Article / Volume 27, Issue 4 / October 2026
What Counts as AI Adoption in SMEs? Construct Validity and Organizational Mechanisms
- Published
- October 2026
- Volume & issue
- 27 (4)
- Pages
- 753–768
- DOI
- 10.24818/RMCI.2026.4.753
Abstract
Reviews of artificial intelligence (AI) adoption in small and medium-sized enterprises (SMEs) may combine studies that do not verify the technology, the firm or the use. This study audits a pre-selected corpus of 30 candidate case reports from a February 2025 Web of Science and Scopus search. Three sequential gates test for qualifying AI, verifiable SME status and use in real work. Twelve reports pass the baseline definition, and 18 fail at their first unmet gate. Strict and permissive AI definitions retain 11 and 16 reports, respectively. The audit therefore yields a reusable decision tree and a screening checklist. The retained reports are then read against organizational theory. Four theory-derived propositions are compatible with these reports to different degrees. A conceptual proposition on employee-initiated generative AI and a rival explanation are offered for future testing. Construct validity therefore changes what counts as evidence of AI adoption in SMEs.
Keywords
JEL classification
M15, O33, L26
How to cite
Adrian George BOZINTAN, Emil-Lucian CRIȘAN, Maria MORTAN (2026). What Counts as AI Adoption in SMEs? Construct Validity and Organizational Mechanisms. Review of International Comparative Management, 27(4), 753–768. DOI: 10.24818/RMCI.2026.4.753