MONOGRAPHY: Artificial Intelligence in Logistics: Evidence from Slovak Enterprises

Artificial intelligence (AI) is widely expected to reshape logistics and supply-chain management, yet enterprise-level evidence on how firms actually deploy it remains thin, fragmented, and drawn mostly from large Western or Asian corporations. Evidence from the small, open, transition economies of Central and Eastern Europe (CEE) is almost absent, and logistics is usually folded into generic accounts of “AI in the supply chain.” This monograph addresses that gap using the AI-IMPACT survey, an organization-level study of AI adoption among enterprises operating in Slovakia (N = 816 deduplicated firms). It focuses on the 142 firms that have begun implementing or testing AI in logistics, the third most frequently named functional domain of AI application among surveyed AI users, and contrasts them with the 420 surveyed AI users that have not. 

Combining descriptive, exploratory, and confirmatory methods (logistic and negative-binomial regression, exploratory factor analysis, a k-means partition, and Wilcoxon signed-rank testing), the study develops one evidence-backed argument: among these firms, adoption has run ahead of embedding. They pilot a broad portfolio of logistics AI applications, but few have converted pilots into regular operational use. The average adopter has 5.3 of 13 applications at least in pilot use and only 1.8 in regular use; 59% have none in regular use at all. A two-group partition separates a minority of “embedded” adopters (35%) from a majority of “experimenters” (65%). What most clearly distinguishes the two is organizational AI maturity, rather than firm size, sector, or perceived barriers. In a multivariate model maturity is the strongest predictor of the breadth of active deployment (incidence-rate ratio ≈ 2.5 per standard deviation), and the association holds when the two maturity dimensions that overlap with usage are removed (IRR ≈ 2.2 for the governance-and-strategy dimensions alone). Perceived benefits exceed perceived risks (Wilcoxon p < .001), though realized gains so far concentrate in growth, service, and productivity rather than hard cost savings, and data security is the foremost concern. Barriers, measured across the whole sample, are higher among non-adopters and predict who adopts; among adopters they do not predict how deeply AI is embedded. All relationships are cross-sectional associations, not causal effects. The monograph reads organizational AI maturity as the firm characteristic most closely tied to the move from experimentation to embedded use, and draws out the implications for managers, technology providers, and CEE digitalization policy. 

Authors:

Peter Štetka & Diana Pallérová