Current research positions AI self-efficacy as instrumental in shaping technopreneurs’ decision-making processes. However, the mechanisms governing how AI self-efficacy interacts with AI decision-making within understudied developing contexts remain underexplored. This study directly addresses this research lacuna by investigating how AI self-efficacy, interest, usage and perceived efficiency mediate the effects of technopreneurs’ AI self-efficacy on AI decision-making. Employing structural equation modelling and regression tests, we conducted a cross-sectional survey of 200 SME technopreneurs in Zimbabwe to explore these relationships. Thus, this research provides critical empirical insights into the technology acceptance pathways specific to technopreneurs in emerging economies. Technopreneurs’ AI self-efficacy was confirmed to shape their AI interest and usage; however, its connection to perceived self-efficacy remained unsubstantiated. Reliance on AI for business decisions was significantly influenced by three factors which include AI interest, AI usage and perceived AI efficiency, alongside self-efficacy. To boost self-efficacy, the study advocates grassroots AI training programmes and the formation of technopreneurs’ AI communities.