Ford Rehires 300 Inspectors After Its AI Quality Bet Missed

Ford has rehired more than 300 veteran quality inspectors after AI driven checks at its factories kept missing defects experienced engineers were brought back to catch. The rehires happened over several years, executives told reporters this week, and form part of a public reversal the company has now tied to its own quality underperformance.

Charles Poon, Ford’s vice president of vehicle hardware engineering, told reporters Ford had overestimated what AI alone could deliver in vehicle assembly. The reversal cleared a path that ends with Ford ranking first among mainstream brands in the J.D. Power 2026 U.S. Initial Quality Study. It is the company’s first such ranking since 2010.

Ford Rehires 300 Inspectors After AI Quality Checks Fall Short

The reversal happened inside a quality system Ford had sold investors on as a productivity breakthrough. Chief operating officer Kumar Galhotra told investors in an earnings call that Ford was “deploying AI across the entire industrial system,” pointing to 900 AI powered cameras in its plants to flag defects at the source. More than 300 veteran quality inspectors have now been brought back to make up for the pitfalls of those automated systems, per Ford’s reversal on its own AI quality checks. Poon told reporters the rehires were needed because the AI quality checks failed to live up to expectations.

The cameras were never supposed to lose the role they had been given. Galhotra told investors the 900 of them would “detect quality issues at the source and help us mitigate supply disruptions.”

Poon’s framing on this week’s call tied the rehires to a single cause. Experienced engineers had been brought back to run mandatory weekly design reviews and rebuild the data sets the AI tools had been trained on. Galhotra said the returning specialists now lead troubleshooting sessions and have overhauled the AI tools to flag defects before parts reach the assembly line.

“Just Ingesting Design Requirements” Was the Bet That Failed

Poon was unusually direct about the bet that did not work. “Artificial intelligence is a fantastic tool,” he told reporters, “but it’s only as good as the information you use to train it.” His more pointed admission was about how Ford had tried to deploy it.

Poon’s explanation cut to the assumption at the heart of the rollout.

Mistakenly, we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that would produce a high-quality product.

Charles Poon, Ford’s vice president of vehicle hardware engineering, told reporters.

He tied the failed push to a related failure: underinvesting in the engineers who understood the products. “Over prior years, we didn’t pay as much attention as we should have to the experience of our most knowledgeable engineers that have been with us through many product cycles,” Poon said. Many of those veterans had already left the company.

The cameras at line speed had become the eyes of a system Ford had hollowed out. Poon said the failure was about how the AI was trained, not the AI itself.

The pattern is showing up in other industries running AI for high stakes validation. AI quality engineering work inside banking software has become a continuous discipline, with teams running test cycles on the same release cadence as the developers they support. Poon’s framing this week was that AI inherits the limits of the experience used to train it. Most of that experience, by Ford’s own admission, had already left the company.

The “Gray Beard” Engineers Ford Wants Back

The engineers Ford is bringing back carry a corporate label. The company calls them “gray beard” engineers, a term that has travelled from internal jargon to public admission. Galhotra told reporters the returning specialists have rebuilt the AI tools’ training and now run troubleshooting sessions on the defects the cameras missed. They also mentor the junior workers who staff the next shift. Ford previously let those experienced engineers leave before the knowledge could be folded into its machine learning systems.

The rehires sit inside a wider leadership shake up. Ford has replaced about two thirds of its senior leaders across engineering, supply chain and manufacturing over the same period. The original report on Ford’s gray beard rehires described the move as unusually human. The wider struggle to scale AI out of pilot shows the same fault line in other sectors.

From 15th in 2023 to First in 2026

The numbers behind the reversal sit on the J.D. Power 2026 U.S. Initial Quality Study. Ford ranked first among mass market brands at 152 problems per 100 vehicles, the highest score in its class. It is the first time since 2010 that Ford has held the position.

Farley framed the number as the payoff for years of unglamorous work. The announcement behind Ford’s 2026 quality ranking lays out the climb: 15th in 2023, then first in 2026. Ford improved by 41 problems per 100 vehicles year over year, the largest such gain among mainstream brands. The index tracks owner reported problems in the first 90 days of ownership, drawing on verified repair orders and customer feedback. Ford beat mass market runner up Nissan at 156 PP100 and Buick at 162 PP100, per the 2026 U.S. Initial Quality Study results.

The result was carried by the lineup. Several Ford models landed at or near the top of their J.D. Power segments, including three that topped theirs outright.

• F 150, Mustang and Super Duty each ranked highest in their segments, the second straight year the three topped their classes.
• Escape, Explorer, Expedition and Maverick also landed in the top three of their segments.
• Seven of Ford’s ten models tested placed in the top three, the highest share of any automaker J.D. Power surveyed.

The seven of ten figure is the most direct measure of how broadly the quality push had reached across Ford’s lineup. The roster of top three finishes spans seven of Ford’s ten models tested this year. “This is a milestone, not a finish line,” Galhotra said.

Recall Costs Still Track the Past

The turnaround is uneven. CBT News, citing Bloomberg, reports Ford expects $1 billion in warranty and materials costs this year. The company remains the most recalled automaker in the United States. Galhotra told reporters the recall count tracks vehicles built under the earlier system.

As vehicles built under the new approach reach customers, Galhotra said, the recall count should fall on its own. Ford merged several of its operations, including electric vehicle and digital teams, into a single Product Creation and Industrialization unit. Galhotra runs it as chief operating officer.

The financial pressure on the old system has produced a financial tailwind on the new one. Farley put a number on the savings this week.

“literally hundreds and hundreds of millions of dollars of a tailwind for Ford on cost.”

Jim Farley, Ford president and CEO, per Bloomberg.

The sum reflects declining warranty and recall expenses flowing through the P&L as the rehired engineers’ fixes reach vehicles already in service.

AI Keeps a Place on Ford’s Plant Floor

AI stays on Ford’s factory floor. Bryce Currie, Ford’s chief manufacturing officer, said the automaker is investing in plant floor tooling that includes AI vision systems designed to flag anomalies. The cameras the system had relied on are being retrained with data the returning engineers now provide.

The shift echoes a comment Farley made a year earlier to author Walter Isaacson. “AI will leave a lot of white collar people behind,” Ford’s boss said then.

Suppliers are folded in earlier in the cycle. Ford’s chief supply chain officer Liz Door said the change drove a 30% reduction in launch issues year over year. “It’s easy to celebrate heroes fixing problems,” Door said. “What we really want is to celebrate zero defects.”

The closing numbers, lined up:

• 152 PP100 on J.D. Power 2026, first among mass market brands.
• First mainstream IQS win for Ford since 2010.
• 41 PP100 year over year improvement, the largest among mainstream brands.
• More than 300 veteran inspectors rehired over several years.
• AI vision systems still being added to plant floors, retrained by the same veterans.

Each line on the list is a defect or training gap the AI cameras on the line once missed. Galhotra has said the next test for the rebuilt system is whether the recall count actually falls as vehicles built under the new approach reach customers.