Ford Gives Line Workers AI-Assisted Inspection
Quick Summary
Ford runs an AI vision system called MAIVS at 17 North American plants that photographs vehicles as they move down the line and returns a pass or fail on wiring, connectors and panel fit in seconds. It has performed 150 million inspections and flagged 400,000 issues a person might have missed.
The Use Case
Human inspectors miss subtle defects, especially in areas like wiring routes and electrical connections that are hard to see and easy to overlook late in a shift. Ford and IBM built a system that runs on ordinary iPhones and IP cameras at operator workstations. The phone photographs the specified area, a trained model on the phone judges it, and the result comes back immediately. Defects are logged to Ford's quality system and fixed on the spot, so the vehicle never leaves the station with a known problem. Ford is adding 40 more inspection systems in 2026.
Real Example
- A vehicle reaches the workstation and the system reads its VIN to know which build variant to expect
- The operator points the phone at the wiring harness connector area
- The model on the phone compares the image to what a correct assembly looks like for that variant
- A red fail appears: one connector is not fully seated
- The defect logs to the Quality Operating System and the operator fixes it in the station
- The in-station process coach clears the vehicle and the line moves on
Why It Matters
- 150 million inspections completed, with 400,000 issues flagged that inspection alone might have missed
- Runs on phones and standard cameras, not specialized hardware, across 1,000+ workstations
- Defects are fixed in the station, before they become warranty claims