AI agents are reshaping academic work and assessment: learning‑management vendors say detection is unreliable, while AI models flag which campus jobs are most exposed. Blackboard executives told educators that reliably detecting agentic AI in coursework is currently impossible, pushing colleges to redesign assessments rather than chase detection technologies. At the same time, analyses from AI models identify entry‑level and administrative roles in higher education as vulnerable to automation—putting pressure on staffing models, professional development and career pathways for early‑career academic workers. Campus leaders are weighing two responses: redesign assessment to require authentic, scaffolded student work and invest in retraining and role redesign to protect institutional capacity. Both approaches require investment and clear policy frameworks around acceptable AI use.