Toward a Unified Understanding of AI Capabilities: Definitions, Operationalization, and Mechanisms to Firm Performance

This study explores how Artificial Intelligence (AI) capabilities contribute to firm performance through a systematic scoping review following PRISMA guidelines and the Arksey & O'Malley (2005) framework. Thirteen empirical and theoretical studies were reviewed from Scopus, Web of Science and IEEE Xplore to clarify how AI capabilities are defined, operationalized, and linked to performance outcomes. Findings reveal conceptual fragmentation but a convergence around three core dimensions: tangible, human, and intangible resources enabling firms to select, orchestrate, and leverage AI-specific assets for value creation. Grounded mainly in the Dynamic Capabilities View (DCV) and the Resource-Based View (RBV), literature highlights mediating mechanisms such as innovation, adaptability, resilience, agility, and learning. Moderators including environmental dynamism, cultural distance, organizational formalization, and AI-oriented leadership shape these relationships. The study proposes a unified conceptual model linking AI capabilities to multiple forms of performance, offering theoretical advancement and actionable insights for digital transformation strategies.