🔒From robots to data analysis, AI could reshape manufacturing in CT

As companies continue to explore the best ways to leverage artificial intelligence, the Italian parent company of a prominent Connecticut manufacturer plans to launch a new AI-related business in the state next year that will help companies modernize their operations.

Italian airfoils compressor manufacturer Pietro Rosa purchased AI-based knowledge management company R-Tree Technologies in 2022, and is implementing the company’s technology in its subsidiary businesses internationally, including at Farmington-based New England Airfoil Products.

While R-Tree is based in Italy, Pietro Rosa plans to open a U.S. headquarters for the AI company in Connecticut, where it will further refine its technology. R-Tree also plans to eventually sell its technology to other U.S. manufacturers, according to Clive Cunliffe, Pietro Rosa’s North American president.

R-Tree uses a company’s data, research and other knowledge sources to create a searchable and all-encompassing database that can generate quick answers to specific employee questions.

For example, if a company plans to do business with a Korean firm, the technology can search its own database to find historical information, like if the company has ever done business in the market before, who led the effort and when, Cunliffe said.

“Because it’s all contained within a knowledge hub within the company,” Cunliffe said, “you can ask, ‘Do we have anybody who speaks Korean? Have we ever done business in that part of the world? Have we ever had any issues there?’ Then it will pull together an … answer to the question that you asked. If you can access the knowledge that you own really quickly, you’ve got something that helps differentiate you in the marketplace.”

Pietro Rosa’s AI investment is one example of how manufacturers are using new technologies to modernize and upgrade their operations.

Accounting firm Marcum released its 2023 National Manufacturing Survey in August, which concluded that AI has moved “from a buzzword to a real-life tool.”

Fifty-five percent of respondents said their company is investing in technology over the next year, including in AI and augmented reality. Companies are using AI to reduce expenses, support automation or improve operations, including logistics, supply chain management and employee training.

“Artificial intelligence is making radical progress and offers promise for easing the labor crunch and driving efficiencies in manufacturing,” said Jonathan Shoop, a partner at Marcum who is also a consumer and industrial products leader.

Still, the technology’s use remains in its early stages, the Marcum survey concluded.

Connecticut stakeholders say manufacturers’ adoption of AI will not only be crucial, but necessary.

“I think all the manufacturers that do not invest in AI will be left behind eventually, because that’s the way to go,” said George Bollas, director of the Pratt & Whitney Institute for Advanced Systems Engineering at the University of Connecticut. “It’s really optimizing processes automatically and can enable increased productivity by who knows how much.”

State industry leaders view AI as a way to augment, not replace, workers in manufacturing and other industries. There are about 92,000 unfilled jobs in Connecticut, many of which are in the manufacturing sector, federal data shows.

“There’s the big question of if AI will replace the workforce,” said Connecticut Chief Manufacturing Officer Paul Lavoie. “Well, on the manufacturing floor it won’t replace the workforce, it will augment the workforce and allow it to work on higher-value tasks. We need to look at how we can use AI, machine learning and robotics to do the work that we can’t find people to do. AI doesn’t replace a worker, but a worker who understands the AI will replace a worker who doesn’t.”

Best AI uses in manufacturing

Artificial intelligence is being used by manufacturers in several ways, experts said. Lavoie said an increasing number of companies are adopting automation and robotics technology to complement their workforce, especially where repeatable tasks are involved.

Robots, for example, can be used to build cardboard boxes and essentially work on an assembly line, freeing up employees to do more high-value tasks, Lavoie said.

Data analysis is another huge use.

“We’re really looking at how to use AI to help drive the interpretation of data to get different results on a shop floor … to come up with better methods,” Lavoie said.

AI systems can help make production lines run more efficiently, said Jeff Orszak, director of business technology and innovation at CONNSTEP, a manufacturing industry consultant affiliated with the Connecticut Business & Industry Association.

Companies can input data into AI-backed systems that recommend, in real time, the best way to move parts through a production facility based on yields, performance and quality controls, Orszak said.

The Connecticut Center for Advanced Technology in East Hartford has been working with colleges and the state’s smaller manufacturers to develop AI models and machine-learning algorithms that companies can adopt and use on their shop floors, said Jackie Garofano, CCAT’s chief technology officer.

CCAT is also helping Pietro Rosa further develop its newly acquired AI technology for use in the U.S.

“We have workforce challenges that we all know and recognize and are trying to tackle, and leveraging these technologies is one way to get there,” Garofano said. “As an R&D center, we want to create turnkey solutions to then put onto the shop floor.”

Machine maintenance is another current AI use.

UConn’s Bollas said he’s been working with CCAT, CONNSTEP and his engineering students on ways to train AI so it can detect when machines aren’t operating at their full capacity.

Companies can use AI to understand and predict when a machine needs maintenance or upgrades, potentially preventing situations where assembly lines go offline for long periods of time, he said.

Future of manufacturing

Orszak said he thinks of AI as a “longer play” of Industry 4.0, a term used to describe the ongoing automation and technological advancement of traditional manufacturing practices.

“My feeling is that AI adoption will really kick in after five to 10 years, and what companies need to start doing now is build a data strategy so that they have an understanding of the data they’re collecting and have put together the information that an AI model can use,” Orszak said. “My message for companies is that these are going to be the productivity-increasing methods of the future, so companies need to start building strategies to put them in place.”

Orszak and Garofano emphasized that not all AI models can be replicated across different companies — a machine shop building aerospace parts might need a different algorithm than a company that develops medical equipment.

And not all companies have the capacity to develop AI algorithms internally, making collaboration important.

“The challenge is more about closing that gap between technology that’s been developed in a higher-ed institution and the technologies that can be deployed at a small manufacturing shop,” Bollas said. “If there are three people that run around and have to produce 1,000 components by the end of the day, they don’t have time to develop AI algorithms.”

Data security is also a concern, especially for companies experimenting with generative AI systems that are available in the public domain.

For example, New England Airfoil Products plans to implement R-Tree’s AI technology after it develops a “shielded version” that protects the information of its aerospace and defense customers, Cunliffe said.

CCAT is also helping with that.

Regardless of the industry, adopting new technologies will be key to driving future business growth, said Lavoie, the state’s chief manufacturing officer.

“We’re never solving the workforce challenge with people — there’s never going to be enough people,” Lavoie said. “If we don’t look at industrial automation, artificial intelligence and machine learning, we’re going to lose significant amounts of GDP growth and manufacturing here in the state and country.”