Why multiple AI models help tackle the most complex manufacturing problems

By | 1 minute read | January 1, 2021

Engineer operates maintenance equipment

Engineer operates maintenance equipment powered by AI

Faster than a spinning servo, AI is changing how factories operate.

According to a recent IBM C-suite study in collaboration with Oxford Economics, 75 percent of executives say they will invest in AI in the next three years to create new business models at the edge, combining intelligent workflows, automation, and edge device interconnectivity.


Investing in AI means bringing together operational technology (commonly called OT), and information technology (our ever-present IT). This pairing is foundational to many Industry 4.0 advances. Applying sophisticated analytics and machine learning on the factory floor comes with complexities, and often requires architects who have experience in both OT and IT.

“A year ago, six months ago, people were applying AI in what I call one dimension; today, they might have a problem that takes multiple models to solve,” Meek said. “An example is our welding project at Magna. We actually have 11 AI models that run near simultaneously to assess the quality, and each of them give a different piece of the puzzle.”

Manufacturers need to know beyond a shadow of a doubt that the product they’re creating is viable. That’s where multiple AI models are coming in. Prior to the integration of AI, manufacturers could only assess one thing at a time—for example, using a computer vision system to make sure a label is applied correctly.

“What we’re seeing now is the tougher problems coming to light,” said Meek. “And manufacturers are less able to solve these problems in-house without advanced technology or the skills needed to deploy these technologies.”

However, nearly 80 percent of manufacturing executives believe the skills gap will significantly impact their ability to increase productivity.


“Although AI is becoming mainstream…large-scale benefits come with the ability to implement this AI at scale,” Lehtonen said in a virtual session on Industry 4.0. “If you want to be really intelligent in manufacturing, you need to have the data and knowledge available from different elements and levels of your manufacturing and IT.”

Want to learn more? Read about how 5G is helping manufacturers realize the promise of data.

More in this series: