Building AI capability from digital orientation: the roles of organizational learning, organizational forgetting, and trust in AI

Abstract

Despite the growing importance of artificial intelligence in business contexts, empirical research on how firms transform digital orientation into AI capability remains limited, particularly when many firms invest in digital and AI initiatives but fail to develop a coherent capability base. Based on the attention based view, this study proposes a framework with mediation and moderation to explain the direct and indirect effects of digital orientation on AI capability, mediated by organizational learning and organizational forgetting, and moderated by the level of AI trust within the firm. We carried out a quantitative study using survey data collected in 2025 from 306 Chinese firms across different industries, ownership types and firm sizes. We applied structural equation modelling, bootstrap procedures and fuzzy set qualitative comparative analysis to examine the variable relationships and configurational paths that explain AI capability formation. The results show that digital orientation is positively associated with AI capability through organizational learning and organizational forgetting. Furthermore, we provide evidence that AI trust strengthens the association between digital orientation and these two knowledge processes, which is consistent with a stronger conversion of digital strategic attention into AI capability. The configurational analysis further reveals two routes to high AI capability, namely a trust based forgetting route and a mature firm learning route. These findings clarify the knowledge mechanism through which digital orientation becomes AI capability and highlight the joint importance of learning, forgetting and trust.