Research on students suggests that repeated feedback in simulated investigative interviews with avatars (computerized children) improves the quality of the interviews conducted in this simulated environment. It remains unclear whether also professional groups (psychologists) benefit from the training and if the effects obtained in the simulated interviews transfer into interviews with real children who have witnessed a mock event. We trained 40 psychologists (Study I) and 69 psychology students (Study II). In both studies, half of the participants received no feedback (control group) while the other half received feedback (experimental group) on their performance during repeated interviews with avatars. Each participant then interviewed two 4-6-year-old children who had each witnessed a different mock event without any feedback being provided. In both studies, interview quality improved in the feedback (vs. control) group during the training session with avatars. The analyses of transfer effects showed that, compared to controls, interview quality was better in the experimental group. More recommended questions were used in both studies, and more correct details were elicited from the children in Study I, during the interviews each participant conducted with two children (N = 76 in Study I; N = 116 in Study II) one week after the training. Although the two studies did not show statistically significant training effects for all investigated variables, we conclude that interview quality can be improved using avatar training and that there is transfer into actual interviews with children at least in the use of recommended questions.