Microsoft Guides
Simple step-by-step instructions that help guide employees to the tools and parts they need teaching them how to use them in real work situations. Photo Work Validation is a feature that gives the ability to check if procedures were completed correctly by operators combining mixed reality and machine learning design.
-
1 Year Project
-
Lead Designer / User Experience & Interface
-
Engineers (2) / Data Science (2) / Project Manager
Process
-
How might we measure if a Guide is completed safely, efficiently, or to quality by an operator?
-
A checkpoint step within a guide that allows an operator to capture a photo for an AI model to validate as correct or incorrect.
-
Reduced 13 days to 2 days of training - Toyota
Reduced 7 weeks to 9 hours of training - Ford
Saw a 40% reduction In errors - Thermo Fisher
Data collection experiments to measure the accuracy of the AI model in different scenarios including lighting, angle, and distance.
Cross functional workshops to understand the end-to-end customer journey and major touchpoints.
Design iterations to explore how this new photo work validation feature can fit into the existing design system we had established.
Early design exploration on how to create an effective guiding mechanism for an operator to take a photo in the same position every time.
Competitive and comparative audits to understand the photo validation space and gather insights from existing solutions.
The design system contributed to and used for Microsoft Guides including Photo Work Validation.
