The dairy industry is one of the largest food industries and is estimated with annual revenue of hundreds of billions of dollars. Cow milk composition varies between individual cows, during lactation period, and due to physical and environmental conditions. For various dairy products (such as: fluid milk, hard and soft cheeses, yogurts and dairy desserts) different milk compositions are required depending on the properties of the final product and its production process. Yet, to date, no prescreening of the milk according to the target product is being done. This project is part of the “Food Big Data-IOT” consortium, in a work package aimed at developing an AI-based tool that will allow prediction of milk functionality from the genetics stage of cows, by sorting cows according to the suitability of their milk, and during milking, to enable efficient routing of milk to final products of three main categories: hard cheeses, acid cheese products and milk drinks/desserts. Within this project, we have thus far performed comprehensive characterization of hundreds of raw milk samples, both for the composition of the milk and for its functionality during acid-curd formation, from individual cows and from herds, for the purpose of creating a broad database. We used the database to build a prediction model using multi-parameter linear regression. The total protein concentration has been found to have the most significant effect in all the models developed. Average casein micelle diameter, acidity, and lactose concentration have been found to be the second most significant factors affecting yield, gelation time, and curd firmness, respectively. In parallel, we collaborate with the team of Prof Avi Gal in developing an AI based model.