Remote visual inspection to identify defects' root causes

Target Audience

Consumer electronics manufacturers, quality control teams, product managers

Challenge

Owlet, a Utah-based baby monitoring device company with approximately 277 employees, faced product failures related to potting material curing defects. With manufacturing overseas in Asia, the company couldn't directly inspect failed units, and traditional root cause analysis was slow and expensive.

Solution Approach

Owlet implemented Instrumental's AI-powered imaging platform with Visual Search capabilities. The system captures images of both failed and new units, using machine learning algorithms to identify anomalies and correlations. The AI identified that potting material wasn't properly cured — validated within just 2 weeks.

Value Add

The implementation delivered $953,000 in annual savings by preventing product replacements, with one-month breakeven on investment. Analysis and improvement cycles became 8X faster, enabling rapid quality resolution and improved customer satisfaction without requiring factory visits.

How to Apply This

Visual quality inspection is a popular use case for factory environments and the manufacturing domain in general. There are various options on how to approach this. In this use case, enabling remote workers to access the relevant information and run inspections was the key. In other scenarios where less deep expert knowledge is necessary (such as detecting scratches on a surface), you can think towards providing custom AI models that are optimized for detecting your recurring quality issues.

Want to explore what computer vision can do for you for quality inspection? Reach out to me and book a free 30-minute feasibility call.