Abstract: In an approach for selectively placing an advertisement in a media item to maximize user engagement, a processor analyzes the media item by performing a static and dynamic scan. A processor determines whether it is more likely the user will engage with an advertisement placed in a foreground section or a background section of the media item. Responsive to determining it is more likely the user engages with the advertisement placed in the foreground section of the media item, a processor loads a set of time intervals associated with the peak regions of the media item. Responsive to determining the level of engagement by the user with the potential advertisement exceeds the threshold when the potential advertisement is associated with the media item, a processor associates the advertisement with the media item. A processor dynamically changes a background music of the advertisement with a sound extracted from the media item.