Organisational AI embeddedness and innovation quality under environmental uncertainty: evidence from Chinese strategic emerging industries

Artificial intelligence (AI) is expected to improve innovation, yet its innovation-quality (IQ) returns remain uneven, particularly in strategic emerging industries marked by technological complexity and environmental uncertainty. To explain this heterogeneity, this study moves beyond AI exposure and examines organisational AI embeddedness (AIE), defined as the integration of AI into strategic attention, technological resources, and business-process routines. Using 9,367 firm-year observations from Chinese A-share firms in strategic emerging industries during 2011–2023, this study constructs a formative AIE index and combines two-way fixed-effects models with fuzzy-set qualitative comparative analysis (fsQCA). AIE is positively associated with IQ. Parallel mediation estimates reveal positive indirect associations through sensing, absorptive, and reconfiguring capabilities, with the indirect association through reconfiguring capability statistically larger than the other two. Market and policy uncertainty strengthen the positive AIE-IQ association. fsQCA identifies multiple high-IQ pathways, with AIE enabling some but not all configurations. These findings explain heterogeneous AI-enabled innovation outcomes through organisational embedding, capability conversion, and environmental contingency.