JOURNAL: AI diffusion as open innovation dynamics: Forecasting adoption, de-adoption, and stationary equilibrium in slovak enterprises

This paper models enterprise adoption of artificial intelligence (AI) as open innovation dynamics, forecasting AI diffusion in Slovak enterprises as a staged process of converting external technological opportunity into organizational capability. Using a cross-sectional survey, the study classifies enterprises as non-adopters, planners, and adopters, and develops a discrete-time Markov chain framework with annual transitions. Survey data were used to map firm-level transition probabilities through a logistic function and aggregated into scenario-specific transition matrices under optimistic, baseline, and conservative assumptions. Diffusion modelling uses an absorbing adoption variant and a de-adoption variant. The results show that AI diffusion is driven by two mechanisms—activation from non-adoption to planning and conversion from planning to adoption—and that scenario differences materially affect diffusion speed and the persistence of transitional segments. When adoption is absorbing, long-run outcomes approach near-saturation; once de-adoption is introduced, diffusion converges instead to a lower stationary equilibrium with a persistent non-adopter pool, showing that AI adoption remains bounded rather than self-sustaining. For research, the framework provides a transparent basis for future longitudinal validation; practically, it supports scenario-based forecasting, bottleneck diagnosis, and the timing of managerial interventions across activation, conversion, and retention levers. Its originality lies in translating open innovation dynamics into a calibrated Markov framework that captures both adoption progression and reversibility.

Authors:

Vladimír Hojdik, Peter Štetka, Nora Grisáková, Zuzana Hajduová, Federico Banda, Diana Pallérová

Read more: https://doi.org/10.1016/j.joitmc.2026.100839