The Institute of Food Technologists (IFT) is leveraging artificial intelligence in its new R&D platform, CoDeveloper, to connect food scientists with the organization’s massive research database.
IFT highlighted the CoDeveloper service in a recent webinar, noting the platform’s ability to quickly solve formulation challenges.
CoDeveloper users also gain access to Sous, a generative AI-powered “co-scientist” equipped not only with the ability to build new products but also to reverse-engineer existing ones, according to IFT.
“The food system is evolving rapidly, and IFT is uniquely positioned for this moment,” said IFT CEO Christie Tarantino-Dean in July. “With more than eight decades of supporting scientific leadership and a global community of experts behind us, CoDeveloper is our bold response to the urgent need for smarter, faster R&D tools, and it reflects our commitment to shaping the future of food through trusted science and information.”
Use case: Prebiotic yogurt
Rene Leber, CoDeveloper customer success manager at IFT, works with food and beverage R&D teams to solve problems such as shelf-stability, reformulation, food safety and regulatory compliance.
The CoDeveloper platform can help scientists develop products faster, she said. Adding fortifying ingredients with health claims to a product in development can lead to sensory off-notes, or unpleasant tastes, smells or textures.
“That’s something where AI can come in and help the formulation, and ideally take some of those steps out as to iterations you might otherwise go through,” she said.
For example, food scientists could prompt CoDeveloper to build a prebiotic yogurt, and the AI tool will fill in the gaps in terms of ingredients needed for the formulation.
“It’s going to look at our trusted resources and it’s going to pull a prebiotic that makes sense for yogurt,” she said. She added that CoDeveloper pulls exclusively from IFT’s 85-year-old database of peer-reviewed studies and does not connect to the internet.
Matching product with function
Developing new food and beverage products is a delicate balance that requires making sure its functional claims align with the product concept and technical feasibility, according to Leber.
“Not all products are equally suited for functional benefits, and successful innovation depends on choosing applications that feel intuitive and credible,” Leber said. “Certain categories such as yogurt with prebiotics or snack bars fortified with protein naturally align with consumer expectations in usage occasions.”
The full composition and structure of the food or beverage, known as the product matrix, interacts with functional ingredients and can reduce the bioavailability, or how much of the active ends up available for absorption and use in the body, according to Leber.
CoDeveloper can save food scientists hours of research simply determining whether the product’s shelf-life will outlast its functional properties, she said.
Sensory qualities matter
Mouthfeel, texture, aroma and appearance are all important considerations for food scientists.
“For example, fibers or proteins may create grittiness or thickness in some products where there otherwise might not be a noticeable texture change, and that’s something that is very off-putting to consumers,” Leber said.
CoDeveloper can more efficiently identify flavors to help mask metallic or bitter notes, she said.
“Techniques such as micro-encapsulation or emulsification can also minimize sensory impacts and improve the stability of bioactive,” she said. “To this end, ultimately, successful formulation requires understanding the target product and then anticipating what the sensory shifts might be and how consumers will perceive them, and working on those to ensure the product matches what the consumer ultimately expects.”
Leber said AI excels at calling out concerns in the sensory shift.
“CoDeveloper can really look at formulations and say, ‘Well, if you’ve added this much protein, you might have a shocking taste,’” she said.
