How Ford flags assembly line defects with IBM's AI-powered inspection tools

Title

How Ford flags assembly line defects with IBM's AI-powered inspection tools

Content

Ford Motor Co. has installed IBM's AI-powered Maximo Visual Inspection technology at 17 of its North American manufacturing facilities to catch assembly-line defects in real time, using pictures taken of vehicle components at various stages of production via iPhone and a cloud-based platform. Ford started its partnership with IBM in 2020. The computer-vision-based inspection technology is helping the automaker reduce warranty claims by identifying and fixing assembly-line defects and avoiding higher rework labor and repair costs to correct issues after a vehicle is fully built. During a presentation at the 2026 AutoTech conference in Novi, Michigan, Jason Barger, Ford's advanced manufacturing IT vision systems and cyber security product manager, and Ed Neubecker, principal automation specialist at IBM, gave a more detailed look at how the system works and how errors can be quickly corrected before they spread to other vehicles further down the line. Ford refers to the inspection system internally as its Mobile AI Vision System, or MAIVS. It has been used to perform 150 million individual inspections at the automaker's manufacturing facilities, flagging 400,000 quality issues that a human worker may have missed. "We needed to give our operators superpowers to see the unseen in real time, and that's where MAIVS comes into play," said Barger. "MAIVS is like giving every operator a tireless set of trained eyes to see every small minute detail on the line. It's able to catch subtle defects, such as a missed component, misaligned assembly, or an improper part that's been installed." As a vehicle moves along the assembly line, an operator workstation near the production lines and equipped with an iPhone for capturing photos first identifies the VIN and vehicle build complexity. The inspection workstation prepares the device to take a picture of a particular area of interest, which could be anything on a vehicle from door panel components to electrical connections and wiring routing. The phones can automatically capture images as vehicles move on the line, but pictures can also be taken using visual triggers. Ford collects both good and bad images to train machine learning models, which are deployed to iPhones where the system runs inference on the image and sends pass/fail results in real time. If an unseated electrical connector is flagged, the concern is logged into Ford's Quality Operating System for corrective action before an issue propagates to other vehicles on the line. If applicable, an operator fixes the defect on the spot and an in-station process coach closes the concern and clears the vehicle for release; Ford can also stop the production line for immediate correction. Ford says it has over 1,000 operator workstations currently deployed at its North American factories. Barger said Ford installed 40 new inspection systems this year alone at its Michigan facilities. At its Kentucky Truck Plant, Ford added 1,200 inspections, 203 new inspectors, 72 new technology tests, and six times the number of AI-powered inspection tools for the 2026 Expedition SUV compared to last year's model launch. In 2021, Ford recognized IBM with an IT Innovation Award for its AI-powered inspection technology.