AI-Driven Economic Analysis: From Data to Decisions

This paper proposes artificial intelligence (AI) architectures for economic analysis, illustrating how AI components integrate from data inputs to decision outputs. Specifically, it reveals layered AI architectures, consisting of a data layer, intelligence layer, and decision layer, varying from traditional AI to generative AI and agentic AI, to address key economic analysis. This paper illustrates how economic data can be used in combination with AI and machine-learning models and applications to transform output into economic insights. It explores the transition from descriptive to predictive insights, from predictive to economic narratives, and from economic narratives to independent decisions. Moreover, it provides an economic evolution view on AI, starting with traditional AI, moving to generative AI and, finally, to agentic AI, and an economic intelligence transition from static analysis to creative modelling and finally autonomous economic reasoning. Furthermore, it develops country-specific use cases for AI-driven economic analysis, such as forecasting inflation and demand for electricity and gas, analyzing the impact of policy measures, and optimizing fiscal allocation in an emerging economy like Pakistan.