Enabling temporal transitions: AI adoption and sequential ambidexterity in Chinese SMEs from 2015–2023

With research about sequential ambidexterity suggesting that firms may approximate ambidexterity over time by alternating between periods of exploration and exploitation, this temporal approach provides a feasible solution for SMEs facing resource limitations to manage the tension between exploration and exploitation. However, little is known about what enables SMEs to successfully enact such temporal transitions over time. In this study, we examine whether and when AI adoption facilitates SMEs’ ability to achieve sequential ambidexterity. By analysing longitudinal data from firms listed on the Small and Medium Enterprise Board of the Shenzhen Stock Exchange in China from 2015 to 2023 and using text analysis, topic modelling approach, and fixed-effects regressions, the findings demonstrate that AI adoption enables firms to dynamically alternate between periods of exploration and exploitation over time. Furthermore, AI learning capability and AI interaction capability strengthen this positive relationship. Our study provides a rich empirical foundation for future research aimed at understanding the dynamics of AI adoption and its impact on ambidexterity within firms.