Previous studies of narrative development based on wordless picture stories indicate that before 7–8 years most children provide descriptive narratives with little inferential content such as explanations and attribution of mental states to the story characters. These components find greater expression in studies where children participated in conversations focused on the causes of the events. In the present study, 84 French-speaking children, from kindergarten to second grade, narrated the Stone story (a wordless five-picture story whose plot is based on a misunderstanding between two characters) before and after a short conversational intervention (SCI) focused on the causes of the events, as well as one week later when they also narrated a new story. Thirty additional children served as the Control group: instead of the SCI they played a Memory game with a set of cards containing the pictures of the Stone story. Children in the SCI group increased the inferential content of the narrative produced after the SCI, thus confirming with a larger sample findings obtained in previous studies. Moreover, results provide new evidence that the immediate improvements in inferential content were still present after a week’s delay and could also be applied to a new story. All narratives produced after the SCI were also longer and contained more markers of causality. The effect was stronger in first and second graders than in kindergarten children. By contrast, no significant improvements were found in the children of the Control group on any of the measures. Such results highlight the effectiveness of the SCI in promoting children’s narrative skills, its usefulness in their assessment, and have important implications for a better understanding of narrative development.