Agentic AI refers to systems that show autonomous, goal-directed behavior in changing environments. Traditional models process inputs passively. Agentic systems act, plan, and adapt with minimal human intervention. Key traits include self-set goals, real-time planning, fast adaptation, and strategy updates based on feedback. LLMs are not agentic by default. They become agentic when placed inside agent frameworks that provide memory, explicit reasoning steps, tool usage, and links to external systems. These architectures let LLM-based agents solve complex tasks, refine their actions, and coordinate subtasks. Figure 8.1 shows an overview.