Artificial Intelligence in Serbian Enterprises: Adoption Levels, Obstacles and Sectoral Applications

Abstract

This paper examines the adoption of artificial intelligence (AI) among Serbian enterprises, with a focus on adoption levels, obstacles, and sectoral applications. Based on a survey of 499 companies conducted in 2025, the findings show that only 29.1% of firms use AI technologies, with adoption marginally higher in manufacturing (31.1%) than in services (29.2%), although the difference is not statistically significant. The main barrier to adoption is the perception that AI is not useful or relevant for business operations, reported by 66.9% of non-adopters, followed by concerns about legal uncertainty, lack of knowledge, and high costs. Among adopters, the most frequent applications are data analysis and interpretation, customer support, and personalization in marketing. Sectoral differences are evident: services focus on customer support, while manufacturing emphasizes personalization and product development. The results suggest that informational barriers and limited internal capacities outweigh purely technological constraints, highlighting the need for targeted policy measures to foster AI adoption.