Barilla is an Italian multinational company active in the food sector, one of the most important internationally recognized for the quality of its products. In particular, it is one of the main international players in the pasta sector, making it one of the most successful brands and testimonials of the Italian Way in the world.
Where Barilla is, there is home
Barilla's imagination is closely linked to that of the home. Cooking at home transmits a series of positive sensations: homemade, genuine and "artisanal" food, all sensations that make the Barilla brand perceive as part of everyday life and of one's family.
Unfortunately, however, in some cases cooking a good dish can be less simple, especially if you are not sure what ingredients are needed. Barilla's need was precisely to devise a system that would inform users on the feasibility of cooking a dish, using an innovative and smart approach.
A social support active 24/7, but smart!
To solve this need, we have thought especially, but not limited to, a target of young off-site who approach their first experience away from the family home. It often happens that you go home without having had the opportunity to go to a supermarket and having to make a virtue of necessity.
For this purpose we have created #SOSPASTA, a chatbot that can be used on the Barilla Twitter social page: by publishing posts using hashtags to send the ingredients available to them, the chatbot automatically replied by suggesting good recipes.
The big challenge was to create a database of keywords and ingredients linked to certain recipes that offered relevant links when these keywords were requested. Furthermore, the algorithm took into account the user's preferences over time, and suggested personalized recipes based on their personal tastes.
The system was also designed to generate an archive that, by analyzing the requests received, would collect the ingredients most used by the community in order to subsequently be able to offer further relevant and quality recipes and create aggregate statistical data to identify the most used ingredients based on specific parameters, such as certain geographic areas.