Cognitive Load, Behavioral Finance and Artificial Intelligence in Supporting Dividend Stock Selection: Architectural Proposition of AI Smart Portfolio

Proposição Arquitetural da Carteira Inteligente IA

Abstract

The growth in the number of individual investors in the Brazilian capital market has expanded access to equities, but it has not removed information asymmetry or the technical limitations involved in interpreting accounting and financial data. In this context, dashboards available on traditional market platforms concentrate multiple indicators, historical series, filters, and screens, imposing a high mental processing effort on users. Consequently, investment decisions may be affected by cognitive overload, decision fatigue, and biases described in Behavioral Finance. In light of this, this article proposes the conceptual architecture of AI Smart Portfolio, a language model designed to support dividend stock selection through natural language interaction and integration with external APIs. Thus, the artifact was outlined based on Design Science Research, combining Cognitive Load Theory, Behavioral Finance, and conversational systems. In addition, the model was conceived to consume Yahoo Finance data via RapidAPI, process indicators such as payout and dividend yield, and return structured, comparable, and traceable reports. Therefore, it is argued that the solution operates as cognitive mediation by reducing visual noise, condensing dispersed data, and supporting the interpretation of financial metrics without transferring the final decision to the system.