How do governments use political communication to facilitate policy reversals of major significance? Dr Yan Xiaojun will discuss how the Chinese state dynamically uses political communication to accomplish rapid discursive shifts in the context of major policy changes. Using a structural topic model for unsupervised machine learning to analyze extensive propaganda materials, this research explores how the Chinese Communist Party facilitated the reversal of its One-Child Policy in 2015 by deploying a systematic, precise, multidimensional, and multi- layered discursive engineering project. Instead of a monolithic and rigid operation, China’s state-directed political communication is a strategic and flexible discursive engineering project tailored to contingent factors, such as the nature of the policy being advocated, the level of the propaganda venue, and the needs, mentality, and tastes of carefully targeted audiences. This coordinated maneuver of state narrative is key for the effective and smooth policy reversal of major significance.