Track and Rail Health Platform by Alara Hakki

Track and Rail Health Platform

Alara Hakki

User Researcher

Product Designer

Product Strategist

Figma

Miro

Kawasaki Japan needed to modernize its rail inspection process, replacing manual methods with an AI/ML-powered solution that could process real-time data from IoT sensors. The goal was to enable predictive maintenance, reducing human error and improving inspection efficiency.

The Challenge

Kawasaki’s existing rail inspection process relied heavily on manual methods, which were inefficient and prone to errors.

The goal was to create an AI-powered platform that integrated IoT sensors to streamline rail inspection and enable predictive maintenance.

__________________________

Design Decisions and Key Insights:

Field interviews revealed that rail inspectors were overwhelmed by large data sets generated by IoT devices. We needed to prioritize critical information while allowing deeper exploration when necessary.

I led the design of a user-friendly dashboard, using wireframes and prototypes to test how inspectors could access real-time data and alerts. The final design used color-coded alerts and customizable data views, ensuring ease of use in field conditions.

The simplified interface focused on quick access to high-priority data with the ability to drill down when necessary. Mobile compatibility was also crucial to support inspectors working in the field.

__________________________

Discoveries

The inspection process is inefficient and labor-intensive.

Managing inspection data is inconsistent and prone to errors.

Maintenance planning is fragmented, leading to missed opportunities for efficiency.

__________________________

Results & Impact

Outcome: The platform improved inspection accuracy and reduced manual work time by 40%. Predictive maintenance features led to fewer service interruptions.

Impact: Rail downtime decreased by 25%, and inspectors found the system intuitive, with 85% reporting that it significantly improved their ability to detect and address track issues.

Like this project