Purpose: The purpose of this paper is describing the seamless integration of the question-answer interaction into automatic lecture recordings (ALRs). This includes the design and implementation of the question management (QM) software for a virtual camera team. Design/methodology/approach: Coming from the human role model the interaction and its management to the virtual world is transferred and integrated it into a virtual camera team. All events are translated into sensor inputs which get processed by the virtual director and are used for the collaboration of the team in order to implement more complex cinematographic rules. Findings: It is found that it is possible to record the whole interaction, to record the original voice of the questioner out of an audience without handing out a microphone or forcing him/her to walk to one, and to record a video of the questioner while asking. So, it is easier to follow a lecture recording as more details are recorded automatically. Practical implications: First experiences on using this software clearly show the small weaknesses of the first version. As mentioned in the outlook, these are currently being addressed, e.g. by looking for a more natural interface. Originality/value: The paper demonstrates how to use the question-answer interaction as sensor input for an automatic lecture recording (ALR) system based on the roles of the according human originals. It ensures that many details of a lecture can be recorded seamlessly to keep the lecture context continuous and therefore to make the lecture recording more vivid and interesting.