Teaching AI to Ops: A Continuous Education

Much has been written about artificial intelligence (AI), its implementation into the food & beverage industry and its potential to revolutionize various parts of the processing world.

Companies have already begun investing — some heavily — in AI systems to perform all sorts of tasks along the production spectrum. AI has worked its way into product formulation, finance and administrative departments, human resources, food safety & quality and operations. In many cases, AI is used to sort, handle and process data, and even formulate an outcome or suggestion, or present a case for change or an actionable solution.

The more AI “learns,” the better it gets. However, while AI can learn by doing, the question begs: How do food & beverage workers learn how to use AI and the information it can provide? Further, how much should companies invest in a system if their employees can’t use it or won’t because they’re not familiar with it?

A culture shift

Processors know all about installing and training on new equipment and lines and in new facilities. Good operations managers and good OEMs can join together and get line workers, supervisors and maintenance teams trained and well-versed on new machinery at a rapid clip nowadays. But AI is a different story.

“The biggest challenge isn't teaching the tech, it's leading the change,” says Jill Stuber, vice president and co-founder of consultancy Catalyst Food Leaders (www.catalystfoodleaders.com). “Adopting AI is more than a systems upgrade; it's a culture shift.”

That culture change, she continues, demands a different kind of leadership. Those atop the organization chart and in charge of training and implementation need to focus on building their human skills first.

“Leaders who invest in building human skills like clarity, empathy, coaching and trust are the ones who win the race on outcomes, retention and engagement,” Stuber says.

This past May, Kellanova gave a few glimpses into how the “AI revolution” is transforming the workplace there and how the company is assisting its workforce through implementation and adoption. The announcement covered all sorts of AI and generative-AI innovations the company had begun working on, whether on the production floor, in the R&D labs or throughout the corporate offices.

Thematically, Kellanova’s tech executives highlighted the company’s strategy to champion a “test-and-learn culture,” allowing employees at any level to flex their curiosity, explore the new options and learn about the opportunities they offer at their own pace.

That approach has led to a good level of acceptance of the program by workers who have participated thus far, the company says — a point that is critical in the old-school landscape of the food & beverage processing industry, says Mike Smith, reliability engineering manager for Life Cycle Engineering (www.lce.com).

“To make implementation effective, we really have to focus a lot on the organizational change management aspect of it,” Smith says, “making sure that we bring them along, help them understand the benefits, show them why it benefits them and the company.”

Stuber agrees, saying that operators really need more than instructions. They need clarity on why the tool matters, particularly how it impacts their role and responsibilities, and what is expected of them. Still, Smith says, every person is different, and the reactions can be night-and-day different at first.

“Some employees will push back against any big change, and you really have to show them and sell them on it,” he adds. “But some will jump right in and be 100% behind what you’re doing; that’s great, but it’s also not the norm.”

Getting maintenance employees fully engaged and understanding and then championing the AI effort can be a significant win for the AI implementation team, Smith says.

“Once maintenance gets on board with it, they buy into it a lot, and most of them are pretty tech-savvy people,” he explains. “Once they buy into it, you have a good support organization — but it will take some time to get there; it’s not instantaneous.”

Operators also would be wise to monitor the message around why the company or plant is implementing AI and the ultimate result of its use, Stuber adds.

“AI can be scary since people have heard it will replace jobs,” she says. “Leaders need to be transparent about what the tool is for and how it supports the team, and then support people through the change.”

An ongoing investment

Even after a company has implemented AI and feels as though its workforce is trained to capitalize on the efficiencies and opportunities, the work isn’t done. Kellanova says it is moving beyond isolated pilots of its AI initiatives and focusing on democratizing access to AI tools and investing in upskilling to “empower people to harness AI’s full potential.”

Stuber would give positive marks to this concept, confirming that AI adoption isn’t a one-time training event.

“What really moves the needle is ongoing support, trust-building and personal connection,” she explains. “You have to reinforce learning over time, check in with people as humans, and make space for them to voice stress, confusion or ideas.”

Commitment to long-term support can be overlooked, leading to an expensive case of “buyer’s remorse” down the road, Smith says.

“I’ve seen companies spend hundreds of thousands of dollars putting sensors on everything they’ve got in the plant, and then this massive amount of data comes in and it doesn’t get analyzed properly,” he says. “They don’t have a process to take that information, turn it into a corrective work order that gets planned, scheduled and ultimately repaired. And if you don’t have those systems, you’ve just spent a lot of cash and you’re not getting anything out of it.”

And, of course, it always helps if a plant or company has an AI champion who fully understands the technology and reasoning for implementation and can give the transition a shot in the arm any time enthusiasm wanes.

“You have to have that person who is the sponsor, who kind of coaches, leads, directs and cheerleads, whatever may need to be done,” Smith says. That champion can help rally workforces that got behind the implementation, helping to maintain any traction the system may have had built up with employees and their motivation to engage and participate.

If the long-term approach matches up with the system implementation — and companies invest in further support to continue training and proper analysis of the inputs — Smith predicts those companies will see longer-term success.