Engagement has been discussed in relation to the effort invested during task accomplishment (Bygate & Samuda, 2009; Helme & Clarke, 2001; Skinner & Pitzer, 2012). Recent studies have been framed by social constructivist views of task processing (Pekrun & Linnenbrink-Garcia, 2012; Philp & Duchesene, 2016) and have approached task engagement more holistically as the effort invested in accomplishing a task, guided by participants’ requirements of task success and involving behavioral, cognitive and attitudinal qualities in interaction with each other and for which the social element permeates the aforementioned three dimensions (Gijsen, 2021). This study addresses the task performance parameter of partner orientation: the effort participants invest in establishing and deepening a relationship with partners to strengthen interaction (Gijsen, 2021). We adopt multimodal (inter)action analysis (Norris, 2004) to investigate how different semiotic resources contribute to partner orientation. Our context is a virtual exchange within the E-LIVE project (Engaging Languages in Intercultural Virtual Exchange) involving 48 trainee language teachers from French, Dutch, and Colombian institutions. Our analysis focuses on the first synchronous interaction via videoconferencing in which teaching experiences were discussed. This followed an asynchronous language biography task. Using multimodal transcriptions from a subset of interactions, we examine three dimensions of partner orientation: The behavioral dimension with performance indicators of providing and accepting support/feedback, giving compliments, using comprehension checks/ production prompts; the attitudinal dimension through comments on working together, task ownership, technological tools used; and the cognitive dimension explored through reflective blogs and questionnaire/focus group data during which participants reflected on the relevance of strengthening the interaction with their partners, the effort they made to reach mutual understanding and address intercultural aspects. We anticipate study findings will advance our understanding of task engagement and help identify effective strategies to maximise engagement, and thus learning, in virtual exchange projects.