Prompt-Based Automation of Agile Requirements Documentation: Design for Limited-Infrastructure Environment

Authors

  • Roman Kruk Institute of automation, cybernetics and computer engineering, National University of water and environmental engineering
  • Nataliia Zhukovska The Department of Computer Science and Applied Mathematics, The National University of Water and Environmental Engineering, NUWEE

DOI:

https://doi.org/10.31713/MCIT.2025.032

Keywords:

Agile, requirements engineering, prompt engineering, Large Language Models, Feature Specification, process automation

Abstract

This study presents a proof-of-concept design and evaluation of a prompt-based automation approach for Agile requirements documentation in a limited-infrastructure environment. Building upon a previously proposed five-step LLM-based analytical framework, the research explores how structured prompt engineering can replicate the behavior of a full middleware system for transforming Jira exports into publishable Feature Specification Documents. The experiment involved 12 iterative prompt-refinement cycles applied to real project data. The final Gem’s configuration achieved 79% successful cases (improved by 27% when project knowledge artifacts were included). The results confirm that prompt-driven automation can effectively bridge unstructured Agile artifacts and formal specification, offering a lightweight and auditable alternative to traditional requirements documentation workflows.

Downloads

Published

2025-11-06

How to Cite

Kruk, R., & Zhukovska, N. (2025). Prompt-Based Automation of Agile Requirements Documentation: Design for Limited-Infrastructure Environment. Modeling, Control and Information Technologies, (8), 113–116. https://doi.org/10.31713/MCIT.2025.032

Issue

No. 8 (2025): Modeling, control and information technologies

Section

Software engineering and information technologies

License

Copyright (c) 2025 Modeling, Control and Information Technologies: Proceedings of International scientific and practical conference

Creative Commons License

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

All materials are distributed under the terms of the Creative Commons Attribution 4.0 International License, which allows others to distribute the work with attribution to the authorship of this work and the first publication in this journal.