A Personalised Cloud-Intelligence (PCI) framework is proposed as an AI-driven system for decision support in small and medium-sized enterprises (SMEs). PCI consists of two primary components: a Digital Twin, which acts as a dynamic virtual representation of the enterprise, and a Cloud Intelligence layer, which integrates machine learning, federated learning, and retrieval-augmented generation to provide actionable insights. The Digital Twin component standardizes and organizes business data such as sales, inventory, finance, and personnel, collected through voice input or document importation. It models real-time operations, supports data imputation, and enables predictive analytics for tasks like demand forecasting and resource planning. The Cloud Intelligence layer enriches the system with external datasets, including market trends and regulatory updates, and employs collaborative model training through federated learning to improve decision-making without compromising data locality. The framework is conceptualized with a modular architecture that includes tools for Extract, Transform, Load (ETL) processes, multi-objective optimization, and explainable AI. The article presents three use cases that illustrate potential applications of PCI in industries such as retail, craft businesses, and independent artistry. It also discusses critical technical issues, including data integration, privacy-preserving methodologies, and user-centric interaction models like conversational interfaces and dashboards. While PCI remains a conceptual framework, it provides a foundation for developing accessible, domain-specific AI technologies that enable SMEs to apply analytics for strategic decision-making.