Techniques are described for delivery and handling of automated user-responsive video sub-content. For example, a user is consuming video content associated with content-of-interest (COI) identifiers corresponding to certain sub-content relevant to the video content during certain playback time windows. During playback, embodiments use one or more sensors detect dynamic user input. At some time during playback of the video content, some of the COI identifiers are considered to be actively relevant, and detected dynamic user input is determined to invoke at least one of the actively relevant COI identifiers. A COI query can automatically be generated based on the invoked one or more COI identifiers, and a corresponding query response can be automatically generated and output to the user based on response data associated with the invoked one or more COI identifiers.