According to Ian Mealey, product marketing director – formulation at Format Solutions, which provides formulation and feed ERP software, AI in the formulation space is still finding its place in pet food.
“So far, we are not seeing many definite instances or requests in relation to the application of AI in our field, and those managing formulas for pet food and livestock feed producers are probably slow adopters,” said Mealey. “Nevertheless, in the formulation field, there are some applications where we believe AI may well develop:
- Understanding the impact of ingredient quality, cost and availability: These may occur simultaneously and individually or together can affect the finished product. AI could be applied to flag to product developers and nutritionists when impacts are beyond acceptable criteria and prompt them to review and perhaps change recipes to maintain quality and cost objectives.
- Helping check and ensure compliance of proposed recipes and their associated declarations — This area often falls under the remit of product developers to ensure that proposed recipes and associated declarations are full and compliant. This requires a lot of knowledge and encapsulates many interactions and complexities. Rules-based applications to check recipes and generate required legal declaration text based on recipe content already exist. AI may well play a role in extending these capabilities to learn from responses to failures, automating formula change tracking and checking compliance of formula and declaration to target market legislations.
- Capturing feedback on recipes and their performance, whether related to production criteria such as finished product quality, customer acceptance, animal ‘performance,’ etc. Individual recipes may meet product specifications but be unsatisfactory or even unacceptable for other reasons. AI could conceivably check proposed recipes for features, such as ingredient combinations, which may not be obvious to the product developer or nutritionist, but which may cause issues, based on previous experience.
- Identifying new product opportunities. One could argue this is perhaps more a marketing function, but it directly relates to the creation of product specifications and the recipes that result. One could foresee an iterative process of technical recipe development related to identified opportunities. This process happens today but will AI technology be applied to accelerate it?”
Of course, it’s not likely that AI will ever truly completely replace the human eye in formulation.
“Given the risk factors involved, and the general pace of recipe change in pet food compared to the livestock industry, perhaps the impetus for AI application will lie elsewhere,” said Mealey. “However, the above examples address complex areas which rely on the knowledge of highly skilled individuals and are candidates for automation and assistance from AI. Well-designed software solutions exist already for many of these areas and they continue to provide valuable solutions for customers. I wonder, therefore, whether AI, due to the data it requires and time taken to ‘learn,’ will practically replace good software solutions that could be viewed as more ‘traditional.’ Perhaps it’s the automation and encapsulation of knowledge and skills that is the key element?”
Read more about AI in pet food manufacturing:
AI has entered the pet food chat: Where will it land?