When AI Adoption Constraints Interact: Organizational Responses and Adoption Trajectories in Tunisian Telecommunications

Purpose – This study examines how technological, organizational and environmental constraints may interact during organizational artificial intelligence (AI) adoption and how firms respond to these interdependencies in the telecommunications sector. Design/methodology/approach – The study draws on 17 semi-structured interviews with managers and professionals directly involved in AI initiatives across three Tunisian telecommunications operators. An abductive thematic analysis combined open coding, the Technology–Organization–Environment (TOE) framework as a sensitizing device, relational coding, and within-case and cross-case comparison. Analytical matrices distinguished participant-stated linkages, process reconstruction and researcher inference. Findings – The analysis identifies four recurrent but non-systematic mechanisms. First, some constraints may propagate across TOE domains, as capability shortages affect data preparation, model reliability and business trust. Second, some environmental pressures become consequential through their translation into internal technological and organizational requirements. Third, some organizational responses may mitigate an initial constraint while generating or intensifying secondary dependencies, costs or coordination burdens. Fourth, the projects examined followed differentiated adoption trajectories, including scaling, partial deployment, prolonged implementation and non-scaling. Originality/value – The study develops a relational and conditional interpretation of TOE by shifting attention from isolated adoption factors to cross-domain propagation, the potential secondary effects of organizational responses and evolving adoption trajectories. Practical implications – Managers should diagnose chains of constraints rather than isolated barriers and assess the possible secondary consequences of mitigation actions before scaling AI initiatives. Stage-specific assessments of data readiness, integration, governance, maintainability and business viability remain important in regulated, infrastructure-intensive organizations operating under resource constraints