AI-supported Bridge Inspection Tool

Role Lead UXUI Designer
Year 2025- (Ongoing)
Key words design

Our infrastructure is ageing. Assets built with concrete and metal are reaching the end of their life cycles, entering a phase of costly interventions that put immense pressure on government budgets. Research indicates that maintaining structures proactively, rather than reactively, can save 30% over the asset life cycle.

To address these challenges, we built a cutting-edge AI-driven bridge inspection and reporting solution.

The platform offers:

App-based on-site digital inspection capture, reducing reliance on manual methods like sketching and note-taking.
AI-driven defect quantification, improving accuracy and efficiency.
360° camera integration, ensuring comprehensive site data capture.
A virtual inspection environment, enabling engineers to review structures remotely instead of relying solely on static PDFs.

Impact:
The application received the Insurance Asia Awards 2025 for Digital Transformation Initiative of the Year, recognizing both technical achievement and user-centered design approach.

My Role

Ethnographic Research and User Discovery: Conducted ten days of ethnographic fieldwork across local government offices in Japan, observing infrastructure maintenance workers' workflows and pain points.

Interface Design: Designed the UI/UX for a bridge maintenance application that coordinates user workflows with AI-powered defect detection systems. Through iterative prototyping with maintenance workers, I translated complex technical capabilities into intuitive interfaces that fit seamlessly into their daily operations.

Cross-Functional Leadership: Bridged communication between engineering teams developing AI detection models and end-users in local government, ensuring technical innovation aligned with real organizational needs and workflows.

Interview and Ethnographic Research

To understand current workflows and pain points, I conducted ten days of ethnographic fieldwork across local government offices in Japan.

Pain Points

Data Standarization

Personnel changes every 5 years
Different contractors handle different bridges -> Crack measurement methods vary by contractor
Without standardised data, it's difficult to track damage progression over time

Photo-damage Linkage

Manual paper-based system creates disconnect between field observations and photo documentation
Significant time wasted on administrative tasks after fieldwork
Risk of mismatching photos with damage descriptions

Inspection Challenges

Inspectors (esp. beginners) struggle to provide comprehensive analysis beyond basic damage description
Incomplete or inaccurate root cause analysis leads to inappropriate maintenance decisions
Cannot effectively prioritize repairs without proper risk assessment

Solution

Database

Create a standardised data collection framework that remains consistent across personnel changes.

On-site application

Allow users to mark defects and immediately upload/associate photos with specific damage points

AI-powered Inspection

Offer AI-powered suggestions for inspection based on similar historical bridge data

User Flow

Based on these insights, I designed AS-IS and TO-BE user flows, mapping workers' current maintenance processes and envisioning how AI-powered capabilities could streamline their workflows while maintaining familiar interaction patterns.

Wireframes & Hifi

I designed wireframes and high-fidelity prototypes for the application, translating research insights into intuitive interfaces that coordinate user workflows with AI-powered defect detection systems.

Award

Our tool is being tested with several local governments, and was recently honored with the Insurance Asia Awards 2025 for Digital Transformation Initiative of the Year.

Credits

Takeharu Harada: Project Manager
Yui Kondo: Lead Designer
Bernardo Perez Orozco: ML lead
Tim de Rooij: ML engineer
Go Maehata: Sales