<p>AI governance frameworks designed for human decision-making fail when consequential outcomes emerge from coordination among machines. Current approaches create governance illusions—interfaces suggest control while algorithmic coordination unfolds beyond effective intervention. This article develops coordination transparency as a governance mechanism grounded in sociomaterial accounts of distributed agency. Instead of restoring centralized human control through better interfaces, coordination transparency targets agent-to-agent interactions directly through four components: interaction logging, live coordination monitoring, intervention hooks, and boundary conditions. This approach resolves a category error in prevailing frameworks that apply human-centered tools to distributed coordination processes. The framework shifts oversight from post hoc explanation of individual outputs to real-time observation and steering of coordination patterns where behavior actually emerges, preserving democratic accountability in systems characterized by distributed rather than centralized agency.</p>