MONOGRAPHY: Artificial Intelligence in Marketing: Evidence from Slovak Enterprises
Artificial intelligence (AI) is entering the marketing function quickly, yet firm-level evidence on how companies adopt AI in marketing, as opposed to how consumers react to a single AI tool, remains thin. The gap is widest for small and medium-sized enterprises (SMEs) and for Central and Eastern Europe (CEE). We address it with the AI-IMPACT Survey Dataset (Slovakia, 2025), an organization-level survey of 816 enterprises, of which 562 already use or pilot AI and 209 have begun implementing or testing AI in Marketing and Sales. Working from the Technology– Organization–Environment (TOE) framework and Rogers’s Diffusion of Innovations, the study addresses seven research questions and twelve hypotheses covering the adoption landscape, the use-case architecture of fifteen marketing applications, the correlates of adoption and of use-case breadth, the way firms weigh benefits against risks, the perceived value and employment effects, and a data-driven segmentation of adopters. The design is cross-sectional and the measures are self-reported, so the results describe associations and perceptions in this sample rather than causal effects or population parameters.
Marketing and Sales is the most frequently selected functional area for AI among the AI-using firms surveyed (37.2%), ahead of production, logistics, and finance, though the option bundles marketing with sales and is broader than the single functions it outranks; a paired test confirms it is selected more often than production (McNemar p = .003). Adoption is recent (68.4% of marketing adopters dated their first AI project to 2023 or later), which mirrors the recency of enterprise AI generally rather than anything specific to marketing. Across the fifteen use cases, customer-facing automation is furthest along, with chatbots, advertising automation, personalized offers, and price optimization in active use by roughly 40–50% of adopters, while data-intensive analytical applications trail and sit mostly in the planning stages. The breadth of active use is most strongly associated with perceived benefit (negative-binomial IRR = 1.51, 95% CI [1.20, 1.89], p < .001); firm size and formal AI maturity show no independent association once benefit is included, although, because benefit and breadth are reported by the same respondent, the association is equally consistent with breadth shaping belief. Benefits and risks do not trade off. Firms that rate benefits highly tend also to rate risks highly (Spearman ρ = .65, falling to about .47 once a response-style adjustment is applied), benefits exceed risks only modestly (3.62 vs. 3.46; Wilcoxon p < .001, d_z = 0.28), and organizational “fear” is essentially unrelated to adoption. More intensive use is associated with more favourable perceived outcomes for revenue, productivity, and customer-service quality (ρ ≈ .27–.42, all q < .001 after correction).
On employment, more adopters report adding or planning AI-related roles (22.5%) than report AI-related layoffs, realized or planned (8.6%), which indicates no sign of marketing-led displacement at this stage rather than a measured net employment effect. A two-segment description, here labelled Advanced Integrators (56%) and Cautious Experimenters (44%), captures most of the structure in the adopter population, though the two groups differ in degree of engagement rather than in kind.The study offers use-case-level evidence on AI in marketing for an under-examined CEE economy, complementing prior CEE work on general MarTech and digital-marketing adoption; it elaborates how TOE and diffusion ideas apply to the breadth and depth of use rather than to adoption alone; and it gives managers, policymakers, and technology providers a grounded read on where the bottlenecks lie. All numerical results, including the multiple-comparison corrections and common-method-bias diagnostics, are documented in an accompanying Master Results Registry, and all literature claims in a Literature Registry of 108 verified sources.
Authors:
Peter Štetka, Federico Banda
