Enhancing Agile Workflows with AI-Driven, Sustainability-Aware Requirements Engineering: A Design Science Approach
: Weerakoon, Oshani; Oyedeji, Shola; Samad, Md Abdus; Abubakar, Abdulkadir; Ishan, Ahsan; Shayan, Muhammad; Mäkilä, Tuomas; Kaila, Erkki
: Herzwurm, Georg; Petrik, Dimitri; Strobel, Gero; Kude, Thomas; Block, Lukas
: International Conference on Software Business
: 2026
Lecture Notes in Business Information Processing
: Software Business : 16th International Conference, ICSOB 2025, Stuttgart, Germany, November 24–26, 2025, Proceedings
: 574
: 212
: 229
: 978-3-032-14517-8
: 978-3-032-14518-5
: 1865-1348
: 1865-1356
DOI: https://doi.org/10.1007/978-3-032-14518-5_17
: https://doi.org/10.1007/978-3-032-14518-5_17
: https://research.utu.fi/converis/portal/detail/Publication/508254637
Generative AI presents growing opportunities across various fields, including Requirements Engineering (RE). RE, which is the backbone of software projects, drives the entire product development toward respective business goals. However, sustainability is not considered a primary need during the requirement elicitation, and as a result, software engineers are usually unable to envision the sustainability impacts of the products they build. To explore this gap, we introduce Reqwire, an AIdriven, sustainability-aware multi-agent system that (i) generates user stories from software requirement documents, (ii) enriches them with sustainability attributes, and (iii) integrates with common agile project management tools, like Jira. The system consists of specialized agents, namely as root, distributor, user story generator, and Jira. We followed a Design Science Research (DSR) approach under seven iterative cycles, incorporating feedback from an industry partner and academia to design and evaluate the Reqwire workflow. Our results indicate that Reqwire reduces manual effort by generating structured user stories, estimating story points, and assigning sustainability tags across five sustainability dimensions: environmental, economic, social, individual, and technical. The multi-agent-based framework enables integration with third-party tools, supporting consistent, systematic project tracking. Reqwire shows promise for enhancing agile workflows and promoting sustainable software practices from initial test rounds with the client.Generative AI presents growing opportunities across various fields, including Requirements Engineering (RE). RE, which is the backbone of software projects, drives the entire product development toward respective business goals. However, sustainability is not considered a primary need during the requirement elicitation, and as a result, software engineers are usually unable to envision the sustainability impacts of the products they build. To explore this gap, we introduce Reqwire, an AIdriven, sustainability-aware multi-agent system that (i) generates user stories from software requirement documents, (ii) enriches them with sustainability attributes, and (iii) integrates with common agile project management tools, like Jira. The system consists of specialized agents, namely as root, distributor, user story generator, and Jira. We followed a Design Science Research (DSR) approach under seven iterative cycles, incorporating feedback from an industry partner and academia to design and evaluate the Reqwire workflow. Our results indicate that Reqwire reduces manual effort by generating structured user stories, estimating story points, and assigning sustainability tags across five sustainability dimensions: environmental, economic, social, individual, and technical. The multi-agent-based framework enables integration with third-party tools, supporting consistent, systematic project tracking. Reqwire shows promise for enhancing agile workflows and promoting sustainable software practices from initial test rounds with the client.
:
This work has been supported by FAST, the Finnish Software Engineering Doctoral Research Network, funded by the Ministry of Education and Culture, Finland. And also, the SE4GD - Software Engineers for Green Deal program under Grant 619839 from the Erasmus Mundus Instrument of the European Union.