Harnessing AI ambidexterity for competitive advantage: the role of dynamic capabilities in digital innovation ecosystems

Purpose Drawing on the theoretical framework of dynamic capabilities, this study examines how small and medium-sized enterprises (SMEs) leverage digital innovation ecosystems (DIEs) to achieve artificial intelligence (AI) ambidexterity, enabling them to sense market opportunities, effectively seize them and transform organizational processes for sustained performance improvement. Design/methodology/approach Data were collected via a structured survey employing validated scales, targeting 102 firms within a European Digital Innovation Hub (EDIH). A partial least squares structural equation modeling (PLS-SEM) approach was applied to assess the impact of AI ambidexterity on dynamic capabilities (i.e. sensing, seizing and transforming) and SME performance. Findings The study reveals that participation in EDIHs – intended as a practical example of DIEs – facilitates a balanced approach to exploring innovative AI solutions and exploiting existing capabilities, resulting in superior performance outcomes through the mediated role of dynamic capabilities. Research limitations/implications This study demonstrates how AI ambidexterity fosters dynamic capabilities, enhancing firm performance within DIEs. Aligning with previous findings, it highlights the role of DIEs in facilitating resource sharing and orchestrating routines, while dynamic capabilities mediate the AI–performance link, emphasizing AI’s dual role in operational efficiency and innovation for competitive advantage. Practical implications By fostering an environment favorable to AI ambidexterity, EDIHs can help managers overcome barriers to advanced technology adoption, particularly for SMEs. This includes providing access to cutting-edge technologies, expert support and collaborative opportunities that lower the costs and risks associated with AI-driven innovation. Originality/value This study contributes to the literature by providing empirical evidence on the relationship between AI ambidexterity, dynamic capabilities and performance within DIEs.