Knowledge at Wharton Where AI Works: Start Small, Think Big – Making AI Stick
Nov 10, 2025
Vivian Sun, Senior Director for Data and AI at Jabil, leads the company's initiatives in computer vision and machine learning. In this engaging discussion, she delves into Jabil's strategic use of AI, starting with computer vision for quality control. Vivian highlights the transformation of manufacturing processes through AI-enhanced inspections, reducing tedious tasks, and improving overall accuracy. She shares insights on fostering AI adoption by securing small wins, emphasizing the importance of scaling and managing change effectively in organizations.
AI Snips
Chapters
Transcript
Episode notes
Pick Deterministic Use Cases First
- Jabil chose computer vision as an entry point because manufacturing outcomes are deterministic and results are easy to validate.
- Deterministic tasks make it easier to demonstrate quick wins and build organizational trust in AI.
Computer Vision Replaced Repetitive Inspections
- Jabil replaced repetitive human visual inspections with AI computer vision on the factory floor.
- This freed inspectors for higher-value tasks and improved inspection quality while easing HR hiring pressure.
Machine Learning Captured Tacit Color Knowledge
- Jabil built a machine learning system to predict color adjustments by instrumenting dye solutions with sensors and analyzing weather effects.
- The project took two years and highlighted that AI often assists expert workers rather than fully replacing them.