Colleges are piloting agentic AI that automates student services and learning supports. Campus experiments range from autonomous agents that manage administrative workflows to a student‑built tool from University of Maryland alumni that teaches case‑study responses and aims to reduce cheating by coaching critical thinking. Administrators highlight efficiency gains and potential to scale advising and tutoring, while faculty and privacy advocates raise questions about accuracy, academic integrity, and data governance. Early pilots show promise for routine tasks but also reveal governance and readiness gaps before agents can be trusted with high‑stakes student outcomes.