In this paper we describe our approach to allow an artificial companion to read texts in an engaging way, and its evaluation. This approach consists in annotating the text with extra-information needed to personalise its narration and make it more interactive; the markup language is generic, i.e. it can be used to annotate any text. However we here focus on story narration to children, and identify narrative strategies tailored for this specific public and context. We give a quick overview of our ongoing implementation of a storytelling module in the Reeti expressive robot, before describing the results of two evaluations. One study concerns the acceptability of different roles of a companion robot, including storytelling, while the other concerns more specifically the believability and engagingness of our strategies.