The Agentic Paradigm in Data Analytics: Architectures, Benchmarks, and Socio-Economic Trajectories

Abstract

The transformation of analytics from traditional human-led methods to automated and AI-drivenprocesses is occurring at a rapid pace. This shift represents the most significant change since the movefrom relational databases to big data architectures. Critical bottlenecks in automated analytics includeSQL generation errors caused by Large Language Model hallucinations and the inability to navigate inconsistent enterprise data.