Advancements in AI, particularly generative models, have facilitated the rise of agentic AI systems, which operate with increasing autonomy, make independent decisions, adapt to dynamic contexts, and proactively influence outcomes. These systems bring distinct considerations and challenges to organizational contexts, distinguishing them from traditional, passive AI systems. This paper examines these unique challenges using a three-pronged qualitative methodology comprising a systematic literature review, expert interviews, and an analysis of professional publications. Through qualitative coding, we delve into the dimensions of agentic AI challenges, focusing on the complexities of their integration into organizations. Our findings reveal challenges in the core human-agentic AI relationship as well as exogenous and internal organizational factors. Our findings offer valuable insights into these complexities and establish a foundation for actionable strategies in future research.