Issue link: https://hub.uberflip.com/i/1359949
69 Let's take a look at the graph below. At right, you can see the percentage of click-through rates on content items presented as recommendations to users, as well as the two types of recommendations—non-personalized and personalized. TOP VISITED ITEMS BOMBORA NLP-BASED RECOMMENDATIONS NON-PERSONALIZED CLICK-THROUGH RATE PERSONALIZED 0 0.5% 1.0% 1.5% 2.0% INSIGHTS ENGAGEMENT ARTIFICIAL INTELLIGENCE The left bar shows the click-through rate of a simple, non-per- sonalized recommender. This recommender doesn't use AI, but instead presents to the user the most viewed items in a Content Hub 12 . By contrast, the right bar shows the click-through rate of a personalized recommender. Using AI, this personalized recommender "learns" from the behavior of visitors with similar interests. It relies on a combination of natural language process- (NLP) ad v e daa (e, by e e a our visitors build as they consume content elsewhere online) from a platform called Bombora.