Generative Ai Adoption and Innovation Capability in Emerging-Market Startups: A Dynamic Capabilities Perspective

Abstract

Artificial intelligence adoption is accelerating, yet its innovation consequences remain uneven and theoretically under-specified. This study extends capability-based research by asking: What is the influence of generative artificial intelligence (AI) capability on innovation performance, as mediated by generative dynamic capabilities and moderated by absorptive capacity and institutional support?  Drawing on dynamic capability theory, it has been applied covariance-based structural equation modelling (SEM) to survey data from 384 tech-savvy startups in an emerging economy.  The results show a strong positive impact of generative AI capability on generative dynamic capabilities, which in turn have a significant impact on innovation capability.  A mediation analysis conducted using 5000 bootstrap resamples show that more than 50% of the total causal effect of AI capability on innovation is mediated through the processes of sensing, seizing, and reconfiguring. Furthermore, results from moderated mediation indicate that the indirect effect of this relationship is significantly increased by absorptive capacity and institutional support, thereby explaining institutional boundary conditions in the process of building AI-enabled capabilities.  Comparative model evaluation confirms the superiority of the proposed mediated framework over the alternative direct-effect framework. By making the generative artificial intelligence capability distinct from generic digital capabilities and empirically portraying the generative era's capability transformation pathway, the study refines dynamic capability theory for the generative era. These results provide theoretically grounded, managerially actionable insights for companies and policymakers striving to achieve sustainable innovation outcomes from successful AI investments.