Whether artificial intelligence can promote service-oriented manufacturing is a widespread concern in both theoretical and practical fields. By using text analysis in machine learning to measure the degree of service-oriented manufacturing for the first time， it conducted an empirical study on the relationship between artificial intelligence and service-oriented manufacturing， based on a data sample of Chinese A-share listed companies from 2008 to 2022. The result indicates that there is no evidence of “AI paradox” in the process of servitization， namely， artificial intelligence is of significantly positive importance to service-oriented manufacturing. It reveals that artificial intelligence capabilities reduce labour skill premium and augment digital innovation ability， thereby facilitate service-oriented manufacturing through cost reduction and efficiency improvement. These mechanisms illustrate a dynamic evolution path of “artificial intelligence→dynamic capabilities→cost reduction and efficiency improvement→service-oriented manufacturing”. Heterogeneous test results show artificial intelligence increases the level of service-oriented manufacturing in competitive industries and policy-oriented servitization more significantly. The research findings offer micro-level evidence of the intrinsic link between artificial intelligence and service-oriented manufacturing， providing significant theoretical and practical insights for the high-quality transformation and upgrading of Chinese enterprises in the era of digital intelligence.