Economies are complex systems in which many entities with various properties and roles interact continuously. The methodological approach started with Leon Walras and Vilfredo Pareto and continued by J.F. Muth, K. J. Arrow, G. Debreu, R. Lucas, and P. Samuelson has been, so far, the fundamental paradigm under which economists work. The traditional approach uses analytically tractable models with a representative perfectly rational agent. In this framework, mathematics has been the main ally of economics. At the same time, economics and finance are experiencing, in recent decades, a new approach that some commentators claim is an attempt at a paradigm shift: from a representative rational agent approach towards a behavioral agent-based approach in which heterogeneous agents are seen as boundedly rational. In this regard, the present paper aims to discuss whether and how Agent-Based Modeling (ABM) is participating in this potential paradigm shift. ABM has gained momentum in recent years thanks to technological advances and the availability of big data. ABM can help address the complexity of economic problems that cannot be solved analytically, allowing researchers to simulate social and relational processes. A critical but non-exhaustive discussion is also devoted to exploring to what extent economics could be on the way to experiencing a paradigmatic change.