Publicação

Opportunities and Challenges of Multi Agent Systems in the Financial Services Industry

Ver documento

Detalhes bibliográficos
Resumo:The rapid advancement of artificial intelligence, particularly large language models (LLMs), has accelerated the adoption of multi-agent systems (MAS) within the financial services industry. Despite growing interest, existing research remains fragmented, either addressing the managerial aspects of AI implementation or offering domain-specific overviews of MAS, without providing an integrated, practical perspective. Using a design science research approach, the study first creates an evaluation matrix, based on Analytical Hierarchy Process (AHP), to systematically assess AI use cases. It then develops and tests a MAS prototype for regulatory change analysis, leveraging retrieval-augmented generation (RAG) and hierarchical agent collaboration. Finally, the research identifies key opportunities and challenges arising from the implementation process. The research combines a comprehensive literature review, expert interviews, and quantitative assessments to identify high-value use cases and evaluate system performance against industry-relevant criteria. Findings highlight both the transformative potential and significant hurdles associated with MAS adoption, including issues of scalability, compliance, and organizational alignment. By synthesizing theoretical and practical insights, this thesis contributes a holistic framework for the implementation and evaluation of MAS in financial services, bridging the gap between academic research and real-world application.
Autores principais:Fernandez, Noah Nikolas Campaña
Assunto:Large Language Model Multi-Agent System Generative Artificial Intelligence Retrieval-Augmented Generation Financial Services Industry SDG 8 - Decent work and economic growth SDG 9 - Industry, innovation and infrastructure
Ano:2025
País:Portugal
Tipo de documento:dissertação de mestrado
Tipo de acesso:acesso aberto
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
Descrição
Resumo:The rapid advancement of artificial intelligence, particularly large language models (LLMs), has accelerated the adoption of multi-agent systems (MAS) within the financial services industry. Despite growing interest, existing research remains fragmented, either addressing the managerial aspects of AI implementation or offering domain-specific overviews of MAS, without providing an integrated, practical perspective. Using a design science research approach, the study first creates an evaluation matrix, based on Analytical Hierarchy Process (AHP), to systematically assess AI use cases. It then develops and tests a MAS prototype for regulatory change analysis, leveraging retrieval-augmented generation (RAG) and hierarchical agent collaboration. Finally, the research identifies key opportunities and challenges arising from the implementation process. The research combines a comprehensive literature review, expert interviews, and quantitative assessments to identify high-value use cases and evaluate system performance against industry-relevant criteria. Findings highlight both the transformative potential and significant hurdles associated with MAS adoption, including issues of scalability, compliance, and organizational alignment. By synthesizing theoretical and practical insights, this thesis contributes a holistic framework for the implementation and evaluation of MAS in financial services, bridging the gap between academic research and real-world application.