This is the dilemma legal teams face: they want the advantages of AI, but they cannot risk undermining the very defensibility their work depends on. And that’s where the distinction between generative AI and agentic AI becomes critical.