Tecnología basada en IA para el diagnóstico de la madurez de la comunicación digital de las PYMES en Cuba

Abstract

The digital transformation of small and medium-sized enterprises (SMEs) requires tools that can diagnose, explain, and improve communication management without relying on massive datasets or expensive infrastructures. This paper develops an artificial-intelligence (AI)-based technology for diagnosing digital communication maturity in Cuban SMEs, integrating external and internal communication, data-quality control, error detection, and algorithmic performance evaluation. The technology retains a weighted seven-indicator maturity index and adds an internal-communication module, an alert system, and an AI Quality Index (AI-QI) that is kept separate from organizational maturity. Two demonstrative SME cases, ElectroStar and FoodAlimento, are used to test arithmetic consistency and diagnostic rules from an aggregate reference matrix; they are not presented as independent empirical validation. The six recalculated IM-GPC scores were reproducible (ElectroStar: 0.209, 0.318, 0.435; FoodAlimento: 0.199, 0.241, 0.259), and a Monte Carlo analysis with ±20% perturbations of the weights showed stable classifications. The diagnostic layer is extended with schema checks, missingness, duplicate detection, logical inconsistencies, outliers, temporal leakage, data drift, model degradation, and task-specific performance metrics for classification, regression, clustering, calibration, uncertainty, and generative-AI factuality. The result is a reproducible and auditable SME-oriented procedure ready for prospective multicenter validation in Cuba.