At APQC’s 2025 KM Conference, I challenged attendees to stop treating AI like magic and start managing it like knowledge. From opaque guardrails to missing metadata, AI systems are increasingly embedded in decision-making—but without KM, they drift into ungoverned complexity. This piece outlines how knowledge professionals must adapt their tools and frameworks to audit prompts, track model changes, align context windows, and govern knowledge flow in retrieval systems. AI isn’t just another tool—it’s a reshaping of the knowledge landscape. KM must now manage not just content, but the systems that generate, retrieve, and interpret it.