Institutional Voids and AI Adoption: Evidence from South Africa and Kenya

This study examines how institutional voids shape firms’ adoption of artificial intelligence (AI) in South Africa and Kenya. Institutional voids in capital, labor, regulatory, and product markets can both constrain and enable technology adoption. Using a survey of 192 decision-makers, we apply partial least squares structural equation modeling (PLS-SEM) to analyze the influence of each void on AI adoption. The findings show that lower capital and product market voids consistently enhance adoption, confirming their role as foundational conditions for technology deployment. In contrast, labor and regulatory voids do not exhibit significant effects in the full sample. However, subgroup analyses reveal important heterogeneity: women and early-career professionals are more likely to perceive the reduction of labor and regulatory voids as enabling AI adoption. The study advances institutional theory by distinguishing between foundational voids that must be reduced and desirable voids that can foster experimentation. Policy implications highlight the need for context-sensitive reforms to support inclusive AI adoption.