Authors:
Rafael Marconi Ramos
1
;
2
;
Pedro Brom
2
;
3
;
João Gabriel de Moraes Souza
4
;
5
;
Li Weigang
2
;
Vinícius Di Oliveira
2
;
6
;
Silvia Reis
5
;
7
;
Jose Salm Junior
5
;
8
;
Vérica Freitas
9
;
Herbert Kimura
7
;
5
;
Daniel Cajueiro
5
;
10
;
Gladston Luiz da Silva
5
and
Victor Celestino
5
;
7
Affiliations:
1
Euro-Americano University Center (Unieuro), Brasília, Brazil
;
2
TransLab, Department of Computer Science, University of Brasília, Campus Darcy Ribeiro, Brasília, Brazil
;
3
Department of Mathematics, Federal Institute of Education, Science and Technology of Brasília, Campus Estrutural, Brasília, Brazil
;
4
Department of Production Engineering, University of Brasília, Campus Darcy Ribeiro, Brasília, Brazil
;
5
LAMFO - Lab. of ML in Finance and Organizations, University of Brasília, Campus Darcy Ribeiro, Brasília, Brazil
;
6
Federal District Secretariat of Economy, Brasília, Brazil
;
7
Department of Business Administration, University of Brasília, Campus Darcy Ribeiro, Brasília, Brazil
;
8
University of the State of Santa Catarina, Florianópolis, Santa Catarina, Brazil
;
9
School of Business and Management, Uberlandia Federal University, Uberlândia, Brazil
;
10
Department of Economics, University of Brasília, Campus Darcy Ribeiro, Brasília, Brazil
Keyword(s):
Multi-Agent Systems, LLM, MoE, Generative AI in Government, Text Rewriting and Simplification, Gov.br.
Abstract:
This paper presents an intelligent multi-agent system to improve clarity, accessibility, and legal compliance of public service descriptions on the Brazilian Gov.br platform. Leveraging large language models (LLMs) like GPT-4, agents with specialized contextual profiles simulate collective deliberation to evaluate, rewrite, and select optimal service texts based on ten linguistic and seven legal criteria. An interactive voting protocol enables consensus-based editorial refinement. Experimental results show the system produces high-quality texts that balance technical accuracy with linguistic simplicity. Implemented as a Mixture of Experts (MoE) architecture through prompt-conditioning and rhetorical configurations within a shared LLM, the approach ensures scalable legal and linguistic compliance. This is among the first MoE applications for institutional text standardization on Gov.br, establishing a state-of-the-art precedent for AI-driven public sector communication.