Discovery's deep content understanding technologies have led to a surge in users watching content with their favorite characters. VK's AI engineers have developed a technology that uses two machine learning models. One model analyzes video frames to identify popular personalities, while the second model recognizes these personalities among the identified ones. This personalized recommendation system has improved user experience by providing more relevant content recommendations.
As a result of these advancements, viewing time for content featuring favorite characters from the "Watch also" section has increased by 10%. This increase in engagement has made the content feed noticeably more tailored and appealing to users.