Beyond AI Adoption: A Human-Centred Decision Intelligence Framework for Sustainable Digital Transformation of French SMEs

An AI-Driven Strategic Technology Management Approach for Improving Investment Decisions, Productivity, Cyber Resilience and Business Competitiveness. Small and medium-sized enterprises (SMEs) are under simultaneous pressure to adopt artificial intelligence (AI), to protect themselves against rising cyber risk, and to reduce the cost and environmental footprint of their operations. Yet the decisions that govern these investments are still commonly made with vendor claims, intuition and spreadsheets, because the analytical machinery of large corporations is neither affordable nor usable at SME scale. This paper argues that the binding problem is not AI adoption but AI investment decision quality, and it proposes a remedy: a human-centred decision intelligence framework, TDIE-SME, that extends the Technology Decision Intelligence Engine (TDIE) to the specific conditions of French SMEs. The framework assesses any candidate technology across five dimensions - AI investment intelligence, SME digital transformation, cyber and technology risk intelligence, human-AI management, and sustainable technology strategy - and integrates multi-criteria decision analysis, predictive analytics and explainable AI behind a human-in-the-loop review gate. For each candidate technology it produces a five-part decision: whether to adopt, defer or reject; the expected financial return; the financial, operational and cyber risk profile; the readiness of people and organisation; and the expected effect on sustainability and competitiveness. The paper specifies the framework's architecture, indicators and decision pipeline, illustrates its use on a clearly hypothetical scenario, and sets out a six-stage research programme - including a minimum viable prototype of a decision intelligence dashboard, presently under development - through which the framework is to be empirically evaluated. No empirical results are claimed at this stage; the contribution is the framework, its decision logic, and a falsifiable evaluation design.