ReqGenie: GPT-Powered Conversational-AI for Requirements Elicitation
Paper i proceeding, 2025

Requirements elicitation is a crucial activity in the software product lifecycle, ensuring a clear understanding of the user needs and project goals. This process guides the design, development, and validation phases, resulting in a successful and functional product. However, eliciting requirements often encounters the challenge of finding experts to gather, analyze, and validate the requirements. To address those challenges, this study aims to investigate the potential of Conversational AI, specifically utilizing OpenAI’s custom GPT, to elicit requirements through stakeholder interviews. Employing an iterative design science approach, ReqGenie was implemented to gather and summarize requirements from text-based interviews. A preliminary performance evaluation of ReqGenie, conducted by comparing its outputs with two existing Software Requirements Specification (SRS) documents, demonstrates promising results. The findings suggest that ReqGenie is particularly supportive for small-scale requirements elicitation tasks, including the software customization in product lines.

Requirements Elicitation

Conversational AI

Large Language Models (LLMs)

Software products

Custom GPT

Författare

Farnaz Fotrousi

Chalmers, Data- och informationsteknik, Interaktionsdesign och Software Engineering

Theocharis Tavantzis

Chalmers, Data- och informationsteknik, Interaktionsdesign och Software Engineering

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

03029743 (ISSN) 16113349 (eISSN)

Vol. 15452 LNCS 352-359
9783031783852 (ISBN)

25th International Conference on Product-Focused Software Process Improvement, PROFES 2024
Tartu, Estonia,

Ämneskategorier

Programvaruteknik

DOI

10.1007/978-3-031-78386-9_25

Mer information

Senast uppdaterat

2024-12-20