In this paper, we address challenges arising from the use of the conversational interface Siri during a literacy activity in a learning context. On one hand, we investigate potentialities of Siri as a virtual assistant and knowledge navigator for task accomplishment. Thus, we attempt to suggest possible lines of mobilizing Siri to afford and improve both students' task elaboration and second language performing during a literacy activity. In human-Siri interactions, a wide range of operations are carried out through closely intertwined oral and written instances of natural language (vocal and visual accounts of 'understanding'). On the other hand, we raise some crucial, critical questions with regard to concrete situations involving Siri as a 'learning' assistant. How do students deal with the unpredictability of the interactions with the virtual conversational agent? How do students handle interaction modalities such as identifying and 'well' pronouncing the right commands and the choice of words that activate the diverse features? Our paper seeks to provide Conversation Analysis (CA) [1] informed insights into human-Siri communication in a learning context.