Sufficient and necessary conditions in the adoption of generative artificial intelligence by entrepreneurs: findings in Spain

Purpose This study assesses the use of generative artificial intelligence (GenAI) among entrepreneurs in Spain, examining both sufficient and necessary conditions for different levels of use. Design/methodology/approach The study uses survey data from the Centro de Investigaciones Sociológicas. Ordered logistic regression is applied to identify sufficient conditions, while Necessary Condition Analysis is used to detect bottlenecks for higher levels of GenAI use. Findings Knowledge of GenAI and perceived economic usefulness are the main predictors of more intensive use. However, the NCA shows that technological optimism and the absence of very high scepticism operate as necessary conditions, especially at regular levels of use. Research limitations/implications The assessment relies on cross-sectional, self-reported data from Spanish self-employed workers. Future research should distinguish between different entrepreneurial profiles and types of GenAI use. The analytical framework may be applicable across different contexts, although the specific associations and bottleneck thresholds observed here should be validated through cross-country research. Practical implications Support policies should combine awareness-raising, applied training, and practical guidance on how GenAI can generate value in entrepreneurial activity. Social implications The findings suggest that responsible GenAI adoption among entrepreneurs requires not only favourable attitudes towards technology but also knowledge, perceived usefulness, and clear guidance on its business applications. Originality/value The study contributes by combining sufficiency and necessity logics to explain actual GenAI use among entrepreneurs, rather than focussing only on behavioural intention, while indicating that technology-readiness dispositions and GenAI-specific behavioural factors may fulfil distinct roles at different levels of use. By distinguishing probabilistic drivers from necessary bottlenecks, the study provides a more complete account of entrepreneurial GenAI adoption and generates more actionable implications for training and policies aimed at increasing effective use.