Brazil and Argentina publish AI studies on generative BI dashboards and agent protocol

Researchers at Brazil's Federal University of Mato Grosso do Sul conducted an exploratory study on applying generative AI to the creation and maintenance of Business Intelligence (BI) dashboards. By comparing academic surveys with industry reports from Uber and Pinterest, they identified convergence on two fronts: grounding AI models in real‑world context through retrieval‑augmented generation, metadata, or semantic layers, and the critical role of human validation when selecting tables, which remains the primary source of errors. The authors propose a three‑stage workflow—preparation, construction, and operation—to help small or inexperienced teams adopt generative AI for BI with controlled risk.

In a separate thesis from Argentina's Universidad Nacional de Río Negro, author Francisco Martin Rivero introduced the Agent‑to‑Agent (A2A) protocol, a standardized framework designed to enable seamless communication and collaboration among heterogeneous AI agents. The protocol allows agents from different organisations to share information, delegate tasks, and coordinate actions securely, addressing current interoperability silos. A practical implementation and evaluation were carried out, demonstrating the protocol's potential benefits for AI system integration.