Co-authored by Ceolla Dillon O’Rourke, Aoife Ní Chíobháin, and Michael Hogan. In education, collaborative learning is valued as a way to build the key collective intelligence skills that societies depend on: coordinating, communicating, sharing knowledge, perspective taking, and solving problems together. However, collaborative learning is not just about putting students in a group and asking them to complete a task. A group also has to manage the process of thinking together: taking turns, noticing confusion, regulating frustration, repairing misunderstandings, and deciding when to speak and when to wait. These social and embodied processes matter because collective intelligence (CI) depends not only on individual ability, but also on how well group members notice and respond to one another. Woolley et al. (2010) found that groups have higher CI when members show greater social sensitivity and share conversational turns more evenly. Successful collaboration depends partly on reading the room. Researchers are increasingly using AI and multimodal learning analytics to make these subtle processes visible (Järvelä et al., 2020). Systems can detect speech patterns, pauses, facial expressions, and …