Agile Management for Foundation Model Based Agentic Systems: A Vision

• Lucas Romao PUC-Rio https://orcid.org/0009-0001-2153-1498
• Yasmin Sandes PUC-Rio
• Gabriel Mariquito PUC-Rio
• Jose M. C. Boaro PUC-Rio https://orcid.org/0000-0002-4456-9050
• Marcos Antonio Alves UFVJM https://orcid.org/0000-0001-6934-6745
• Marcos Kalinowski PUC-Rio https://orcid.org/0000-0003-1445-3425

DOI: https://doi.org/10.5753/se4as.2026.32449

Resumo

[Context] Foundation Model (FM)-based agentic systems extend the reasoning capabilities of FMs to enable autonomous software components. Yet, while these systems open new possibilities for software, managing their development through agile methodologies is not straightforward: agent development is inherently experimental, traditional agile frameworks offer no suitable artifacts for it, and non-technical stakeholders often lack AI literacy. [Goal] To address these challenges, in this vision paper we discuss how agile management can be extended to support the development of FMbased agentic systems. [Method] We ground our vision in lessons learned from agile management of Machine Learning (ML)-enabled systems development, extending it to an agentic perspective. [Results] Our vision comprises guidance organized around practices, artifacts, and roles. In terms of practices, we suggest that teams refine agent development through sprint-bound experimentation, with an initial specification of agent requirements and an offset cadence between AI and Software Engineers. Regarding artifacts, we propose specifying agent-intensive agile requirements through Agent and Data Stories, with Layers of Done versioning, such as an Agent Demo API and a Minimum Viable Agent. On roles, we advocate that Product Owners specify Agent and Data Stories collaboratively with AI Engineers. [Conclusion] We acknowledge that this guidance still require empirical evaluation. Even so, we propose it as a starting point for practitioners and researchers seeking to extend agile management to the development of agentic systems.

Palavras-chave: Agile Management, Foundation Model Based Agents

Referências

Júlia Condé Araújo, Marina Condé Araújo, José MC Boaro, and Marcos Kalinowski. 2026. Towards an Approach for Specifying Intelligent Systems Involving FoundationModel Based Agents. In 5th International Conference on AI Engineering – CAIN’26 (Rio de Janeiro, Brazil).

Rishi Bommasani et al. 2021. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258 (2021).

Martin Fowler, Jim Highsmith, et al. 2001. The agile manifesto. Software development 9, 8 (2001), 28–35.

Christian Kästner. 2022. Machine Learning in Production: FromModels to Products.

Marco Kuhrmann, Paolo Tell, et al. 2021. What Makes Agile Software Development Agile? IEEE Transactions on Software Engineering 48 (2021).

Mirko Perkusich, Danyllo Albuquerque, Allysson Allex Araújo, et al. 2026. Adoption of Large Language Models in Scrum Management: Insights from Brazilian Practitioners. InAgile Processes in Software Engineering and Extreme Programming. Springer Nature Switzerland.

Lucas Romao, Hugo Villamizar, Romeu Oliveira, Silvio Alonso, and Marcos Kalinowski. 2026. Agile Management for Machine Learning: A Systematic Mapping Study. In Software Engineering and Advanced Applications, Davide Taibi and Darja Smite (Eds.). Springer Nature Switzerland.

Lucas Romao, Luiz Xavier, Júlia Condé Araújo, Marina Condé Araújo, Ariane Rodrigues, and Marcos Kalinowski. 2026. Applying a Requirements-Focused Agile Management Approach for Machine Learning-Enabled Systems. In 5th International Conference on AI Engineering – CAIN (Rio de Janeiro, Brazil).

Davide Taibi, Henry Muccini, Karthik Vaidhyanathan, Marcos Kalinowski, et al. 2026. A research agenda on agents and software engineering: Outcomes from the rio A2SE seminar. arXiv preprint arXiv:2605.11720 (2026).

Karthik Vaidhyanathan, Anish Chandran, HenryMuccini, and Regi Roy. 2022. Agile4MLS—Leveraging Agile Practices for Developing Machine Learning-Enabled Systems: An Industrial Experience. IEEE Software 39 (2022).

Lei Wang et al. 2024. A survey on large language model based autonomous agents. Frontiers of Computer Science 18, 6 (2024), 186345.

Zheying Zhang et al. 2025. AI and Agile Software Development: A Research Roadmap from the XP2025 Workshop. arXiv preprint arXiv:2508.20563 (2025).