5 minutes read time

AI-driven improvements to automotive manufacturing, from QC to maintenance and more

Innovations like AI-enabled quality control, predictive maintenance, and collaborative robots (or “cobots”) have already shown great ROI in the factories where they’ve been installed. BMW, one of the leading-edge adopters, reports saving more than 500 minutes of unplanned stoppages per year through predictive maintenance. Meanwhile, Hyundai and Daimler Trucks North America (DTNA) are using sensors and AI algorithms to track supply chains, receive parts shipments, and prep production cells. All of this adds up to saved time, resources, and ultimately money.

These kinds of benefits are, if anything, likely to grow in the coming years. The technology is getting smarter, more mature, and more affordable as adoption spreads. And some growth categories like EV, with vehicles sporting fewer parts and more prefabrication, are particularly well-suited to benefit from smarter automation. Automotive battery manufacturers also stand to benefit enormously thanks to AI vision systems that ensure correct assembly and adhesive application.

Yet success stories like BMW’s and Hyundai’s are still just leading indicators rather than the industry-wide standard, and AI adoption is still far from universal. So why the lag?

Information and complexity are the biggest obstacles

In part, you can blame poor communication. Success stories like these aren’t as widely known as they should be. The lack of familiarity with the technologies can also be daunting, even for veteran engineers: the transition to AI can feel like a much bigger leap than the incremental improvements brought by earlier waves of innovation.

The perceived complexity might be an even bigger factor. In a recent survey, 65% of automation engineers reported dealing with long setup times for product changes of all sorts. The real problem, it turns out, may not necessarily be with AI, but with making changes in general.

This is why rapid reconfiguration is so crucial to the future of AI-enabled automation. One of the great promises of AI—besides efficiency and uptime—is flexibility. Especially as motor vehicles themselves become more advanced, being able to rapidly reconfigure a factory to accommodate improvements is critical. But for many factories, throwing AI into the mix adds yet another layer of complexity, making it appear less nimble, not more.

There’s an AI for that too!

Fortunately, some of the most promising advances in AI automation are pointed in exactly that direction. French startup Inbolt is a good example. Using AI vision, they’ve developed a system that can match real-world configurations to CAD models or 3D scans, cutting reconfiguration times from days down to hours, while allowing robots to work in far less structured environments. In many cases, custom jigs and indexing tools can be dispensed with entirely.

Another big obstacle to reconfiguration is the human element: training. Gen AI offers much promise here too. Thanks to digital twins, virtual simulations of factories are becoming more detailed and immersive, and can be built faster and more cheaply than ever. Imagine, then, a virtual factory, accessed through VR or screen-based interfaces, that perfectly simulates a reconfiguration and guides the factory team through their new tasks—all while actual physical improvements are being installed. The result is quicker, more effective training, so teams can hit the ground running as soon as installation is complete.

This is one of the things that sets AI-enabled automation apart. It’s not just a technology that improves speed, quality, and reliability. It also uses that same flexibility to ease adoption. “AI automation helping you implement AI automation” might sound a little sci-fi, but that’s exactly what the market has in store in the next few years.