As AI systems evolve from tools to autonomous actors, the stakes around oversight, safety and accountability intensify. This roundtable examines how organisations can define meaningful human control, set credible safety thresholds and shape practical approaches to agentic behaviour before autonomy outpaces governance.