AI Diffusion, Implementation Capacity, and Firm Performance
IMF Working Papers · 2026-11 · 2026 (253) · 1
DOI: 10.5089/9798229066600.001 ↗Kurz gefasst
Fachartikel in der Zeitschrift IMF Working Papers, 2026-11. Verfasst von Nordine Abidi und 2 weiteren Autor:innen.
Abstract (Originalsprache)
This paper studies artificial intelligence (AI) diffusion, firm-level implementation capacity, and performance using the World Bank Enterprise Surveys, covering more than 100,000 firms across 158 economies during 2022–2025. We combine a text-based indicator of explicit AI mentions in innovation narratives with an implementation-capacity index capturing digital readiness, workforce skills, and innovation capabilities. Explicit AI mentions are exceptionally rare, averaging about 0.6 percent across firms, while implementation capacity varies substantially across regions and the firm distribution. Average firm implementation capacity is positively but imperfectly associated with national AI preparedness, suggesting that national benchmarks provide an incomplete picture of firm capabilities. Within countries, higher implementation capacity is associated with higher sales and employment growth and lower probabilities of substantial contractions. The findings document substantial differences in complementary capabilities relevant to AI and broader technological absorption, without identifying the performance effects of AI use itself.
Zitieren
APA
Abidi, N., Herradi, M. E., & Sakha, S. (2026). AI Diffusion, Implementation Capacity, and Firm Performance. IMF Working Papers, 2026(253), 1. https://doi.org/10.5089/9798229066600.001
Vancouver
Abidi N, Herradi ME, Sakha S. AI Diffusion, Implementation Capacity, and Firm Performance. IMF Working Papers. 2026;2026(253):1. doi:10.5089/9798229066600.001
BibTeX
@article{Abidi2026,
author = {Abidi, Nordine and Herradi, Mehdi El and Sakha, Sahra},
title = {AI Diffusion, Implementation Capacity, and Firm Performance},
journal = {IMF Working Papers},
year = {2026},
volume = {2026},
number = {253},
pages = {1},
doi = {10.5089/9798229066600.001}
}
EUVIMED