Ford has rehired former engineers to correct errors caused by its automated systems. This comes despite the company ranking first in the J.D. Power Initial Quality Study among mass-market automakers and openly discussing the challenges it has faced in recent years, particularly regarding its reliance on automated systems in manufacturing and design.
It became clear that these automated systems were not as reliable as previously thought, prompting Ford to hire experienced technical specialists—sometimes bringing back former employees—to rectify mistakes made by the company’s robots. According to Ford, while artificial intelligence (AI) is powerful, it is also prone to flaws, and its effectiveness relies heavily on the quality of the data used to train AI models. The automaker also underestimated the importance of the institutional knowledge held by its seasoned engineers, who had participated in multiple vehicle development cycles. This combination of factors contributed to a decline in the quality of Ford vehicles.
“We mistakenly believed that simply implementing artificial intelligence and adjusting our design requirements would ensure a high-quality product,” said Charles Pun, vice president of vehicle hardware engineering, during a briefing with journalists this week. Pun noted that some of Ford’s most experienced employees left before all their accumulated knowledge could be fully transferred into the company’s automated systems. This necessitated the return of some of these workers to retrain the systems or, in some cases, to mentor younger engineers who are now responsible for maintaining the quality of Ford vehicles.
Pun indicated that Ford has hired, promoted, or rehired over 350 experienced engineers to restore this level of expertise. In addition to mentoring younger engineers, they are also tasked with enhancing data collection and AI training that support Ford’s automated systems. “This is where some of our most experienced engineers had experience solving and identifying these problems before they seep into the system,” said Pun.
Changing Approach to Quality and Software
Currently, Ford leads the industry in the number of recalls, and its quality ratings have declined over the past few years. These challenges have become more pronounced recently due to difficulties with the launch of the Explorer and Aviator models, supply chain disruptions during the COVID-19 pandemic, and a notable increase in vehicle recalls. According to Ford’s Chief Operating Officer Kumar Galhotra, the automaker ultimately concluded that its approach to quality had become too fragmented. Different departments were operating in isolation, and the company largely relied on a “find and fix” philosophy that focused on identifying defects after they appeared and addressing them as quickly as possible.
While this approach could resolve immediate issues, it did not prevent them from arising. “We are moving from a ‘find and fix’ mentality to preventing problems before they occur,” said Galhotra. “We are focused on contributing factors and early indicators rather than end results. Stop watching the problem and start solving it.”
The transformation extends beyond vehicle hardware. Software and digital technology development teams are now collaborating more closely with vehicle engineering, manufacturing, and supply chain teams, executives reported. Ford is now attempting to combine the speed and flexibility associated with software development with the rigor and validation requirements of automotive engineering. Historically, this has not always been the case.
Pun explained that Ford was detecting software bugs only in the late stages of the process because it did not fully utilize the available cycles of rapid iteration. However, the automaker could not release software updates as quickly as consumer electronics companies with a “move fast and fix later” mentality, Pun noted. Cars, unlike smartphones, operate in a safety-critical environment where customers depend on the software functioning correctly from the moment the vehicle is delivered.
To address this issue, Ford established a specialized team of 40 people dedicated to software quality assurance, whose sole responsibility is to prevent problems before they occur. However, this does not mean that Ford is not looking to integrate AI into more of its processes.
The automaker claims to have significantly expanded its automated testing capabilities, adding over 100,000 new AI-based tests designed to detect edge cases and stress software systems under a variety of conditions. Because the testing system is highly automated, software changes can be quickly revalidated even in the late stages of development, ensuring that modifications do not introduce new defects. “Because these tests are highly automated, even if we have a late change in the software, we can quickly go through the entire validation process to ensure it works perfectly before it reaches the customer,” said Pun. “We have established software reliability as a separate strict discipline with rigorous metrics.”
Source: The Verge