Sub-agentic AI architecture: a conceptual review of specialization and hierarchical orchestration

Authors

DOI:

https://doi.org/10.26439/interfases2026.n023.8714

Keywords:

agents, architecture, artificial inteligence , ILIA, subagents

Abstract

During the second half of 2025, artificial intelligence (AI) research has shown a shift towards multi-agent architectures, in which sub-agents acquire a central structural role. Unlike monolithic approaches, these architectures allow complex tasks to be broken down into specialized functional units, coordinated by an orchestrating agent, which promotes scalability, parallelism, and more efficient context management. This article defines multi-agent architectures, web agents, and sub-agents. The conceptual framework Multi-Agent Data Mining (MADM), the architecture of the complex task tree, and the kitchen metaphor are analyzed. Benefits in terms of efficiency, resilience, and adaptability are identified, both in scientific environments and on commercial platforms. However, challenges remain related to orchestration, security, computational costs, academic training, and ethical governance. In this context, the Latin American Artificial Intelligence Index (ILIA 2025) is examined as a guide for integrating sub-agentic architectures into educational models with an ethical, interdisciplinary, and sustainable focus, aimed at fostering more responsible and inclusive regional innovation.

Downloads

Download data is not yet available.

Author Biographies

  • Olda Bustillos Ortega, Escuela de ingeniería informática, Universidad Internacional de las Américas, Costa Rica

    Es magíster en Informática con especialidad en Auditoría Informática por la Universidad Latina de Costa Rica. Actualmente, se desempeña como directora de la Escuela de Ingeniería Informática y docente investigadora en la Universidad Internacional de las Américas. Ha participado en la XI Conferencia “Servicios creativos y modernos para el comercio y desarrollo sostenible” de la Red Latinoamericana y Caribeña para la Investigación en Servicios (REDLAS), en The 2022 International Conference on Information Technology & Systems (ICITS’22) y fue distinguida como honorary academician of social sciences por la International Academy of Social Sciences de Palm Beach, Florida, Estados Unidos. Sus intereses de investigación se centran en la inteligencia artificial, la ciberseguridad y el desarrollo de sistemas inteligentes.

  • Jorge Murillo Gamboa, Escuela de Ingeniería Informática, Universidad Internacional de las Américas, Costa Rica

    Es magíster en Computación con especialidad en Telemática por el Instituto Tecnológico de Costa Rica. Actualmente, se desempeña como docente investigador. Ha recibido los reconocimientos Valued Member and Supporter (2016) y Certificate of Appreciation (2014), otorgados por la IEEE Computer Society. Sus intereses de investigación se centran en la inteligencia artificial, la ciberseguridad y el desarrollo de habilidades y competencias digitales según el modelo SFIA.

  • Olman Núñez Peralta, Escuela de Ingeniería Informática, Universidad Internacional de las Américas, Costa Rica

    Es magíster en Base de Datos por la Universidad Cenfotec. Actualmente, se desempeña como docente investigador. Sus intereses de investigación se centran en bases de datos, machine learning e inteligencia artificial.

  • Fabián Rodríguez Sibaja, Escuela de Ingeniería Informática, Universidad Internacional de las Américas, Costa Rica

    Es magíster en Administración de Proyectos por la Universidad para la Cooperación Internacional. Actualmente, se desempeña como docente investigador. Sus intereses de investigación se centran en las metodologías de desarrollo de software, la inteligencia artificial y los sistemas de información de apoyo a los negocios.

  • Daniel Mena Bocker, Escuela de Ingeniería Informática, Universidad Internacional de las Américas, Costa Rica

    Es magíster por la Universidad Magister. Actualmente, se desempeña como docente investigador. Sus intereses de investigación se centran en la ciberseguridad y la inteligencia artificial.

References

Auto-GPT. (s. f.). Auto-GPT – The next evolution of data driven Chat AI. https://auto-gpt.ai/

Babaei, G., & Giudici, P. (2025). Explainable artificial intelligence (XAI) in investment decision-making. Academia AI and Applications, 1(2). https://doi.org/10.20935/AcadAI8017

Bustillos Ortega, O., Murillo Gamboa, J., Núñez Peralta, O., & Rodríguez Sibaja, F. (2025). Creando asistentes, agentes y plataformas colaborativas de inteligencia artificial. Interfases, (21), 35-57. https://doi.org/10.26439/interfases2025.n021.7801

Carrascosa, C., Terrasa, A., Fabregat, J., & Botti, V. (2005). Gestión de comportamientos en agentes de tiempo real. Inteligencia Artificial: Revista Iberoamericana de Inteligencia Artificial, 9(25), 39-48. https://dialnet.unirioja.es/servlet/articulo?codigo=1212985

Chaimontree, S., Atkinson, K., & Coenen, F. (2010). Multi-agent based clustering: Towards generic multi-agent data mining. En P. Perner (Ed.), Advances in data mining: Applications and theoretical aspects (pp. 115-127). Springer. https://doi.org/10.1007/978-3-642-14400-4_9

Chan, A., Ezell, C., Kaufmann, M., Wei, K., Hammond, L., Bradley, H., Bluemke, E., Rajkumar, N., Krueger, D., Kolt, N., Heim, L., & Anderljung, M. (2024). Visibility into AI agents. En FAccT ’24: The 2024 ACM conference on fairness, accountability, and transparency (pp. 958-973). Association for Computing Machinery. https://doi.org/10.1145/3630106.3658948

Chevrot, A., Vernotte, A., Falleri, J.-R., Blanc, X., Legeard, B., & Cretin, A. (2025). Are autonomous web agents good testers? Proceedings of the ACM on Software Engineering, 2(ISSTA), 206-228. https://doi.org/10.1145/3728879

Claude Code Docs. (2025). Crear subagentes personalizados. https://code.claude.com/docs/es/sub-agents

Deng, Z., Guo, Y., Han, C., Ma, W., Xiong, J., Wen, S., & Xiang, Y. (2025). AI agents under threat: A survey of key security challenges and future pathways. ACM Computing Surveys, 57(7), Artículo 182. https://doi.org/10.1145/3716628

Durán, R., Moreno, A., Adasme, S., Rovira, S., Jordán, V., & Poveda, L. (Coords.). (2025). Índice Latinoamericano de inteligencia artificial (ILIA) ٢٠٢٥. Comisión Económica para América Latina y el Caribe. https://www.cepal.org/es/publicaciones/82514-indice-latinoamericano-inteligencia-artificial-ilia-2025

Elmahalawy, A. M. (2015). A new approach in learning for intelligent multi agent systems. En Proceedings 21st European Conference on Modelling and Simulation. European Council for Modelling and Simulation. https://www.academia.edu/70025062/Intelligent_Agent_Technology_Intelligent_Multi_Agent

Github. (s. f.). Github Copilot. Domina tu oficio. https://github.com/features/copilot?locale=es-419

Gonçalves, E. J. T., Cortés, M. I., Campos, G. A. L., Lopes, Y. S., Freire, E. S. S., Da Silva, V. T., De Oliveira, K. S. F., & De Oliveira, M. A. (2015). MAS-ML 2.0: Supporting the modelling of multi-agent systems with different agent architectures. Journal of Systems and Software, 108, 77-109. https://doi.org/10.1016/j.jss.2015.06.008

Ning, L., Liang, Z., Jiang, Z., Qu, H., Ding, Y., Fan, W., Wei, X. Y., Lin, S., Liu, H., Yu, P. S., & Li, Q. (2025). A survey of web agents: Towards next-generation AI agents for web automation with large foundation models. En Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 6140-6150). Association for Computing Machinery. https://doi.org/10.1145/3711896.3736555

Rother, D., Pajarinen, J., Peters, J., & Weisswange, T. H. (2025). Open-ended coordination for multi-agent systems using modular open policies. Autonomous Agents and Multi-Agent Systems, 39, Artículo 40. https://doi.org/10.1007/s10458-025-09723-7

Samdani, G., Viswanathan, G., & Dasu Jegadeesh, A. (2025). Human-AI collaboration: Balancing agentic ai and autonomy in hybrid systems. International Journal on Cloud Computing: Services and Architecture, 15(1). https://doi.org/10.5121/ijccsa.2025.15101

Soto Arriaza, A., & Durán Rojas, R. (Dirs.). (2023). Índice latinoamericano de inteligencia artificial. Centro Nacional de Inteligencia Artificial. https://indicelatam.cl/wp-content/uploads/2023/08/ILIA-2023.pdf

White, R. W. (2024). Advancing the search frontier with AI agents. Communications of the ACM, 67(9), 54-65. https://doi.org/10.1145/3655615

Zhu, Z., Tan, Y., Yamashita, N., Lee, Y. C., & Zhang, R. (2025, 26 de abril). The benefits of prosociality towards AI agents: Examining the effects of helping AI agents on human well-being. En Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (pp. 1-18). Association for Computing Machinery. https://doi.org/10.1145/3706598.3713116

Downloads

Published

2026-07-30

Issue

No. 023 (2026)

Section

Research papers

License

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

Authors who publish with this journal agree to the following terms:

Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under an Attribution 4.0 International (CC BY 4.0) License. that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.

Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.

Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).

 

Last updated 03/05/21

How to Cite

Bustillos Ortega, O., Murillo Gamboa, J., Núñez Peralta, O. ., Rodríguez Sibaja, F., & Mena Bocker, D. (2026). Sub-agentic AI architecture: a conceptual review of specialization and hierarchical orchestration. Interfases, 023, 62-85. https://doi.org/10.26439/interfases2026.n023.8714

Most read articles by the same author(s)