Agile requirement analysis approach using artificial intelligent technologies
An approach to requirements analysis using artificial intelligence technologies, taking into account the specifics of the AGILE methodology is proposed in this paper. The approach corresponds to the Model-Driven Methodology, in which the main artifacts of software development are software models rep...
Збережено в:
| Дата: | 2024 |
|---|---|
| Автор: | |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
PROBLEMS IN PROGRAMMING
2024
|
| Теми: | |
| Онлайн доступ: | https://pp.isofts.kiev.ua/index.php/ojs1/article/view/630 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Problems in programming |
| Завантажити файл: |  |
Репозитарії
Problems in programming| id |
pp_isofts_kiev_ua-article-630 |
|---|---|
| record_format |
ojs |
| resource_txt_mv |
ppisoftskievua/ce/beea46c665e6830580ae826e94db7fce.pdf |
| spelling |
pp_isofts_kiev_ua-article-6302025-02-15T11:48:23Z Agile requirement analysis approach using artificial intelligent technologies AGILE підхід аналізу вимог із використанням технологій штучного інтелекту Chebanyuk, O.V. Artificial Intelligence; AGILE; Model-Driven Development; Requirement Analysis; PlantUML; Text to Model Transformation; Requirement Visualization UDC 004.415.2.045 (076.5) штучний інтелект; AGILE; модельно-орієнтована розробка програмного заьезпечення; аналіз вимог до програмного заьезпечення; PlantUML; перетворення «текст в модель»; візуалізація вимог УДК 004.415.2.045 (076.5) An approach to requirements analysis using artificial intelligence technologies, taking into account the specifics of the AGILE methodology is proposed in this paper. The approach corresponds to the Model-Driven Methodology, in which the main artifacts of software development are software models represented by UML diagrams. The proposed approach corresponds to the key ideas of the AGILE manifesto, and is oriented towards the fact that AGILE has a priority to satisfy a customer when he changes requirements. Artificial intelligence technologies serve to prepare initial information for the “Text to Model Transformation” of the requirements specification into those types of UML diagrams (Use Case and Sequence), which are used for requirements analysis. The choice of the UML diagram visualization environment is substantiated.Problems in programming 2024; 2-3: 140-146 У роботі пропонується підхід до аналізу вимог за допомогою технологій штучного інтелекту, враховуючи особливості методології AGILE. Підхід відповідає модельно-оріентованій методології розробки програмного забезпечення, у якому основними артефактами розробки програмного забезпечення є моделі програмного забезпечення, що представляються UML діаграмами. Запропонований підхід відповідає ключовим ідеям AGILE manifesto і орієнтований на те, що вимоги замовника можуть часто змінюватися. Технології штучного інтелекту служать для підготовки початкової інформації для "перетворення з тексту у модель" специфікації вимог у ті види діаграм UML (Use Case та Sequence), які використовуються для аналізу вимог. Обґрунтовано вибір середовища візуалізації UML діаграм.Problems in programming 2024; 2-3: 140-146 PROBLEMS IN PROGRAMMING ПРОБЛЕМЫ ПРОГРАММИРОВАНИЯ ПРОБЛЕМИ ПРОГРАМУВАННЯ 2024-12-17 Article Article application/pdf https://pp.isofts.kiev.ua/index.php/ojs1/article/view/630 10.15407/pp2024.02-03.140 PROBLEMS IN PROGRAMMING; No 2-3 (2024); 140-146 ПРОБЛЕМЫ ПРОГРАММИРОВАНИЯ; No 2-3 (2024); 140-146 ПРОБЛЕМИ ПРОГРАМУВАННЯ; No 2-3 (2024); 140-146 1727-4907 10.15407/pp2024.02-03 en https://pp.isofts.kiev.ua/index.php/ojs1/article/view/630/682 Copyright (c) 2024 PROBLEMS IN PROGRAMMING |
| institution |
Problems in programming |
| baseUrl_str |
https://pp.isofts.kiev.ua/index.php/ojs1/oai |
| datestamp_date |
2025-02-15T11:48:23Z |
| collection |
OJS |
| language |
English |
| topic |
Artificial Intelligence AGILE Model-Driven Development Requirement Analysis PlantUML Text to Model Transformation Requirement Visualization UDC 004.415.2.045 (076.5) |
| spellingShingle |
Artificial Intelligence AGILE Model-Driven Development Requirement Analysis PlantUML Text to Model Transformation Requirement Visualization UDC 004.415.2.045 (076.5) Chebanyuk, O.V. Agile requirement analysis approach using artificial intelligent technologies |
| topic_facet |
Artificial Intelligence AGILE Model-Driven Development Requirement Analysis PlantUML Text to Model Transformation Requirement Visualization UDC 004.415.2.045 (076.5) штучний інтелект AGILE модельно-орієнтована розробка програмного заьезпечення аналіз вимог до програмного заьезпечення PlantUML перетворення «текст в модель» візуалізація вимог УДК 004.415.2.045 (076.5) |
| format |
Article |
| author |
Chebanyuk, O.V. |
| author_facet |
Chebanyuk, O.V. |
| author_sort |
Chebanyuk, O.V. |
| title |
Agile requirement analysis approach using artificial intelligent technologies |
| title_short |
Agile requirement analysis approach using artificial intelligent technologies |
| title_full |
Agile requirement analysis approach using artificial intelligent technologies |
| title_fullStr |
Agile requirement analysis approach using artificial intelligent technologies |
| title_full_unstemmed |
Agile requirement analysis approach using artificial intelligent technologies |
| title_sort |
agile requirement analysis approach using artificial intelligent technologies |
| title_alt |
AGILE підхід аналізу вимог із використанням технологій штучного інтелекту |
| description |
An approach to requirements analysis using artificial intelligence technologies, taking into account the specifics of the AGILE methodology is proposed in this paper. The approach corresponds to the Model-Driven Methodology, in which the main artifacts of software development are software models represented by UML diagrams. The proposed approach corresponds to the key ideas of the AGILE manifesto, and is oriented towards the fact that AGILE has a priority to satisfy a customer when he changes requirements. Artificial intelligence technologies serve to prepare initial information for the “Text to Model Transformation” of the requirements specification into those types of UML diagrams (Use Case and Sequence), which are used for requirements analysis. The choice of the UML diagram visualization environment is substantiated.Problems in programming 2024; 2-3: 140-146 |
| publisher |
PROBLEMS IN PROGRAMMING |
| publishDate |
2024 |
| url |
https://pp.isofts.kiev.ua/index.php/ojs1/article/view/630 |
| work_keys_str_mv |
AT chebanyukov agilerequirementanalysisapproachusingartificialintelligenttechnologies AT chebanyukov agilepídhídanalízuvimogízvikoristannâmtehnologíjštučnogoíntelektu |
| first_indexed |
2025-07-17T09:52:31Z |
| last_indexed |
2025-07-17T09:52:31Z |
| _version_ |
1850410095249195008 |
| fulltext |
140
Методи та засоби програмної інженерії
UDC 004.415.2.045 (076.5) http://doi.org/10.15407/pp2024.02-03.140
O.V. Chebanyuk
AGILE APPROACH TO REQUIREMENT ANALYSIS USING
ARTIFICIAL INTELLIGENT TECHNOLOGIES
An approach to requirements analysis using artificial intelligence technologies, taking into account the specifics of
the AGILE methodology is proposed in this paper. The approach corresponds to the Model-Driven Methodology, in
which the main artifacts of software development are software models represented by UML diagrams. The proposed
approach corresponds to the key ideas of the AGILE manifesto, and is oriented towards the fact that AGILE has a
priority to satisfy a customer when he changes requirements.
Artificial intelligence technologies serve to prepare initial information for the “Text to Model Transformation” of the
requirements specification into those types of UML diagrams (Use Case and Sequence), which are used for
requirements analysis.
The choice of the UML diagram visualization environment is substantiated.
Keywords: Artificial Intelligence, AGILE, Model-Driven Development, Requirement Analysis, PlantUML, Text to
Model Transformation, Requirement Visualization.
О.В. Чебанюк
AGILE ПІДХІД АНАЛІЗУ ВИМОГ ІЗ ВИКОРИСТАННЯМ
ТЕХНОЛОГІЙ ШТУЧНОГО ІНТЕЛЕКТУ
У роботі пропонується підхід до аналізу вимог за допомогою технологій штучного інтелекту, враховуючи
особливості методології AGILE. Підхід відповідає модельно-оріентованій методології розробки програмного
забезпечення, у якому основними артефактами розробки програмного забезпечення є моделі програмного за-
безпечення, що представляються UML діаграмами.
Запропонований підхід відповідає ключовим ідеям AGILE manifesto і орієнтований на те, що вимоги замов-
ника можуть часто змінюватися. Технології штучного інтелекту служать для підготовки початкової інфор-
мації для “перетворення з тексту у модель” специфікації вимог у ті види діаграм UML (Use Case та Sequence),
які використовуються для аналізу вимог.
Обґрунтовано вибір середовища візуалізації UML діаграм.
Ключові слова: штучний інтелект, AGILE, модельно-орієнтована розробка програмного заьезпечення, аналіз
вимог до програмного заьезпечення, PlantUML, перетворення «текст в модель», візуалізація вимог.
Introduction
One of the requirements of AGILE is
a quick response to changes in customer re-
quirements, which can happen quite often.
This implies changing all software develop-
ment artefacts that are associated with a spe-
cific requirement. The information contained
in these artefacts must quickly be synchro-
nized with the changes made by the customer.
To successfully implement this approach, the
process of synchronizing changes needs to be
automated, which can significantly speed up
development and improve the quality of the
source code.
For today, perspective and modern ap-
proaches use of methods that combine the ar-
tificial intelligence technologies and funda-
mental knowledge of model-driven engineer-
ing [1]. This paper proposes such an ap-
proach.
Review papers
Foundations of requirement analysis
by means of chat bots and agents were de-
scribed in the paper [2]. Researchers have ex-
plored the development of Bots or agents that
can engage in natural language conversations
© O.В. Чебанюк, 2024
ISSN 1727-4907. Проблеми програмування. 2024. №2-3
141
Методи та засоби програмної інженерії
with stakeholders (Figure 1). These automat-
ed agents assist in the process of eliciting re-
quirements from users and stakeholders. By
imitating human interviewers, these bots can
ask questions, provide clarifications, and
guide stakeholders through the requirements
gathering process.
One specific example mentioned in
the paper is LadderBot. LadderBot mimics a
human interviewer by conversing in natural
language. It formulates questions and pro-
vides assistance during the requirements elici-
tation process. The main steps of requirement
analysis using LadderBot are the next
- Natural Language Interaction:
LadderBot interacts with users
through a chat interface, mimick-
ing human conversation.
- AI-Driven Analysis: LadderBot
uses Artificial Intelligence (AI) al-
gorithms to process user input.
- LadderBot identifies patterns,
keywords, and context to under-
stand user needs.
- The AI extracts the information re-
lated to requirements.
- Dynamic Adaptation: Based on
user responses, LadderBot dynam-
ically adapts its conversation.
- It formulates follow-up questions
to explore different aspects of re-
quirements.
Task and research questions
Task: to propose an approach for
AGILE requirement analysis using artificial
intelligence tools. In order to perform this
task it is necessary to solve the next research
questions (RQs):
RQ1: Ground a choice of a visualiza-
tion environment to perform “text to model
transformation” operation.
RQ2: Propose the important steps of
AGILE requirement analysis approach.
RQ3: Conduct an experiment with
several AI tools to compare the obtained re-
sults of requests.
Fig. 1. Description of the process using chat bot for requirement analysis
142
Методи та засоби програмної інженерії
RQ4: Analyse the experimental result
and ground a choice of an AI tool to perform
the AGILE requirement analysis approach.
Analysis of the research questions al-
lows us to formulate the scientific novelty of
the conducted research.
The paper proposes a requirement
analysis approach that allows to avoid a hu-
man factor and save time performing the next
activities:
- to prepare a full and non-
contradictory requirement specifi-
cation from any product vision
document using artificial intelli-
gence technologies.
- It also allows to design software
models for requirement analysis
that correspond to the requirement
specification using artificial intel-
ligence technologies.
Model-Driven Engineering
foundations of the proposed
approach
Analysis of “text to model transfor-
mation” modelling environments.
Aim of this analysis is to select model-
ling environment with the simplest represen-
tation of textual description of UML diagram.
Simple representation requires minimum ef-
forts to teach AI tools to prepare a correct
and full text representation of UML diagram.
Figure 2 represents a classic model to model
transformation scheme with propositions
(blue text labels) of elements’ names that par-
ticipate in the proposed approach.
The text to model transformation is
done by modelling environment. The next
modelling environments were considered:
- Visual studio plug-in for class dia-
gram generating;
- DrawIO;
- Luquidchart;
- PlantUML;
- ASTAH UML.
Because of limited value of paper, the
detailed analysis of modelling environments
is not represented.
As a result of modelling environment
analysis, PlantUML was chosen [8]. The cri-
teria that ChatGTP and Gemini were learned
to generate correct and full text description of
PlantUML Sequence and Use Case diagrams.
Fig. 2. Classic “Model to Model transformation” scheme with description of key
elements necessary for transformation. Figure is taken from [7]
143
Методи та засоби програмної інженерії
AGILE Requirement Analysis
approach
In order to realize the proposed ap-
proach the next actors are involved: Custom-
er, Requirement Engineer, AI, and Domain
Analyst. UML Sequence Diagram is repre-
sented on figure 3. The description of the
main ideas of the proposed approach is pre-
sented by roles of every actor.
Customer: The customer’s role is to
prepare the Product Vision Document and to
provide feedbacks during the Scrum meeting
if requirement clarification is needed.
Requirements Engineer: receives the
Product Vision Document from the Customer,
then verifies UML diagram, obtained after the
next iteration of a domain analysis and ex-
plains the UML diagrams to the customer dur-
ing the Scrum meeting.
AI: The AI is involved in designing
Epics, User Stories, and UML diagrams (Use
Case and Sequence Diagrams) based on the
instructions from the Domain Analyst. It also
helps in formulating and evaluating key ques-
tions to the client, finding answers to these
questions, and creating the Requirement
Specification. During the Requirement Clari-
fication loop, the AI refines the requirement
specification and UML diagrams as per the
Domain Analyst’s instructions.
Domain Analyst: The Domain Ana-
lyst instructs the AI to design User Stories
and UML diagrams, formulates key questions
to the Customer. The Domain Analyst also
evaluates and corrects the questions formulat-
ed by the AI. During the Scrum meeting, the
Domain Analyst considers Customers’ and
Requirements’ Engineer notes about require-
ment specification and UML diagrams. It
gives input information for the next stage of
the Requirement Clarification. The next activ-
ity of the Domain Analyst is to instruct the AI
to refine the User Stories and UML diagrams,
and corrects the Requirement Specification if
needed.
Experimental research of the
proposed approach
Consider requirement specification of
software system, describing rental processes
of sport equipment.
Fig. 3. UML sequence diagram of the proposed AGILE requirement analysis approach
144
Методи та засоби програмної інженерії
The system presents various sports
equipment on the company's website, each
with a specific name, price, and unit of meas-
urement. Customers can rent equipment, and
their standard questionnaire data, phone, and
email address are collected for communica-
tion. The system automatically records the
customer, equipment, quantity, rental date,
and return date for each rental. The rental sys-
tem manages the availability and condition of
the equipment. After each return, the equip-
ment undergoes a thorough cleaning and in-
spection process. Any necessary repairs are
carried out immediately. If the equipment is
damaged beyond repair, it is replaced. The
system also handles issues related to the lack
of information about the availability of the
necessary equipment in the warehouse in the
required quantity.
Customers can track their rental histo-
ry online. This feature provides detailed in-
formation about their past rentals, including
the types of equipment rented, rental dates, re-
turn dates, and costs. Based on the total cost
of the order, the system provides additional
discounts. These features allow customers to
manage their rentals effectively and plan for
future ones.
Requirement analysis is provided with
three different chat bots. Bing Copilot,
AIchatting, and AI Chat.
Because of limited value of paper, on-
ly essential prompts to AI tools and analysis
of their answers are represented below.
Domain Analyst activities:
Prompt 1 Hello I have a description of
software system. Write please epics, user sto-
ries and acceptance criteria for them.
Rental Process and Customer Interac-
tion: {text of the requirement specification.}
Result: All AI networks have de-
signed well and clear user stories with differ-
ent level of description (see Table 1).
Prompt 2 Please generate me ten the
most important questions about problem do-
main having User Stories and epics.
Prompt 2 1. Please define to which ep-
ic and a user story are related to which ques-
tion.
Prompt 3 Please find answers to these
questions and verify the description of soft-
ware system. Please mark changes of the de-
scription by bold font and do not forget about
the references. Thank you!!!
Result: Bing Copilot has changed the
text of specification correctly adding details
to description of the product vision document.
Bing Copilot Gemini Aichatting
Prompt 1
Number of User Stories Six user stories Ten user stories Three user stories
Accuracy of User stories
description (from 0 to 10)
4
Very common
descriptions
10
Well systematization
2
Two user stories and only
couple of aspects are
covered
Prompt 2 and Prompt 3
Generated questions are
related to
Four user stories All epics + additional
user stories
All three user stories were
précised
Good precision Weak precision
Prompt 4
Estimation of UML
diagrams
Clear UML diagrams, describing user
stories (Sequence diagrams) and epics
(Use Case diagrams)
Stop to work (limited
number of prompts with
registered account)
Таble 1
Analysis of answers of different AI tools
145
Методи та засоби програмної інженерії
Gemini generated additional user stories then
structured them correctly too.
Prompt 4 May you generate a
PlantUml description of use case and se-
quence diagrams from the improved require-
ment specification of the renting system?
Result: Gemini got correct templates
of the UML diagrams and then improved de-
scription of user stories. Bing Copilot pro-
posed correct and clear description of UML
diagrams.
Estimation of answers for different AI
tools is represented in Table 1.
Conclusion
The paper presents the AGILE re-
quirements engineering approach, which al-
lows the use of artificial intelligence tools.
The approach effectively solves the following
key tasks of requirements analysis:
- synchronization of customer re-
quirements with the content of the
requirements specification, epics,
user stories, and UML use case
and sequence diagrams;
- using of artificial intelligence tools
for the design software develop-
ment artefacts that are used for re-
quirement analysis. Experimental
analysis, aimed to define AI tools
with the best capabilities for refin-
ing software development arte-
facts, is represented. The most ap-
propriate results were obtained by
Bing Copilot and Gemini.
Acknowledgement
This paper is performed as a part of a
research project “Ingeniería de dominio para
los desarrollos de inteligencia artificial” (Do-
main engineering for artificial intelligence
developments) de Instituto de Investigación
en Inteligencia Artificial (IIIA, Catalonia,
Spain), Consejo Superior de Investigaciones
Científicas (CSIC, Spain).
References
1. Rietz, T., 2019. Designing a conversa-
tional requirements elicitation system
for end-users. In: 2019 IEEE 27th In-
ternational Requirements Engineering
Conference (RE), pp. 1-6.
doi:10.1109/RE.2019.00061.
2. Kaur, K. & Kaur, P., 2024. The appli-
cation of AI techniques in require-
ments classification: a systematic
mapping. Artificial Intelligence Re-
view. Available at:
https://link.springer.com/chapter/10.1
007/978-981-15-7907-3_20 (Ac-
cessed: 09 April 2024).
3. Mognon, F. & Stadzisz, P.C., 2017.
Modeling in Agile Software Devel-
opment: A Systematic Literature Re-
view. In: Agile Methods (WBMA
2016), pp. 50–59. Springer. Available
at:
https://link.springer.com/chapter/10.1
007/978-3-319-55907-0_5.
4. Bugeja, M., 2024. Artificial Intelli-
gence. Science News, 22 February.
Available at:
https://www.sciencenews.org/topic/art
ificial-intelligence. Publisher: Society
for Science & the Public.
5. Stapleton, J. & Subramaniam, M.,
2023. Impact of Agile Methodology
Use on Project Success in Organiza-
tions - A Systematic Literature Re-
view. In: Agile Project Management,
pp. 239-260. Springer. Available at:
https://link.springer.com/chapter/10.1
007/978-3-030-63322-6_21. Publish-
er: Springer Nature.
6. Ciccozzi, F., Malavolta, I. & Selic, B.,
2019. Execution of UML models: a
systematic review of research and
practice. Software Systems Modeling,
18, pp.2313–2360. Available at:
https://doi.org/10.1007/s10270-018-
0675-4
7. Figure model to model transformation
is taken from
https://wiki.eclipse.org/images/9/90/O
MCW_chapter10_Modelplex-WP6-
Training_IntroductionToM2M.pdf
8. PlantUML https://plantuml.com/
Одержано: 10.04.2024
Внутрішня рецензія отримана: 19.04.2024
Зовнішня рецензія отримана: 28.04.2024
146
Методи та засоби програмної інженерії
Про автора:
1 Olena Chebanyuk,
Contract researcher
https://orcid.org/0000-0002-9873-6010
Місце роботи автора:
1 Instituto de Investigación en
Inteligencia Artificial (IIIA)
Consejo Superior de Investigaciones
Científicas (CSIC).
E-mail: elena.chebanyuk@iiia.csic.es
Сайт:
https://www.iiia.csic.es/es/people/person/?per
son_id=207
|