This chapter argues that the question of whether AI is truly social misframes the phenomenon at stake. Across the cases of Beau, Areca, and ShamAIn, social experience does not begin with belief, attribution, or recognition. It begins with a response. People act, wait, adjust, and continue interacting and only later describe what has already taken shape as social. The chapter therefore treats sociality not as a property of an entity, but as an outcome produced through repeated interaction. In these cases, social effects did not depend on whether AI possessed intention, consciousness, or understanding. They emerged through the stabilization of attention, expectation, temporal rhythm, and responsibility. Debates centered on authenticity or deception are thus insufficient, not because they are wrong, but because they overlook how interaction persists in practice under acknowledged uncertainty. Drawing on relational perspectives from Science and Technology Studies (STS), the chapter reframes the social actor as a position that emerges within relations rather than as a preexisting subject. This move does not collapse the distinction between humans and AI. Instead, it shifts analytic attention toward how relations are organized and how effects such as intimacy, hierarchy, care, and authority are produced in specific contexts. From this perspective, design concerns the conditions under which relations form and endure, rather than attributing social qualities to systems. The three cases are presented not as proof that AI belongs in society, but as sites for examining how sociality is configured, stabilized, and unsettled.