An experiment with 24 subjects was performed. The subjects were split in three groups, and asked to extract information from the AdApt dialogue system for somewhat over 30 minutes per subject. The system configuration varied in that one group had turn-taking gestures from an animated talking head, another had an hourglass symbol to signal when the system was busy, and the third had no turn-taking feedback at all. The results show that although the hourglass setup showed no decrease in efficiency compared to the facial gestures, it made the subjects less satisfied. The lack of turn-taking feedback was noticed and mentioned by half of the subjects in that group.