The growing integration of artificial intelligence (AI) in team decision-making processes has raised important questions about how to effectively implement these technologies to maximize team performance. While AI systems can augment human cognition and decision-making capabilities, little is known about how the timing of AI assistance impacts team dynamics and outcomes. Through a series of three online experiments, this study investigates whether introducing AI assistance at different stages of team collaboration leads to different outcomes in team decision-making. Across all three studies, the findings consistently demonstrate that the timing of AI assistance significantly influences team performance. Teams receiving AI support during the initial stages of collaboration outperformed those receiving assistance in later stages. Early AI assistance generated a ripple effect that enhanced cognitive diversity and information elaboration among team members, leading to improved decision-making outcomes. When AI was introduced in later stages of team collaboration, reflexivity interventions proved effective in enhancing team performance, suggesting potential strategies for implementing AI in established teams. These findings suggest that organizations should carefully consider the timing of AI implementation in team settings, particularly favoring early integration during team formation, while implementing appropriate interventions when introducing AI to established teams.
