Purpose This study offers an integrated perspective on how organizations restructure decision-making models and adopt strategies to balance algorithmic risk, innovation, and transparency in the Artificial Intelligence (AI) era. Design/methodology/approach A qualitative research design was employed based on semi-structured interviews with managers, analysts, HR professionals, and public and private sector experts. Data were analysed using the Gioia Methodology to build a theoretical framework grounded in practice. Findings The analysis identified three key organizational dimensions: algorithmic governance (macro level), algorithmic auditing (meso level), and AI literacy (micro level). These dimensions are functionally interconnected and guided by the transversal principle of responsible innovation. Together, they form a multilevel and holistic framework for AI management in organizations. This framework also reflects how organizations dynamically sense, seize, and transform themselves in response to AI-related challenges, enriching the literature on dynamic capabilities in the digital age. Research limitations/implications The framework provides actionable guidance for self-assessment, policy development, and strategic investment in AI governance and training. It also highlights the relevance of new professional roles and cross-functional teams in supporting ethical and sustainable AI adoption. Since based on in-depth qualitative interviews within a European context, the findings are not intended to be statistically generalizable. Future studies could empirically test the framework in diverse sectors and national contexts, explore how organizations of different sizes operationalize AI literacy, and examine how explainability tools can be embedded into governance structures. Originality/value This study contributes a novel conceptualization that integrates ethical concerns, control mechanisms, and organizational culture into a unified model. Originality lies in the way individual dimensions interact, are organizationally embedded, and aligned with emerging regulatory and strategic demands. Moreover, the study expands the debate on digital transformation and hybrid decision-making structures by proposing an organizationally embedded approach to AI risk management and innovation.
Romeo et al. (Thu,) studied this question.