Listening to a conversation between two talkers and intending to remember the content later is a common situation. However, in cognitive-psychological research on verbal short-term memory, the standard stimulus material consists of unrelated words or isolated sentences. When connected sentences are employed in experimental studies, which is rather uncommon, diotic sound presentation is used without visual cues from the talkers. In the present studies, we investigated whether the spatial position of two talkers and static or animated faces as visual cues can reduce listening effort and improve memory of the conversation’s content. In a series of three experiments, participants completed a dual-task paradigm that comprised a primary listening task in which a conversation between two conversing talkers was presented. In Experiment 1 (n = 29), the audio signals of the two talkers were presented via headphones, either spatially separated (+/- 60°) or co-located (0°). In Experiment 2 (n = 30), static faces and in Experiment 3 (n = 28), animated faces were presented on computer screens and the corresponding audio signals were played through spatially separated (+/- 30°) loudspeakers. Parallel to the primary listening task, a vibrotactile pattern recognition task was administered as a secondary task to measure listening effort. In all experiments, we assessed working memory capacity, attention, and processing speed to account for person-specific factors, included as covariates, in the data analysis. We discuss our findings on the role of spatial auditory and visual cues on listening effort and short-term memory in conversations.