Although systems based on artificial intelligence (AI) can collaborate with humans on various complex tasks, little is known about how AI systems can successfully collaborate with human teams (team-AI collaboration). Team performance research states that team composition and intragroup processes are important predictors of team performance. However, it is not clear how intragroup processes differ in team-AI collaboration from human teams and if this is reflected in differences in team performance. To answer these questions, we synthesize evidence from 18 empirical articles. Results indicate that intragroup processes like communication and coordination are less effective in team-AI collaboration. Moreover, whether team cognition and trust are higher in team-AI collaboration compared to human teams is not clear, since studies find conflicting results. Likewise, the results on team performance differences between team-AI collaboration and human teams are inconsistent. With this article we offer a foundation for future research on team-AI collaboration.