Ford had to hire old engineers to fix the mistakes made by their automatic systems

To celebrate its new status as No. 1 in JD Power’s first quality rating among mainstream automakers, Ford opens up about the issues it has faced in recent years, particularly regarding its reliance on automated systems in manufacturing and design. It turns out that those automated systems weren’t as powerful as previously thought, requiring Ford to hire experienced technicians — sometimes bringing back former employees — to fix the mistakes made by the company’s robots.

In Ford’s view, AI is powerful and prone to pitfalls. Its effectiveness depends entirely on the quality of the data used to train the AI ​​models. In addition, the automaker has underestimated the value of institutional knowledge gathered by its veteran engineers who have worked through multiple vehicle development cycles. And this combination of circumstances led to a decline in the quality of Ford cars.

“Inadvertently, we thought that by introducing artificial intelligence and fixing the design requirements we had, that would produce a superior product,” said Charles Poon, VP of automotive mechanical engineering, at a press conference this week.

According to Poon, some of the company’s most experienced employees left before all their knowledge was fully transferred to Ford’s automated systems. That made it necessary to bring back some of those workers to retrain those systems, or in some cases, to enrich the junior engineers who were struggling to maintain the quality of Ford vehicles. Poon said Ford has hired, promoted, or brought back more than 350 experienced engineers to rebuild that technology layer. In addition to mentoring junior engineers, they were also tasked with developing data collection and AI training that supports Ford’s automated systems.

“That’s where some of our more experienced engineers have the knowledge to solve and identify those problems before they get into the system,” said Poon.

Ford currently leads the industry in the number of recalls, and its quality ratings have declined over the past few years. Those challenges have become more prominent recently, with difficulties associated with the launch of the Explorer and Aviator, supply disruptions during the covid crisis, and a noticeable increase in the number of its recalled vehicles.

According to the COO of Ford, Kumar Galhotra, the car maker finally concluded that the way he looks at quality has become very different. Different departments worked in silos, and the company relied heavily on a “find and fix” philosophy that focused on identifying errors after they occurred and fixing them as quickly as possible. While that approach may have addressed immediate problems, it has not prevented those problems from occurring in the first place.

“We’re moving from that catch-and-fix mentality to preventing problems before they happen,” Galhotra said. “We focus on enablers and early indicators versus outcomes. Stop loving the problem and start solving it.”

The change extends beyond the vehicle’s hardware. Software and digital teams now work more closely with automotive engineering, manufacturing, and supply chain teams, executives said. And Ford is now trying to combine the speed and flexibility associated with software development with the strict requirements and validation of automotive-grade engineering.

Historically, this has not always been the case. Ford was finding software bugs late in the process because it wasn’t taking full advantage of the fastest cycles available, Poon said. That said, the automaker hasn’t been able to roll out software updates as quickly as electronics companies with the mentality that they can “move fast and fix later,” Poon said. Cars, unlike smartphones, operate in a safety-critical environment where customers depend on software that works well from the moment the car is delivered. To fix this, Ford created a dedicated 40-person software quality assurance team whose sole responsibility is to prevent problems before they happen.

But don’t think Ford isn’t committed to integrating AI into many of its processes. The automaker says it has dramatically expanded its automated testing capabilities, adding more than 100,000 new AI-powered tests designed to identify edge cases and stress software systems under a wide variety of conditions. Because the testing framework is highly automated, software changes can be reviewed quickly even later in development, to ensure that changes do not introduce new defects.

“Because these tests are so automated, even if we have a late change in the software, we can run back through the entire validation process to make sure it’s working properly before it reaches the customer,” Poon said. “We’ve established software reliability as a solid one with solid metrics.”