Ford recently revealed that the company had to rehire about 350 experienced "greybeard" engineers over the past three years to mentor younger employees and rewrite previously underperforming diagnostic systems and AI tools to make up for glaring flaws in quality control. Charles Poon, Ford's vice president of vehicle hardware engineering, said in an interview that management had previously underestimated the deep experience accumulated by senior engineers who have experienced multiple rounds of product iterations, and simply replacing them with AI "is a huge mistake."
Poon emphasized that AI is “a very good tool,” but it “can only be as good as the information used to train it.” Without high-quality empirical data and correct application scenarios, it is difficult for algorithms alone to support the reliability requirements of complex automotive products. The old engineers who were hired back are now arranged to host mandatory regular meetings, focus on troubleshooting vehicle problems, and reprogram automated engineering software and AI tools to eliminate potential defects as much as possible before parts enter the production line.
The focus of these technical experts is to identify weaknesses in design and manufacturing processes at an early stage to reduce losses caused by large-scale recalls and quality defects. Ford has reportedly spent billions of dollars on quality issues and recalls, and the company aims to cut $1 billion in expenses this year, so improving quality is seen as a key to survival and profitability.
Judging from external indicators, this adjustment has begun to have an effect. In last year's J.D. Power New Vehicle Quality Survey, an annual study that evaluates a vehicle's quality performance during its first three months of ownership, Ford ranked only 10th among mainstream brands, scoring below the industry average. However, in the same list this year, J.D. Power ranked Ford No. 1 among mainstream brands, surpassing competitors such as Toyota and Honda. Ford attributed this "leap-forward" improvement directly to the return of professional capabilities and experience of rehired engineers.
Ford's experience is also seen as a microcosm of the current wave of large-scale "AI replacement" in enterprises. Careerminds, a workplace service organization, previously conducted statistics on companies that had implemented "AI-driven layoffs" and found that 35.6% of them later had to rehire more than half of the laid off employees, and another 32.7% rehired 25% to 50% of the original employees. This shows that in practice, many companies find that AI cannot fully undertake complex work functions originally completed by humans, and have to make corrections between cost and business continuity.
Fintech company Klarna is another prime example. In 2024, the company's CEO Sebastian Siemiatkowski announced in a high-profile manner that the newly launched chatbot took on the workload equivalent to 700 full-time customer service personnel in its first month of launch. Accordingly, the company froze recruitment and eliminated hundreds of positions. However, by mid-2025 and 2026, Klarna began to step up its recruitment of human customer service personnel as customer satisfaction dropped significantly.
It turns out that AI is excellent at handling simple, standardized questions such as checking account balances, but once faced with complex user appeals that require understanding context, emotion, and nuance, it can easily slide into mechanical, blunt, and "robot-like" answers full of corporate jargon that fail to truly solve the problem. In this case, users' patience for pure AI services is quickly exhausted, and companies have to re-examine the radical strategy of "total AI".
Ford's admission of mistakes and return to "human-machine collaboration" has been seen by many industry observers as a wake-up call: Even with the rapid development of AI technology, experienced human experts still play a key role in complex system design, quality control and customer service. For those companies that hope to lay off large numbers of employees and directly replace them with AI, how to find a balance between efficiency improvements and professional experience may be more important than simply pursuing automation.