Researchers from UC Davis, including IIFH-affiliated faculty member Ilias Tagkopoulos, have demonstrated how artificial intelligence can help predict consumer preferences for coffee using a surprisingly small set of sensory inputs. Published in npj Science of Food, the study combined machine learning, sensory science, and consumer data to identify the key factors that drive liking of black drip coffee. Analyzing responses from 118 consumers across 27 coffee preparations, the team found that perceived acidity, flavor intensity, and sweetness were the strongest predictors of consumer acceptance. The resulting AI models were able to predict whether consumers would like or dislike a coffee sample with strong accuracy, even when using only a few sensory measures.
The work highlights the potential of AI to transform how food and beverage products are developed and optimized. Beyond identifying drivers of liking, the researchers introduced a novel approach for uncovering distinct consumer preference segments, revealing that different groups value different sensory characteristics and even perceive the same product differently. For UC Davis and the Innovation Institute for Food and Health, the research exemplifies how advanced data science can bridge fundamental food research and commercial application—providing food companies with new tools to better understand consumers, guide product innovation, and accelerate evidence-based decision-making across the food system.
