I keep hearing about "Agentic AI" in education. But what does that even mean? I wanted to explore the pedagogical decisions we think the AI is making (or we're hoping it's making), what infrastructure and capabilities AI might need, and whether those decisions actually might help people learn. I took the 26 research-backed learning principles from ABCs of How We Learn and mapped each one to a bot concept, then asked: does this thing actually need agentic infrastructure, or is it just a good prompt?