Ford Over-Automated, Rehired 350 Engineers, Won Quality

Ford has risen to the top spot in JD Power’s 2026 initial quality survey for mainstream brands. And their journey towards the top is a lesson in why blind faith in artificial intelligence (AI) can lead to disaster. Over the past few years, Ford has increasingly relied on automated tools to make decisions about manufacturing processes and car design. These tools proved far less accurate than anticipated, leading to declines in quality. As a result, Ford had to go scrounging for retired engineers and begging them to return.

As Ford’s VP of vehicle hardware engineering Charles Poon told reporters during a press conference this week, “We mistakenly believed that if we just introduced artificial intelligence and changed our design parameters, we would get a high-quality vehicle.” They didn’t. The automated tools didn’t have the same depth of experience as seasoned engineers. Years of working through different car builds, understanding what pitfalls to avoid, which suppliers to watch out for, which alerts to ignore and which ones to take seriously. And before that expertise could be fully captured and fed into the AI models, many of the engineers who knew those things left the company.

1350 engineers, and what they actually did

So Ford started bringing back, promoting, or hiring more than 350 experienced engineers to fill the void. Some of them helped train younger engineers who were having trouble meeting quality benchmarks. Others focused on refining the data gathering and model training procedures that drive the automated tools. Meanwhile, Ford formed a 40-person software quality assurance team dedicated solely to finding issues before products go into production. This is a departure from the traditional “find and fix” approach, in which defects are identified once they emerge. “We’re moving from a find-and-fix mentality to preventing issues,” said COO Kumar Galhotra.

Ford also introduced over 100,000 new AI-driven tests to help uncover edge cases and push systems to their limits across various scenarios. The automated testing process allows software updates to be quickly validated even in the late stages of development, preventing defects from snowballing into costly recalls. Right now, Ford holds the unenviable title of having the most recalls in history (a title they’re desperately trying to lose). And quality issues have been exacerbated by pandemic-related supply chain disruptions and troubled launches of vehicles like the Explorer and Aviator.

2The lesson that applies well beyond car factories

This is a classic example of how AI performance depends on the quality and context of the input it receives, and how it’s difficult to capture expert knowledge without the experts themselves. Ford’s automated tools made mistakes not because AI is flawed but because they were fed incomplete information and not closely monitored. Rather than giving up on AI, Ford insists that they still believe in automation and have significantly ramped up their use of AI-driven testing. Instead, they brought in experienced humans to supplement the technology, using them as trainers and verifiers rather than substitutes.

This combination of AI for scalability and speed, and human experts for judgment and edge cases, is becoming the winning formula for many high-consequence applications. Unfortunately, Ford had to learn this lesson the hard way, losing some quality points along the way.