PRISEBOT - a chatbot to assist in the development of istar extensions

Abstract

Requirements Engineering is a sub-area of software engineering that is important for identifying the requirements and functionalities of a system. In turn, Goal-Oriented Requirements Engineering (GORE) introduces a new paradigm to requirements elicitation, analysis, and specification, considering the goal as the central concept. iStar is a goal-oriented modeling language representing system requirements through actors, intentional elements, and their relationships. iStar has been widely used to model requirements in academic and industrial projects. Therefore, iStar has been extended to represent new domain-specific concepts and practical aspects. In this context, a process has been proposed to support the creation of new extensions, the PRISE (PRocess to support iStar Extensions) process. PRISE aims to guide development proposals for iStar extensions, making them more complete, consistent, and conflict-free. PRISE is available through BPMN (Business Process Model and Notation) modeling, which makes it possible to navigate the extension statically. On the other hand, interactivity has become feasible with the rise of development and the use of intelligent chats to support a given body of knowledge. Thus, supporting PRISE using a chatbot can make its use more iterative, contributing to greater adoption of the process by extenders. This work proposes interactive and clear support for developing the iStar extension based on PRISE through a chatbot. This solution is accessible via the World Wide Web (WEB) as a support tool for clarifying questions related to PRISE and indicating existing extensions and constructs for various domains. PriseBot was developed using RASA, an open-source framework commonly used to build chat and voice-based AI assistants and Retrieval-Augmented Generation (RAG). We have illustrated PriseBot usage by creating an iStar extension to specify the requirements of chatbot development. Finally, we performed the PriseBot evaluation in two ways: a case study with an extender that used PriseBot during the development of a new iStar extension and a survey with experts in iStar extensions based on the Unified Theory of Acceptance and Use of Technology (UTAUT) method.