The potential of synchronous technology-mediated oral interaction for second language acquisition by young learners is of interest to researchers and classroom teachers alike. However, several publications underline the lack of research concerning telecollaborative interaction with young learners (Pérez & Vinagre, 2007; Dooly & Vinagre, 2022). In addition, as Pennock-Speck & Clavel-Arroitia (2023) highlight, existing studies are often limited to descriptions of the design and implementation of learning scenarios, and focus on learner motivation rather than language learning (Ligorio and Van Veen, 2006; García-Martínez and Gracia-Téllez 2018; Mont & Masats, 2018).Only a handful of studies have examined potential learning outcomes and participant roles in achieving these in interaction (Austin et al. 2017; Whyte & Wigham, 2022). In this small body of research on learner-learner interaction, some have taken a multimodal approach to focus on micro-analyses of short interaction sequences (Austin et al 2017) while others have taken a quantitative approach to conversation analysis (Pennock-Speck & Clavel-Arroitia 2023). Our work contributes to this important strand of young learner research by focusing on live oral interaction as a potential locus of L2 acquisition, and our ongoing research highlights the influence of both task design and teacher engagement in creating opportunities for language learning in synchronous video communication (Whyte & Wigham, 2022). The present study examines task implementation for an information-gap task conducted by different teachers in two primary school telecollaboration projects (CEFR A1). The “who’s who?” task involved identifying pupils in the remote class by through question and answer sessions using identity cards created and exchanged in a pre-task phase. The linguistic objectives were for learners to ask and answer questions about themselves and the task was conducted via videoconferencing. Data include hors-champ video recordings of task interaction as well as post-task interviews with learners and teachers. The research questions for the present study are as follows: 1. How is learner autonomy manifested in task-as-process? 2. What is the role of the teacher in creating opportunities for learner autonomy? To address these questions, we adopt a mixed methods approach, combining quantitative analysis of verbal and co-verbal actions with fine-grained multimodal (inter)action analysis of illustrative episodes (Norris, 2019). The interaction data recordings (30min and 55min) were transcribed and imported into ELAN (Sloetjes & Wittenburg, 2008). Analysis included annotated episodes when the learners depended on the teacher, took initiative, asked and answered questions, used formulaic language, and introduced new constructions. Our presentation will first give an overview of the two task instantiations. We then analyse selected episodes to illustrate different levels of learner autonomy as manifested in task-as-process, showing how teacher role both in task-as-workplan and task-in-process (Breen, 1989) offers and/or constrains opportunities for learner autonomy. We conclude with reflections for practitioners on guiding young learners during synchronous oral telecollaboration tasks.