CHICAGO — Artificial intelligence (AI) offers a lot of promise for food manufacturers across the value chain: predicting consumer behavior, reducing experiment cycles when reformulating and making manufacturing more efficient.

At the Institute of Food Technologist (IFT) FIRST 2024, held in Chicago, July 15-17, Eric Krums, senior solution consultant at Infor, Sandeep Anand, senior director, head of machine learning solution at Infor, and Michael Warter, senior vice president and chief information officer at Ruiz Food Products, Denison, Texas, shared how AI machine learning and product life cycle management offers opportunities to optimize product development and maximize return on investment. The panel also shared some of the challenges and first steps food manufacturers need to take to set themselves up for success when using AI.  

AI machine learning does have a wealth of potential to assist food manufacturers across the entire value chain, Anand said. It can accelerate new product development and even ingredient substitutions based on pricing and availability. It can also drive reactions to consumer behaviors. 

“Machine learning allows you to take the data and make predictive models,” he said. “With AI you want to look at the relationship between ingredients, recipes and customer demands but also manufacturing.”

Food manufacturers cannot take advantage of technology like AI machine learning. Clean data is critical to building reliable predictive models that AI can work from. Product Lifecycle Management (PLM) systems help manufacturers with formula management and regulatory compliance. They can even interact with suppliers and integrate with the existing ERP system in a food manufacturer’s facility. 

“This is the foundational element that everyone needs to get to before they can start thinking about AI,” Krums explained. “This is where you can store your data, have the correct labels and ensure you’re in regulatory compliance. Once that’s set up you can play with fun tools like AI.” 

By incorporating many of these foundational tools for data collection and management, Ruiz Foods was able to keep R&D connected to manufacturing. That integration is key to creating faster processes and sets Ruiz Foods up for success as it experiments with AI machine learning. However, Warter noted several challenges that must be addressed about AI in the food system. 

First is helping people understand what AI is and how it can be used. AI machine learning is just one type of AI, the other two being deep learning and generative AI. Machine learning makes predictions based on models and data, while generative AI can sometimes create information that isn’t true, so it must always be verified. 

The second challenge is that AI is not a project relegated to the IT department. 

“If you’re going to drive change with AI, it starts with really understanding your processes and the data they create so you can build effective models for it,” Warter explained. 

That means getting operations, R&D and QA teams involved. 

The third challenge that Warter cited is the lack of regulation and security around generative AI. 

“I can see the promise, but how do we ensure that what we put in there stays secure to us,” he said. 

Krums agreed, stating: “How do we keep the secret sauce secret?” 

AI machine learning has a lot of potential to increase efficiency across an entire bakery or snack operation, but it takes a foundation of a clean and robust data set to build an effective model for machine learning to work from. 

“The value is being able to use AI across the entire ecosystem and being able to do this data-driven decision making that can be leveraged to create an interconnected ecosystem,” Anand concluded.