Abstract: In an answer view, a first video segment is selected based on a first natural language input and displayed in a main display region, and a self-video image of a user is displayed in a peripheral display region having a smaller area than the main display region. To transition from the answer view to a question view, the self-video image is enlarged to replace the first video segment in the main display region. A second natural language input is received. To transition from the question view to the answer view, the self-video image is reduced to occupy the peripheral display region and the self-video image is replaced in the main display region with a second video segment selected based on the second natural language input. The video segments are pre-recorded video response segments spoken by the same person. Enlarging the self-video image masks the transition between the video segments.