Upgrading the production line

Optimized manufacturing pairs highly skilled workers with advanced technologies

Smart manufacturing can helps mitigate labor shortages through faster, more accurate workflows.

Manufacturers are responding to ongoing supply-chain disruptions with AI investments to create “smart factories” that are more resilient and agile. The result is an explosion of interest in new technologies that analysts are calling “Industry 4.0.”

“Industry 1.0” was the original Industrial Revolution of the late 18th and early 19th centuries, which transformed manufacturing via steam and water power, machine tools, and mechanized factories. Industry 2.0 introduced electricity to the production line. The most recent round of industrial innovation, known as Industry 3.0, was characterized by the adoption of physical robots—think human-programmed robotic arms that build cars on a factory floor.

In Industry 4.0, the robots get smarter and more independent. The robots building cars take in data—how many cars they can build per hour, how many parts they need to build each car—and use AI to interpret that data and make sure the production line runs without a hiccup. AI can fine-tune its own work without human involvement.

According to a global survey from ServiceNow and ThoughtLab, which polled 900 executives in 13 countries on the state of optimization in industry, manufacturers are investing more in optimization technologies such as AI than any other vertical except financial services.

Thorsten Wuest, an associate professor of smart manufacturing at West Virginia University, says manufacturing companies have access to more data and better analytical tools than ever before. “The factory floor is now connected with IoT [internet of things] devices and sensors, so manufacturers have access to more real-time data,” he told Workflow. “At the same time, processing has improved. We have better algorithms, open-source programming languages like Python, and low-code tools, all making processing more accessible.”

Since the start of the pandemic, a quarter of manufacturers have made major strides optimizing production using AI and machine learning, according to survey results. That percentage is set to grow to more than half in the next few years. Executives reported that they were investing in modernizing IT platforms and technology to better share data across organizations.

Smart factories vs. labor shortage

At the beginning of the pandemic, about 1.4 million Americans lost manufacturing jobs. Even as the industry has rehired many workers, hundreds of thousands of positions remain unfilled. According to a report from Deloitte and the Manufacturing Institute, manufacturers are having trouble finding entry level and skilled labor for factory jobs. This skills gap, according to the report, is projected to leave more than two million jobs unfilled by 2030, costing the U.S. economy as much as $1 trillion.

1/4

Manufacturers that have made major strides in optimizing production with AI and ML during the pandemic

Smart factories could help mitigate that problem. Wuest highlights the importance of robotic process automation (RPA), which uses software bots to automate repetitive computer tasks like changing passwords or routine data entry. The result is faster, more accurate workflows that require fewer human workers. “Automation helps us avoid injuries and boring work on the physical side, AI is helping us do that on the cognitive side,” Wuest says.

Factory equipment like “cobots,” or collaborative robots, carry out routine tasks on the factory floor without human help. Although cobots require configuration, they are programmed using low code, which breaks down complicated programming languages into simple building blocks that non-technical people can use. Cobots free up developers to carry out more difficult tasks on the factory floor.

The fear that these automation tools will displace human jobs is misplaced, according to Wuest. Technologies like RPA and cobots are designed to work with human workers, not to replace them. “With AI that performs predictive maintenance,” he says, “we know when a tool will break. But humans maintain the robots and the equipment.”

Although manufacturing is known as a conservative industry, the pandemic has forced it to adopt new technologies. “Manufacturers have been using Zoom, training personnel remotely, taking meetings virtually,” says Adrian Dima, a technology consultant and engineer. “The pandemic was an accelerator.”

Reaching the skeptics

Despite the potential, many manufacturers are reluctant to implement AI and ML systems due to the high costs associated with such investments, says Rui Alves, an associate professor of economics at the Polytechnic Institute of Setubal in Portugal. “When we’re talking about breakthrough technologies like ML and AI, we’re talking about radical innovations,” he says. “And for conservative sectors, this radical innovation can sometimes be a barrier.”

2M

Number of unfilled manufacturing jobs by 2030

Moreover, most IT tools designed for manufacturing aren’t user-friendly. And as many of those who work in these fields lack an IT background, learning such tools requires a heavy investment of time and money. “It can be expensive to implement these technologies,” Alves says. “Change has a cost, especially in an industry where resources are very important.”

To reach skeptics, Dima advises developers to educate potential manufacturing clients on why, how, and when to use these tools in their own language. Operational technology developers often communicate with manufacturers as though they share an advanced understanding of the technology, he says, which isn’t always the case. “We have to educate them on their level,” Dimas says. “We can’t talk to them like they’re in IT.”

Before executives invest in these technologies, they should understand the processes they want to automate and what they hope to get out of it, says Wuest. He also cautions executives to allow experts who understand the technology to have the space to use it innovatively on the production line. “This process shouldn’t be top down or bottom up,” he says. “You should have the support from the top but the freedom to implement from the bottom.”