Artificial intelligence as a driver of innovation: Exploring the entrepreneurial ecosystem in Sudan

Abstract

Purpose: This study examines the associations between entrepreneurship education, innovative mindset, community support, and entrepreneurial innovation among young entrepreneurs in Sudan, with AI utilization as a mediating mechanism in a fragile, post-conflict institutional context.

Design/methodology/approach: A quantitative, cross-sectional design was employed using PLS-SEM with data from 404 young Sudanese entrepreneurs. Common method bias was assessed via Harman's single-factor test and collinearity via VIF diagnostics.

Findings: Community support exerts the strongest direct positive association with entrepreneurial innovation (β = 0.796) yet simultaneously acts as a digital deterrent through AI utilization (β = −0.040), reflecting a suppression pattern rooted in relational network embeddedness. Innovative mindset yields a significant direct effect (β = 0.256) and the largest indirect effect via AI utilization (β = 1.044), though the high inter-construct correlation (r = 0.982) warrants cautious interpretation. Entrepreneurship education shows a modest direct effect (β = 0.111) but non-significant mediation through AI utilization (β = −0.030, p = 0.098), signalling a critical gap between educational content and AI-mediated innovation.

Research limitations/implications: Cross-sectional design and convenience sampling limit causal inference. Future research should employ longitudinal designs and behaviorally anchored measures.

Practical implications: Fostering AI-enabled entrepreneurship in post-conflict settings requires simultaneous investment in digital infrastructure, curriculum reform, community norm change, and individual capacity building.

Social implications: Community social capital, while enabling conventional innovation, may constrain digital transformation in fragile economies.

Originality/value: This study demonstrates that AI's mediating role in entrepreneurial innovation is contingent and context-dependent, challenging frameworks derived from technologically mature settings.