which aspects of the workflow were creating the greatest friction. Research was conducted to identify the primary barriers to efficiency and uncover opportunities to improve the driver experience while supporting operational performance.
How might we reduce the time and effort required to complete daily vehicle inspections while maintaining compliance and safety requirements?
Research Plan
To understand the root causes of inefficiencies in the driver load-out process, I used a mixed-methods research approach that combined existing feedback analysis, surveys, interviews, and usability testing. The research was conducted in four phases:
Research Methods
Feedback Analysis • Surveys • Semi-Structured Interviews • Usability Testing • Pilot Evaluation
I synthesized findings using thematic analysis, and digital affinity mapping to identify patterns across feedback data, surveys, and interviews. By triangulating insights across multiple research methods and stakeholder groups, I validated that vehicle inspections not the broader UTR workflow were the primary source of friction. This process increased confidence in the findings and helped the team focus on an automation-focused solution.
Based on the research findings, I partnered with product, design, and engineering stakeholders to explore opportunities for reducing friction within the vehicle inspection workflow. The research identified manual inspections as the primary bottleneck within the broader UTR process, leading the team to pursue a modular, AI-powered vehicle inspection concept capable of automatically detecting vehicle damage and syncing findings directly to fleet management systems
By automating inspection tasks, the concept aimed to reduce inspection time, improve the driver experience, and support station compliance goals
Pilot Evaluation
I conducted a 5-station pilot with 40 delivery drivers to evaluate the usability and operational impact of an automated vehicle inspection experience over a three-week period. The pilot combined task-based usability testing, post-pilot surveys, and operational performance metrics to validate both the user experience and business impact of the solution.
- Damage verification controls to allow drivers to review and confirm detected damages.
- Removal of redundant VIN scanning steps to further streamline the workflow.
- Clear inspection completion indicators to improve user confidence and reduce uncertainty.
This demonstrated that reducing inspection friction produced measurable operational improvements.