The AI Implementation Paradox in Danish SMEs: Why Positive Intentions Fail to Translate into Sustained Use

A substantial gap persists between organizational intention to adopt artificial intelligence (AI) marketing tools and their sustained use. This study investigates why Danish small and medium-sized enterprises (SMEs) with clear implementation intentions nonetheless fail to implement or scale AI adoption. Drawing on a qualitative multiple case study of six Danish SMEs and integrating the Technology Acceptance Model (TAM), Resource-Based View (RBV), and organizational AI capability research, we identify five mechanisms that systematically undermine intention-to-use translation: expectation-reality mismatches, time paradoxes, organizational structure constraints, tool proliferation, and support vacuums. We introduce the “intention-expansion gap,” in which successful initial adoption stabilizes at partial use despite persistent expansion intentions. We extend TAM by positioning implementation capability as a moderator of the intention-behavior relationship and reconceptualize perceived ease of use as dynamically evolving; we extend RBV by showing resource constraints operate as implementation rather than adoption barriers.