ABSTRACTBackgroundCollaborative problem‐solving (CPS), encompassing social and cognitive domains, is recognised as a critical competency for current and future learners. While the cognitive domain, rooted in individual problem‐solving, is relatively well understood, the social domain, manifested through social interaction, remains a challenge, especially in remote settings.ObjectivesThis study explored the complex social domain of CPS through theoretical and empirical understanding of joint attention (JA) behaviour. JA was conceptualised as a continuum of attentional levels, from lower (monitoring and common) to higher levels (mutual and shared). While gaze alignment was considered foundational to JA, capturing higher attentional levels, crucial for productive CPS, required extending the focus to communication over the shared objects of attention—were they external (e.g., events) or internal (e.g., thoughts).MethodsDual eye‐tracking data from 12 dyads (aged 12–13) engaged in remote CPS were first analysed for the strength of eye‐gaze coupling, used as a proxy for lower JA levels and a potential indicator of higher levels. This was followed by qualitative analysis of interaction (logfiles, eye‐event videos) in three dyads with strong eye‐gaze coupling, to examine how these measures related to actualised JA.Results and ConclusionsThe results highlighted the relevance of JA in understanding the social domain of CPS, even in remote settings. Informed by quantitative measures, the qualitative analysis revealed dyadic behaviour linked to higher JA levels. Overall, the findings provide preliminary insights into how JA relates to productive CPS, offering a groundwork for further research on supporting learners to “improve collectively”.