Recommender systems emerge as a way for companies to anticipate customer preferences and needs. This is possible with Artificial Intelligence using various data sources that must contain at least the entity to which the recommendation is provided (customer), and the product to be recommended (item). The common principle of all recommendations is that there are important dependencies between customer-centred and item-centred activity.
The recommender analysis is often based on previous interaction between clients and items, and/or recommendations based on knowledge of requirements specified (directly or indirectly) by the client or a group of clients. In the first case, past interests and trends are often good indicators of future options, while in the second case the categories of items may show significant correlations that can be exploited to make more precise recommendations.
In addition, recommender systems are continuously calibrated according to customer preferences, resulting in more and more customers being retained over time and showing relevant items tailored to each customer. In other words, they improve the customer experience and increase sales.
