Imagine an interactive map where each product appears as a dot. The position of each dot is determined by its characteristics – i.e. color, size, price and many more. The embedding model works like an intelligent cartographer that arranges these dots in such a way that relations and similarities between them become visible. 

Now, a product is no longer only an individual dot, but part of a larger network. Due to the distances between the dots, you can see what products are similar to each other. As a consequence, personalized recommendations can be given. However, the capabilities of embedding models go far beyond product data. They can be applied to almost all data types – from images and tracking data to user analyses. This means that new ways open up to better understand customers, optimize processes and develop innovative solutions.