Integrating AI-Based Forecasting with Corporate Performance Management: Architecture, Methods, and Results

The thesis, carried out at Var Group S.p.A., examines the integration of two in-house solutions: Foresight, an AI platform for demand forecasting, and CPM (Corporate Performance Management) used for planning and performance management. Historically, the tools operated as separate systems, causing fragmented outcomes, and weak alignment between operational forecasts and economic and financial objectives. The research goal is to design an integrated architecture that connects AI-generated forecasts to CPM planning models, reducing manual activities and updating latency while enabling unified, cross-functional decision governance. The methodology combines a theoretical review of demand planning and AI forecasting techniques, an analysis of current market solutions, and the development of a unified data model enabling forecast-to-plan flows. A real case study is simulated using an integrated prototype application to assess feasibility and benefits. Results indicate effective harmonization between the two models, a reduction in manual and import/export errors, more timely updates, and the creation of a single forecasting scenario shared between the sales planning, demand planning, and finance departments. The added value for Var Group is the evolution of its offering toward a truly integrated platform; for client companies, the benefits are faster, more accurate and more collaborative decisions. Future developments include process automation via APIs and orchestrators, near real-time data synchronization, and the activation of a feedback loop for the model retraining, supporting the industrial scalability of the solution.