Menu
Brasil
home Sumário navigate_before Anterior Atual Seguinte navigate_next
home Sumário
Research Article • Rev. Digit. Bibl. and Cienc. Inf. 23 • 2025 • https://doi.org/10.20396/rdbci.v23i00.8679339/en linkcopy

Open-access Designing chatbots: perspectives and intersections in Information Science

AuthorshipSCIMAGO INSTITUTIONS RANKINGS

    ABSTRACT

    Introduction:  Chatbots originated in the 1960s, but in today's technological landscape, they have emerged as important tools, improving the user experience and providing an agile and personalized service. From the perspective of Information Science (IS), chatbots represent a new channel of communication and interaction with information to facilitate access to large volumes of data, assist in the organization and classification of content, and even in the creation of information recommendation and curation systems. Thus, investigating how CI is being related to the development of chatbots is the question of this research.

    Objective:  To investigate, in the academic-scientific literature, methodologies and technologies for developing chatbots, as well as their main applications and relations with CI. Methodology: A literature review was carried out between July and August 2024.

    Results:  55 documents were analyzed, segmented into five categories: Methodologies, Models, Languages, Platforms and Processing techniques. The categories were designed to explore the diversity of approaches and techniques present in the literature, as well as to understand how different areas are using chatbots to optimize processes and services.

    Results:  Fifty-five documents were analyzed, divided into five categories: Methodologies, Models, Languages, Platforms, and Processing Techniques. The categories were designed to explore the diversity of approaches and techniques present in the literature, as well as to understand how different areas are using chatbots to optimize processes and services.

    Conclusion:  The constant evolution of AI and information retrieval technologies opens up a promising horizon for the development of chatbots. However, despite being related to various techniques and processes, CI is still not given much prominence. Future work suggests that CI techniques and methods be formally integrated into the development of chatbots.

    KEYWORDS
    Chatbot; Organization of information; Information retrieval; Information Science.

    RESUMO

    Introdução:  Os chatbots têm sua origem nos anos 1960, mas, no cenário tecnológico atual, emergem como ferramentas importantes, aprimorando a experiência do usuário e proporcionando um atendimento ágil e personalizado. Do ponto de vista da Ciência da Informação (CI), os chatbots representam um novo canal de comunicação e interação com a informação para facilitar o acesso a grandes volumes de dados, auxiliar na organização e classificação de conteúdos e, até mesmo, na criação de sistemas de recomendação e curadoria de informações. Assim, investigar como a CI está sendo relacionada ao desenvolvimento dos chatbots é o questionamento desta pesquisa.

    Objetivo:  Investigar, na literatura acadêmico-científica, metodologias e tecnologias para o desenvolvimento de chatbots, bem como suas principais aplicações e relações com a CI. Metodologia: Foi realizada uma revisão de literatura cujo protocolo foi executado entre os meses de julho e agosto de 2024.

    Resultados:  Foram analisados 55 documentos, com segmentos em cinco categorias: Metodologias, Modelos, Linguagens, Plataformas e Técnicas de processamento. As categorias foram desenhadas com o intuito de explorar a diversidade de abordagens e técnicas presentes na literatura, além de compreender como as diferentes áreas estão utilizando os chatbots para otimizar processos e serviços.

    Conclusão:  A evolução constante das tecnologias de IA e recuperação da informação abre um horizonte promissor para o desenvolvimento de chatbots. No entanto, apesar de estar relacionada com várias técnicas e processos, a CI ainda ganha pouco destaque. Como trabalhos futuros, sugere-se que técnicas e métodos da CI sejam integrados de maneira formal no desenvolvimento de chatbots.

    PALAVRAS-CHAVE
    Chatbot; Organização da informação; Recuperação da informação; Ciência da Informação.

    1 INTRODUCTION

    Chatbots were first invented in the 1960s by Joseph Weizenbaum with his chatbot ELIZA. Inspired by the Turing test, ELIZA simulated a Rogerian psychologist and engaged in simple conversations with users. Despite its limitations, ELIZA sparked interest in natural language processing (NLP) research, paving the way for future advances (Adamopoulou & Moussiades, 2020). Following the Turing test and the creation of the first chatbots, these applications have been refined and evolved over the years.

    The 1990s were marked by the internet boom, which drove the development and adoption of chatbots. In 1995, ALICE, a chatbot, stood out for its ability to engage in open and complex conversations using NLP and machine learning techniques. Rule-based chatbots such as SmarterChild, which debuted in 2001, also gained popularity by offering customer support and information on various websites (Guimarães, 2022). The rise of instant messaging platforms in the early 2000s revolutionized the chatbot landscape. Another important milestone occurred in November 2022 with the launch of generative artificial intelligence models that combined today's technological landscape. Chatbots have emerged as important tools that improve the user experience by providing fast, personalized service. This innovative technology uses artificial intelligence to interact with users by answering questions, assisting with frequently asked questions, and offering real-time recommendations (Batllori, 2023).

    From an Information Science (IS) perspective, chatbots represent a new channel for communicating and interacting with information. They can facilitate access to large volumes of data, assist in organizing and classifying content, and contribute to the creation of recommendation and information curation systems. Chatbots' customization and learning capabilities allow them to adapt to users' individual needs, offering a more efficient and intuitive experience for searching for and retrieving information (Castor et al., 2021).

    To investigate the relationship between IS and the development of chatbots, this research considered the following guiding questions: What methodologies are used in chatbot development? What are the main technologies and platforms used? What are the main areas of application for chatbots? This study investigated methodologies and technologies for developing chatbots, as well as their main applications and relationships with IS, in the academic and scientific literature.

    The study aims to map, understand, and provide an overview of the evolution of chatbots and their modeling techniques. It also integrates ongoing academic research on chatbots interrelated with IS and information systems, building on the work of Soares and Silva (2024). This work aims to identify gaps and opportunities in the context of IS information retrieval with natural language interaction with users.

    2 LITERATURE REVIEW

    Chatbots and artificial intelligence are shaping the future of human interaction with technology. While they present numerous opportunities, they also bring challenges that need to be addressed. A crucial limitation lies in the ongoing need for up-to-date information to ensure that chatbots remain relevant, useful, and maintain performance over time (Oliveira; Matos, 2023).

    The integration of chatbots in the contemporary digital context highlights the importance of these three pillars of IS: organization, representation, and retrieval of information, which are fundamental topics for the understanding, retrieval, and efficient use of data, both in chatbots and in other contexts. Chatbots designed to simulate human conversations depend on a well-defined information structure to provide accurate and contextualized responses.

    According to Barreto (2002), the objective of the information organization process is to enable and facilitate access to information, which, in turn, has the competence and intention to produce knowledge. The field of Information Organization (IO) encompasses all studies related to the processes and tools used in the organization of information resources of any kind, with the aim of enabling the information needs of a given community of users to be met (Lima, 2020).

    As Hjorland (2010) pointed out, information representation involves describing documents physically and intellectually and organizing these descriptions so users can access and retrieve them. Using metadata and data representation standards, such as natural language, allows chatbots to interpret questions and queries more intuitively. Additionally, implementing NLP and machine learning techniques improves chatbots' ability to understand nuances and contexts in interactions, resulting in responses that are more appropriate.

    Information retrieval in chatbots refers to the ability to locate and present relevant information from a structured database (Moore et al., 2023). Search algorithms and machine learning-based retrieval techniques are often used to optimize this process, ensuring the chatbot can quickly access effectively organized and represented data.

    Regardless of the approach or structure adopted for building chatbots, the relationship between the organization, representation, and retrieval of information is fundamental to these systems' performance. Proper data structuring improves the user experience and increases the effectiveness of chatbots in knowledge management. As technology advances, it becomes necessary to continuously evaluate these dimensions to expand the capacity and usefulness of chatbots in different contexts (Matthews, 2020). In this context, this review investigates chatbot development from an information science perspective, emphasizing the role of information practices in these systems' quality and performance.

    3 METHODOLOY

    According to Gil (2002), a literature review is a critical component of academic work. It provides an overview and summary of relevant research on a specific topic, problem, or time. Kitchenham (2004) defines a literature review as a method that identifies, evaluates, and interprets research relevant to a particular research question, topic, or phenomenon of interest. It also establishes a search strategy with well-defined criteria.

    For this literature review on chatbots, we adapted a protocol based on literature in the field. We considered the following criteria: general objective, questions to be answered, information sources researched, eligibility criteria, inclusion and exclusion criteria, search fields, search expressions, general string, procedures for selecting retrieved documents, analysis procedures, and exclusion criteria after document analysis and processing. See chart 1 for more information.

    Thumbnail
    Chart 1
    Literature Review Protocol

    First, the research sources were defined, and databases with broad coverage in various areas of knowledge were chosen, such as Scopus, Web of Science, SciELO, and the Digital Library of Theses and Dissertations (Biblioteca Digital de Teses e Dissertações - BDTD). The expression and search strings were defined considering the syntax of each database. The terms "chatbot" and its synonyms were used, as well as the terms "methodology" or "construction" and their synonyms, and "information." The following search fields were used: title, abstract, and keywords in the Scopus, Web of Science (WOS), and BDTD databases; and title and keywords in the SciELO database.

    The languages English, Spanish, and Portuguese were selected for the eligibility criteria since most articles are in English, which covers a large part of scientific production in this area. No date limits were established for the searches to avoid restricting the research and to allow retrieval of historical information on the topic. The following document types were selected for the sample: journal articles, event publications, theses and dissertations, and book chapters.

    The inclusion criteria were defined to select appropriate study documents, including those with abstracts, written in English, Spanish, or Portuguese, and related to the development and construction of chatbots. Three exclusion criteria were defined: exclusion of documents not written in the selected languages, exclusion of documents whose titles, abstracts, and keywords did not address the topic of chatbots, and exclusion of documents that were not available in full.

    In the selection procedures, we read the titles and abstracts of the retrieved documents to verify their relevance to the review protocol's overall objective. During the analysis, we sought to identify the concepts, definitions, and applications of chatbots, as well as elements such as construction and methodology. This allowed us to retrieve documents that were strictly relevant to the study. After analyzing the documents, the exclusion criteria were used to discard works without a conceptual, theoretical, or methodological approach.

    The results analysis procedure involved using Rayyan software to screen the abstracts, locate, and exclude duplicate studies. Next, the documents were read, and the titles and abstracts were considered to verify the relevance of the documents to the research. To understand the phenomena investigated, the content analysis technique described by Bardin (2011) was used to treat the results, inferences, and interpretations.

    Bardin (2011) defines content analysis as a set of research techniques that interpret and analyze the content and meaning of various types of documents. Content analysis uses both quantitative (e.g., word frequency) and qualitative (e.g., recurring themes and relationships between concepts) indicators. These indicators describe the content of messages and make inferences about the context in which they were produced (Bardin, 2011).

    Thus, during the thorough reading of the documents, recurring themes were identified and recorded in an Excel spreadsheet for each document. Finally, the identified themes were grouped according to their conceptual relationships, and a faceted classification of the documents was performed. Based on the occurrence of these indicators in the results, the following categories were considered: methodologies, models, languages, platforms, and natural language processing (NLP) techniques.

    The literature review protocol was carried out between July and August of 2024. The data analysis used a mixed methodology based on qualitative and quantitative methods.

    4 RESULTS AND DISCUSSION

    A search of the four selected databases yielded 2,090 documents. Using the Rayyan software, 538 duplicate documents were removed, leaving 1,552 papers for the initial screening. After analyzing the titles and abstracts, a significant 1,478 documents were discarded because they did not meet the inclusion/exclusion criteria defined in the protocol. Reasons for exclusion included focus on the analysis and evaluation of generative language models (e.g., ChatGPT, Bard, and Copilot), use of a language other than English, Portuguese, or Spanish, and lack of access to the full text. Thus, 74 documents were selected for a full reading to verify their relevance to the review's overall objective. The analysis procedures sought to identify the concepts, definitions, and methodological aspects and applications of chatbots. After a thorough reading, 19 documents were excluded due to gaps in data relevant to the review, lack of details and tools for implementation, and limited information on the purpose and application of the chatbot tool. This left a total of 55 documents for analysis and discussion, as illustrated in the document selection flow (Figure 1) and chart (Chart 2).

    Thumbnail
    Chart 2
    Selected documents

    thumbnail of Figure 1
    Figure 1
    Flowchart of document selection and eligibility for review

    thumbnail of Chart 1
    Chart 1
    Distribution of publications by year

    The 55 studies in the sample that presented elements and tools used to build chatbots were composed of journal articles, conference papers, and a master's thesis. The publication dates of the retrieved studies suggest that the topic of chatbots is relatively new, with an increasing number of publications recently, particularly since 2020. As shown in Graph 1, most of the works in the sample were published between 2021 and 2023. Graph 1 shows the distribution of publication frequency over the years.

    The review sought to answer the following question regarding the area of application of chatbots: What are the main areas in which chatbots are used, according to the literature? Several areas have benefited from chatbots' ability to provide fast service, reduce operating expenses, and improve the user experience through automated, personalized interactions. Most of the chatbots described in the study, 26 (47.27%), are concentrated in education. They are used to assist students with questions, provide instant feedback, make study materials available, and support distance learning.

    The second largest number of domains applied to healthcare were identified and developed for activities such as symptom screening, appointment scheduling, medication reminders, smoking cessation, and providing wellness and mental health information. Seven (12.72%) of the chatbots presented in the studies fall into this category. KI appears in third place with four studies (7.27%), followed by computer science with three studies (5.45%). The other areas of application had one study each (1.81%). Virtual assistants in libraries and information services use IS for content curation, catalog navigation support, research guidance, and information literacy. They guide users on how to conduct research, evaluate the credibility of sources, and understand ethical information use practices.

    The authors' affiliations are relevant information for analyzing academic contributions because they can identify the geographical distribution of studies and indicate the institutions leading research in a given area. Figure 2 illustrates the geographical distribution of publications with a heat map. It highlights India as the country with the highest concentration of research, with eight studies, followed by Brazil and Indonesia, with six studies each. The intensity of the blue color reflects the concentration of studies, indicating more intense academic activity in India and a significant research focus in Indonesia and Brazil. Countries such as Thailand, China, and the US are represented in green and show moderate publication output. Countries in beige have fewer studies and suggest a limited contribution to the topic, possibly due to a lack of resources or interest in the area. Finally, white indicates countries that did not contribute any publications to the analyzed sample.

    thumbnail of Figure 2
    Figure 2
    Heat map (authors' affiliations)

    The scope of the studies (see Graph 2) was examined to understand the various approaches to chatbot research and to clarify the discipline in which this technology has been most extensively studied. Computer science (CS) was identified as the field with the most developed and published studies on chatbots, demonstrating its central role in advancing and innovating technological tools. CS has distinguished itself through the creation of new natural language processing (NLP) architectures, machine learning techniques, and approaches that enable the continuous improvement of chatbots in terms of both effectiveness and applicability in different contexts (Mitchell, 2002).

    thumbnail of Graph 2
    Graph 2
    Area of knowledge (authors' categorization)

    Next, in sections 4.2.1 to 4.2.5, the following categories identified will be discussed: Methodologies, Models, Languages, Platforms, and NLP Techniques.

    4.1 Methodologies used

    To investigate the methodologies used in chatbot implementation, we analyzed retrieved documents and identified methodologies and processes. We highlighted not only agile methods, such as Scrum and Design Thinking, but also traditional software development methods, such as the Software Development Life Cycle (SDLC), as well as modeling tools, such as Business Process Model and Notation (BPMN).

    According to Moore et al. (2023), chatbot development involves various methodologies and approaches that facilitate their creation and implementation. Oliveira and Matos (2023) state that chatbot development methods can vary according to the tool's objectives and target audience and include agile approaches, which favor rapid iterations and continuous feedback, and more traditional methods, which follow well-defined stages of planning, development, and evaluation.

    4.1.1 Agile approaches

    Agile approaches emerged in response to the need for greater flexibility and adaptability in software development. In contrast to traditional methods, which proved to be slow and inefficient in the face of constant changes, agile approaches were developed (Schwaber & Sutherland, 2020; Jurafsky & Martin, 2023).

    In 2001, 17 software developers gathered in Utah to explore new approaches that could replace traditional methods. The outcome of this meeting was the "Manifesto for Agile Software Development," a document outlining the core values and principles of agile methodologies (Beck et al., 2001).

    These methodologies have been applied in various fields, offering a flexible, iterative approach to complex projects involving the collaboration of developers, researchers, analysts, scientists, and other professionals (Schwaber & Sutherland, 2020). Agile methods are frequently chosen for contexts requiring quick responses and solutions aligned with user needs due to their ability to adapt and focus on continuous value delivery. Silva's (2022) study reiterates the value of agile methods in IS. In this context, Scrum and Design Thinking emerged as prominent agile methodologies, receiving attention in five studies.

    The Scrum method has proven effective for developing chatbots. Iparraguirre-Villanueva et al. (2023) demonstrated its successful application in developing a chatbot that optimized IT support and incident management. Similarly, Alqaidi, Alharbi, and Almatrafi (2021) described the "Student Assistant," a chatbot with features including automated responses, lost and found item forums, announcements, and volunteer recruitment. They developed it under the Scrum methodology. Muhyiddin and Setiawan (2021) presented the UNYSA chatbot, which aims to improve academic performance and reduce advisors' workload. They also developed it based on Scrum.

    Scrum uses an iterative and incremental approach to help with risk management, improve project predictability, and emphasize transparency, inspection, and adaptation as key principles (Schwaber & Sutherland, 2020).

    Design Thinking, as explored in studies by Chan et al. (2018) and Daniel et al. (2019), is a complementary approach that puts users at the center of the design process. It emphasizes empathy and rapid prototyping to facilitate the creation of chatbots that focus on the user experience. Design Thinking's relevance in chatbot development aligns with IS principles, as Guimarães and Rocha (2021) point out in their research on information practices and Design Thinking. This research addresses the importance of understanding users' needs and contexts to create effective information systems.

    Design thinking involves problem definition, ideation, prototyping, and testing with a focus on empathy and a profound understanding of the user to create more intuitive experiences (Combelles, Ebert, & Lucena, 2020).

    4.1.2 Traditional approaches

    Traditional approaches, also known as "classic" or "prescriptive" models, offer a more structured and sequential method of software development. One of the most widely used models is the Waterfall model, which follows a linear sequence of phases: requirements gathering, design, implementation, testing, and maintenance (Sommerville, 2016). Traditional software development methods are characterized by detailed planning and extensive documentation before development begins.

    In their studies, the authors Kingchang, Chatwattana, and Wannapiroon (2024) and Feitosa et al. (2020) opted to use traditional methods to implement chatbots, demonstrating that these structured approaches are effective, especially in projects requiring greater control and predictability. The Software Development Life Cycle (SDLC) offers a structured framework that ensures greater control and predictability for the project, with well-defined phases of planning, analysis, design, implementation, testing, and maintenance. Integration with existing systems is also an important factor to consider; SDLC helps ensure that the chatbot integrates seamlessly with the organization's systems, avoiding conflicts and failures.

    Sectors such as healthcare and finance benefit from the structured approach of SDLC because reliability is essential in these sectors, and errors can have serious consequences. SDLC minimizes risks and ensures the quality of the chatbot (Feitosa et al., 2020).

    Studies by López et al. (2019) and Feitoza (2021) describe the Business Process Model and Notation (BPMN) methodology for developing chatbots. López et al. (2019) used BPMN to guide processes, and Feitoza (2021) applied the same methodology to develop the Cecílio chatbot for users in federal higher education institutions.

    BPMN is a graphical notation that standardizes business process modeling. It provides clear flowcharts that make processes easy to understand, regardless of technical knowledge. BPMN provides a standardized visual language that enables the modeling, analysis, and optimization of information processes. Similar to CI, BPMN covers the complete information cycle: generation, collection, organization, storage, retrieval, dissemination, and use (López et al., 2019).

    Less frequently found in the analyzed studies and identified in only one study each are Rapid Application Development (RAD), which was described by Rukhiran and Netinant (2022) and by Nasharuddin et al. (2021) in their studies on automated retrieval of postgraduate information and services using the chatbot system and designing an educational chatbot, respectively. A Case Study of CikguAIBot. RAD is a software development methodology that prioritizes rapid prototyping and regular user feedback to deliver applications quickly and efficiently.

    Object-oriented programming (OOP) was described in "The Combination of Natural Language Processing and Entity Extraction for an Academic Chatbot," a paper published by Tommy, Kirana, and Riska (2020). According to Figueiredo et al. (2015), OOP is a programming paradigm that organizes software around objects representing abstractions of the real world, though it is also considered a methodology. These objects combine data (attributes) and behaviors (methods) into a single structure. The main goal is to create modular, reusable, and easy-to-maintain systems. OOP can be used with various development methodologies, such as Scrum, for project management and code structuring. The choice of methodology depends on factors such as project size, problem complexity, and team preferences (Figueiredo et al., 2015).

    The build-and-test methodology is an iterative approach to software development that focuses on continuously building and testing small parts of the information system. This approach prioritizes the user and places them at the center of the process, allowing for the creation of quality products that meet their needs and expectations. In the study "Developing a Chatbot for College Student Program," Chan et al. (2018) used this method to develop the EaSelective chatbot to help students choose elective courses.

    In the methodologies category, agile and iterative methodologies-especially Scrum-stood out as the most commonly used for building chatbots. This reinforces the importance of flexible processes that allow for quick adaptations. Meanwhile, user-centered approaches, such as Design Thinking, and traditional methodologies, such as SDLC and BPMN, also emerged as prominent choices, suggesting a combination of agile practices and structured models. These choices indicate a desire for constant interaction with users and continuous development cycles, both of which are important factors for the success of chatbot projects in a rapidly evolving technological landscape.

    Based on the relationship between these methodologies and the field of IS, we examined how they address the information lifecycle, from creation and organization to retrieval and effective use by users. Incorporating agile and user-centered approaches, such as Scrum and Design Thinking, enhances human-computer interaction, highlighting users as a relevant theme in IS. Similarly, structured practices such as SDLC and BPMN promote a systematic view of data development and management processes, reflecting IS's interdisciplinary nature (Borko, 1968). This nature integrates techniques for classifying, analyzing, and using data to improve decision-making and accessibility of information.

    CI has the potential to play a prominent role in the creation of chatbot systems, since the development methodologies applied are directly influenced by the way data is organized, represented, and retrieved. Thus, the logical organization of information through a planned structure is vital for the chatbot to understand and use information accurately, ensuring more effective interactions with users (Carvalho, 2022).

    CI offers theoretical and practical principles that ensure proper structuring, such as metadata (Alves et al., 2018), enabling chatbots to understand and respond accurately to user interactions.

    For Guimarães (2022), the relationship between KI and the development of artifacts can be considered a transdisciplinary approach, as it allows the expansion of KI's boundaries of contact with fields that go beyond the scientific, such as the development of technological artifacts in the areas of design and engineering.

    4.2 Model

    Natural language processing (NLP)-based models enable chatbots to understand and respond to user questions in a contextual and relevant manner. The models and methodologies applied in bot construction and development should be defined based on the project's specific needs, such as the desired interaction's complexity, cost, and nature (Hyrmet & Arbana, 2021).

    The RASA model was identified in four studies in the results of the review. The authors of the study "Using Artificial Intelligence Methods to Create a Chatbot for University Questions and Answers" (Ramalakshmi et al., 2023) highlight the effectiveness of using RASA for chatbot development, the importance of dynamic information retrieval, and the role of chatbots in improving user engagement in educational contexts. Vasilevich et al. (2022) used the RASA model to develop SmartBot in their study, "Language-Agnostic Knowledge Graphs for Smarter Multilingual Chatbots." SmartBot aims to provide relevant recommendations based on user contributions.

    The studies "Contextual Bot: A Study of WASABI" by Agrawal et al. (2022) and "An Intelligent Chatbot System Based on Entity Extraction Using RASA NLU and Neural Network" by Jiao (2020) presented RASA as an open-source framework for building chatbots and virtual assistants with a focus on natural language processing (NLP) and machine learning (ML). According to the authors, RASA enables the development of robust and sophisticated conversational interfaces with the capacity to comprehend natural language and sustain more organic and engaging conversations.

    In this review, RASA was categorized as both a model (4.2.2) and a platform (4.2.4) because, according to Meshram et al. (2021), it can be understood in both ways when building chatbots. As a model, RASA uses advanced machine learning algorithms, such as reinforcement learning and neural networks, to process natural language, understand user intent, and generate relevant responses. As a platform, RASA offers tools and resources that facilitate chatbot development, deployment, and management. It also provides integration with various messaging platforms, application-programming interfaces (APIs), and databases. This makes RASA a solution for building robust, scalable chatbots (Meshram et al., 2021).

    The BERT model was addressed in four recent studies: by Moore et al. (2023) in A Comprehensive Solution to Retrieval-Based Chatbot Construction; by Mishra et al. (2023) in Advancements in Conversational AI: Building a Mental Health Chatbot with the BERT Model; by Elsayed et al. (2023) in "Chatbot as a Virtual Assistant to Retrieve Information from Datasheets Using Memory Controllers Domain Knowledge"; and by Wang, Liu, and Song (2022) in "Designing an Educational Chatbot with Joint Intent Classification and Slot Filling." BERT is a language model based on the Transformer architecture that stands out for enabling the parallel processing of text sequences and capturing bidirectional relationships between words, considering the context to the left and right. This innovative approach marks a significant advancement in the field of NLP, offering flexibility and precision in understanding the linguistic context. Its ability to be adjusted for a wide variety of NLP tasks without substantial changes to its architecture reinforces its relevance in both research and practical applications.

    Another notable model is the Large Language Model (LLM), also known as the Large-Scale Language Model. LLM models are advanced artificial intelligence models designed to understand and generate text in a manner similar to human language. LLM models are based on deep learning architectures and are trained using large amounts of data, such as texts from books, articles, and websites (Sejnowski, 2023).

    Foosherian et al. (2023) highlight the improvement of pipeline-based chatbots with LLMs in "Enhancing Pipeline-Based Conversational Agents with Large Language Models." In "Introducing a Chatbot to the Web Portal of a Higher Education Institution to Enhance Student Interaction," Oliveira and Matos (2023) present the implementation of a chatbot aimed at improving student interaction within an educational institution. In "Language-Agnostic Knowledge Graphs for Smarter Multilingual Chatbots," Vasilevich et al. (2022) highlight the SmartBot chatbot, which uses a natural language understanding model to extract intentions and entities from user inputs through knowledge graphs. Unlike traditional models, which are based on fixed structures and explicit rules, LLMs rely on deep learning and are trained on large volumes of data. This enables them to interpret complex queries and provide contextualized responses (Sejnowski, 2023). However, LLMs' semantic flexibility presents challenges, such as possible hallucinations and higher computational demand.

    The contrastive learning model has proven to be efficient for developing retrieval-based chatbots, as highlighted by Moore et al. (2023) in A comprehensive solution to Retrieval Based Chatbot Construction. The study highlights the potential of contrastive learning to overcome challenges such as disambiguation, entity recognition, and sentiment analysis, thus enabling chatbots to respond more accurately and naturally to user interactions..

    Contrastive learning models resemble the fundamentals of information retrieval (IR), particularly when it comes to identifying semantic patterns in large amounts of data. Baeza-Yates and Ribeiro-Neto (2013) argue that the effectiveness of IR systems hinges on adequately representing information and ranking documents based on their relevance to user queries. Retrieval quality is associated with the efficient modeling of semantic aspects. Contrastive learning focuses on the distinction between similar and dissimilar examples to improve the semantic representation of information and facilitate the system's understanding of natural language.

    Latent Dirichlet Allocation (LDA) is a probabilistic model that uses natural language processing (NLP) techniques to estimate the probability of word sequences occurring in a text. Toumi et al. (2020) describe how this approach can be applied to construct recommendation chatbots in their study, "Intelligent Chatbot-LDA Recommender System," allowing the identification of relevant topics and linguistic patterns to improve responses.

    Baizal et al. (2023) highlighted the N-gram probabilistic model in their study, "Movie recommender chatbot based on Dialogflow," when implementing an entertainment chatbot for movie recommendations. Widely used in natural language processing (NLP), this model estimates the probability of a word occurring based on N-1 previous words in a text sequence. It analyzes large volumes of text (corpora) and identifies frequency patterns, enabling the prediction of the most likely terms or phrases in certain contexts.

    From an information retrieval (IR) perspective, adopting the N-gram model contributes directly to organizing and retrieving information in automated systems, such as chatbots and digital repositories. N-gram modeling recognizes linguistic patterns that facilitate the identification of user intentions and the accurate recommendation of content, thereby optimizing the search for and access to relevant information. In environments such as digital libraries and recommendation systems, this technique can improve document representation and facilitate mediation between users and available information resources, promoting more contextualized and efficient retrieval.

    Both the N-gram and latent Dirichlet allocation (LDA) models use a probabilistic approach. N-grams estimate the probability of word sequences, and LDA calculates the distribution probabilities of topics in documents (Touimi et al., 2020).

    In the study by Daniel et al. (2019), "Multi-Platform Chatbot Modeling and Deployment with the Jarvis Framework," the Model-Driven Engineering (MDE) model was highlighted. This software development model offers advantages over traditional approaches. By placing models at the center of the process, MDE enables development teams to create complex systems more efficiently and at a lower cost while maintaining high quality.

    In the models category, the analyzed works revealed that the RASA and BERT models were the most adopted, with four occurrences each. This demonstrates the importance of open frameworks and pre-trained models for chatbot development. The growth of Large Language Model (LLM)-based applications, present in three documents, reinforces the trend toward using large-scale language models, in line with advances in AI techniques. Methods such as contrastive learning, latent Dirichlet allocation (LDA), probabilistic N-grams, and machine-driven engineering (MDE) appear less frequently, suggesting they are used for specific modeling needs. Together, these findings demonstrate the diversity of practices and the search for robust solutions in the development of conversational tools.

    The models found are related to IS regarding information organization and retrieval. The model's ability to understand complex linguistic contexts improves information retrieval. According to Saracevic (1995), the relationship between IS and information technologies lies in the application of computational resources aimed at information retrieval and the creation of products and services that promote user satisfaction. Guimarães (2020) agrees, pointing out that the construction of technological artifacts, especially informational ones, is deeply linked to the representation, organization, use, and retrieval of information - aspects traditionally addressed by IS.

    Ferneda (2012) highlights that advances in computer technology have significantly improved information retrieval, enabling more efficient management and access to large volumes of data. For chatbots, information retrieval is essential for providing accurate and relevant responses to users. This reinforces the consensus among authors on the importance of IS in developing technological artifacts.

    4.3 Languages

    Programming languages are conceptualized as formal communication systems that enable programmers to instruct computers to perform specific tasks. These languages play a role in the creation of chatbots. They provide the tools to express algorithms, manipulate data, and control the behavior of hardware and software (Mitchell, 2002). Each language has its syntax and semantics, which define how commands are written and interpreted. In the context of chatbot development, programming languages are essential to modeling and implementing chatbots and computer systems.

    They determine how logic, natural language understanding, and interaction interfaces are implemented, playing a role in the development of chatbots. The choice of language depends on several factors, including the type of chatbot, the environment in which it will be used, and the degree of intelligence required (Srivastava & Prabhakar, 2020).

    To investigate the technologies used, the current and next categories were created to answer the following question: What are the main technologies and platforms used in chatbot development?

    The analysis of the results revealed that Python is a popular language for developing chatbots due to its robustness and the wide range of available NLP and ML libraries. Several studies in the healthcare field, including those by Singla et al. (2024), Mishra et al. (2023), and Bendotti et al. (2024), have demonstrated Python's effectiveness in developing chatbots for medical assistance and patient monitoring. In an educational context, researchers such as Ramalakshmi et al. (2023), Nasharuddin et al. (2021), Muhyidin and Setiawan (2021), and Touimi et al. (2020) have explored using Python for teaching and learning purposes.

    Notably, Nasharuddin et al. (2021) investigated integrating Python with the Telegram platform to enable the development of virtual assistants with continuous learning capabilities that contribute to dynamic adaptation to user needs. These features make chatbots more flexible and effective by improving interaction and response personalization over time in both healthcare and education.

    The studies revealed that JavaScript was the second most notable language for developing chatbots due to its versatility and ability to create interactive web applications. Studies by Lynnyk, Krestyanpol, and Rozvod (2024); Wang, Liu, and Song (2022); Santana et al. (2021); López et al. (2019); and Bendotti et al. (2024), among others, highlight JavaScript's application in various contexts, including chatbots for customer service, technical support, and social interaction. These studies demonstrate the language's wide range of uses, from creating interactive interfaces to implementing complex backend functionalities. This highlights JavaScript's ability to meet the growing technological demands of modern software development.

    PHP (Hypertext Preprocessor), found in five of the reviewed studies, is a popular programming language for web development, particularly for creating dynamic and interactive content. In the context of chatbots, PHP has been used to integrate these tools into web pages and online platforms. Studies by Iparraguirre-Villanueva et al. (2023), López et al. (2019), Agus Santoso et al. (2018), Mendoza et al. (2022), and Sanjaya and Santoso (2021) demonstrate the effectiveness of PHP for various applications, including e-commerce, online services, and interactive communication. Its integration with web servers enables the creation of dynamic, interactive pages, which are essential for modern, responsive platforms. Even with the evolution of technologies, the literature shows that PHP remains an effective choice due to its ability to create interactive experiences and facilitate continuous interaction between users and systems, especially in projects requiring flexibility and dynamism.

    Other languages, such as Cascading Style Sheets (CSS) and Cython, also stood out in specific studies. UNIBOT, a chatbot developed to assist with queries and provide quick responses to users of an educational institution, used CSS to develop its front end. Cython, in turn, was the language used to develop a bot's features for quotes and price, limit, and stock volume queries.

    Therefore, Python stood out as the most used language in the category (eight occurrences), possibly due to its wide range of libraries and tools focused on AI and machine learning, which are essential for developing chatbots. Next, JavaScript and PHP appeared with five references each, reflecting their relevance in the front and back ends of web applications, i.e., full stack. CSS and Cython appear only once, indicating their use for interface styling and performance optimization, respectively. These data reinforce the predominance of languages offering robustness and flexibility to handle the inherent complexity of conversational systems.

    The appropriate programming language for developing chatbots depends on several variables, including demand for interactivity, type of information the system needs to process, need for integration with other platforms, and complexity of desired features. This results in various options offering different solutions to challenges developers face.

    As pointed out by Borko (1968) and Saracevic (1995) in the CI field, programming languages are indispensable tools for developing data organization, retrieval, and analysis systems. Python, for example, is widely adopted for data mining and machine-learning projects, which are strategic areas for dealing with large volumes of information, as Lancaster (2004) has highlighted since the 1970s.

    Additionally, using different programming languages enables the creation of innovative information interfaces, such as interactive catalogs, digital libraries, specialized search systems, and virtual reference services. These technological solutions improve information access and mediation, contributing to user experience and the efficiency of data organization and retrieval processes.

    4.4 Platform

    Chatbot creation platforms offer a set of tools that enable developers to more easily create, train, implement, and manage chatbots. While some platforms allow for advanced customization, they generally do not require in-depth programming knowledge (Touimi et al., 2020). Features such as natural language processing (NLP), scalability, analytics, a user-friendly interface, support for multiple languages, and access to information and services facilitate chatbot management and development.

    Research retrieved in this study indicates that the main benefits of using chatbot development platforms include improved customer experience, reduced costs, and modernized traditional business processes through efficient, artificial intelligence-based interactions (Baizal et al., 2023; Chumkaew, 2023; Iparraguirre-Villanueva et al., 2023).

    The results demonstrate the predominance of Google Cloud's Dialogflow (17 occurrences), followed by RASA (five occurrences) and IBM Watson (four occurrences). Kommunicate and LINE were mentioned in two studies each, while ENGATI, Amazon Lex, the Django Software Foundation, Heroku, Arisa Nest, Discord, Facebook, NeoRouter, and MOOC EAD were mentioned in one study each.

    Tanwar et al. (2023) and Thaiprasert et al. (2023) support this in their studies on the applicability of Dialogflow in highly interactive environments that require immediate responses. The studies by Alqaidi, Alharbi, and Almatrafi (2021); Nasharuddin et al. (2021); and Erekata et al. (2021) confirm the adoption of Google Dialogflow due to its ability to integrate with other technological tools and platforms, such as Google Cloud and external application programming interfaces (APIs). Nasharuddin et al. (2021) and Muhyidin and Setiawan (2021), for example, report that using Dialogflow with data analysis technologies and backend systems improves chatbot effectiveness, making them more responsive and capable of offering personalized experiences.

    Note that RASA was allocated to both category 4.2.2, "Model," and category 4.2.4, "Platform," as described in the studies by Elsayed et al. (2023), "Chatbot as a Virtual Assistant to Retrieve Information from Datasheets Using Memory Controllers Domain Knowledge"; Ramalakshmi et al. (2023), "Using Artificial Intelligence Methods to Create a Chatbot for University Questions and Answers" and Oliveira and Matos (2023), "Introducing a Chatbot to the Web Portal of a Higher Education Institution to Enhance Student Interaction."

    Therefore, Google DialogFlow predominates in the platform category, highlighting its capacity and widespread adoption in chatbot projects. RASA and IBM Watson then emerge as relevant alternatives, and other platforms, such as Kommunicate, LINE, Engati, and Amazon Lex, were mentioned, contributing to the diversity of available options.

    Chatbot creation platforms align with CI practices by enabling the organization, retrieval, and dissemination of knowledge in dynamic, personalized ways. As Borko (1968) pointed out, CI involves generating, collecting, organizing, and using information. Conversational interaction technologies, such as chatbots, enhance these processes by creating environments that immediately respond to users' information needs. Solutions such as DialogFlow, RASA, and IBM Watson are examples of initiatives that aim to improve communication between people and systems and facilitate access to and interpretation of data in multiple contexts. Furthermore, adopting these platforms aligns with Lancaster's (2004) principle that automated systems are essential for advancing information retrieval.

    Chatbots employ NLP techniques and artificial intelligence algorithms to make information flows more efficient and scalable, meeting the growing demands of the digital age (Carvalho, 2022). This perspective reinforces the value of practices related to planning, managing, and evaluating information systems. These practices are essential for developing methodologies and tools that enable increasingly satisfactory interactions between people and metadata. This is in line with the objectives of CI.

    4.5 Natural Language Processing (NLP) Techniques

    Natural language processing (NLP) is an artificial intelligence field dedicated to studying interactions between computers and human language. It involves understanding texts and speech, as well as automatically generating natural language (Nadkarni, Ohno-Machado, & Chapman, 2011). It encompasses computational techniques and models that analyze linguistic structures such as syntax and semantics. It also has practical applications in tasks like machine translation, text classification, chatbots, and automatic summarization (Caseli & Nunes, 2024; Comarella & Café, 2008). By combining knowledge of linguistics, statistics, and machine learning, NLP enables computer systems to efficiently handle the ambiguities, variations, and contexts inherent in human language.

    The following techniques stood out in the reviewed studies: Term Frequency-Inverse Document Frequency (TF-IDF), which measures the relevance of each term in a set of documents; Sequence-to-Sequence (Seq2seq), a neural network-based architecture geared toward tasks such as translation and text generation; and Bag of Words, an approach that converts texts into sets of words, facilitating the identification of lexical and statistical patterns.

    4.5.1 Term Frequency-Inverse Document Frequency (TF-IDF)

    Matthews (2020) presents TF-IDF as a classic and still useful method for information retrieval in simple or keyword-based chatbot systems. The method serves as a point of contrast to show the evolution of techniques in chatbot development, proving to be efficient and widely implemented, and relevant in contexts where the computational costs of LLMs are not feasible. Indexing-based models, such as TF-IDF, have been pillars in information retrieval. The method measures the relevance of a term in a document based on its frequency (TF) and rarity across the corpus (IDF) (Manning, Raghavan, & Schütze, 2008). TF-IDF is a widely used technique for representing text in vector form, considering the frequency of terms in a document and their rarity in the corpus. In retrieval-based chatbots, techniques such as TF-IDF are used to measure the similarity between the user's query and pre-existing documents in the database (Salton, 1968). Despite its simplicity, TF-IDF is effective for queries involving exact keywords and specific knowledge bases.

    In the review studies, authors Baizal et al. (2023), entitled Movie recommender chatbot based on Dialogflow, Tanwar et al. (2023), with AI Based Chatbot for Healthcare using Machine Learning, Thaiprasert et al. (2023), with AI Based Chatbot for Healthcare using Machine Learning, Thaiprasert et al. (2023), with Development of a Class Materials Search System using LINE Chatbot, Santana et al. (2021), com A Chatbot to Support Basic Students Questions, Prakasam et al. (2023), with Design and Development of AI-Powered Healthcare WhatsApp Chatbot, and Attigeri; Agrawal e Kolekar (2024), with Advanced NLP Models for Technical University Information Chatbots: Development and Comparative Analysis, highlighted the use of TF-IDF to implement chatbots in education, healthcare, and entertainment, respectively. The studies mention that chatbots can benefit from TF-IDF, mainly in the area of natural language understanding and information retrieval.

    4.5.2 Seq 2 seq

    The sequence-to-sequence (Seq2Seq) technique was introduced by Sutskever, Vinyals, and Le (2014). It describes a neural network architecture that can transform an input sequence into an output sequence. Seq2seq uses neural networks for this transformation and is applied to activities such as machine translation, summarization, and image captioning (Sutskever, Vinyals, & Le, 2014). Generally, the encoder reads the provided elements (words, characters, or tokens) and condenses the information into a context vector. The decoder then uses this vector to generate the output sequence, one element at a time. This strategy has been used for natural language processing (NLP) tasks, such as machine translation, summarization, and text generation (Karri & Kumar, 2020). Improved versions of Seq2seq use attention mechanisms that allow the decoder to focus on specific parts of the context vector, producing more accurate sequence generation results.

    Karri and Kumar (2020) mentioned the Seq2seq technique in their study, "Deep Learning Techniques for Implementation of Chatbots," which also highlighted parameters that assist in the creation of useful chatbots. These parameters include scalability, interoperability, speed, and the ability to pass the Turing test. The Turing test is a measure of a machine's ability to exhibit behavior indistinguishable from that of a human.

    4.5.3 Bag of words

    The "bag of words" technique converts text into a set of words without considering their order or dependency relationships (Rudkowsky et al., 2018). Though simple, this technique is valuable for classification and topic detection because it provides a statistical representation of linguistic data (Jurafsky & Martin, 2023). In this context, each word becomes a unit, and the text is represented by the frequency with which these terms occur or by their presence or absence (Manning, Raghavan, & Schütze, 2008). Although simple, this approach has proven effective in natural language processing (NLP) applications, such as document classification, sentiment analysis, and topic detection. It provides a statistical representation based on the vocabulary used in the text. This approach was used with Seq2Seq by Karri and Kumar (2020) in Deep Learning Techniques for Implementation of Chatbots.

    In the context of information retrieval (IR), the bag-of-words approach can be useful for organizing and retrieving information in environments that deal with large volumes of textual data, such as institutional repositories, digital libraries, and document search systems (Manning, Raghavan, & Schütze, 2008). The technique contributes to topic mapping, document clustering, and recommending relevant materials to users by enabling automatic content indexing and classification by textual similarity (Jurafsky & Martin, 2023).

    However, the technique's main semantic limitation is that it disregards the context in which words appear. This can result in superficial or inaccurate interpretations, especially when meaning depends on the relationship between terms (Manning, Raghavan, & Schütze, 2008). Thus, although Bag of Words is efficient for exploratory analysis and the initial organization of large text collections, it may not capture semantic nuances that are important for accurate information retrieval.

    In this category, we highlight NLP techniques due to their relevance and prevalence in certain studies. According to Caseli and Nunes (2024), natural language processing (NLP) is a field of research that aims to investigate and propose methods and systems for the computational processing of human language. The term "natural" in the acronym refers to languages spoken by humans, distinguishing them from other types of languages (e.g., mathematical, visual, gestural, and programming).

    Natural language processing (NLP) can be divided into two major subareas: Natural Language Understanding (NLU) and Natural Language Generation (NLG). NLP involves understanding user input and generating corresponding output, which includes components such as automatic text recognition and dialogue management (Caseli & Nunes, 2024, p. 10). These approaches align with Rukhiran and Netinant's (2022) ideas. In their study, "Automated Retrieval of Graduate Information and Services Using the Chatbot System," they highlight that NLP and Information Retrieval (IR) share objectives related to the organization, representation, and retrieval of textual and linguistic information. According to Taskın et al. (2019), NLP develops methods and tools that enable machines to understand, process, and generate human language, thereby expanding the interaction between users and information systems. In turn, IS provides the theoretical basis necessary to structure and retrieve information while considering aspects such as relevance, context, and user needs.

    The interaction between NLP and IS is evident in applications such as search engines, recommendation systems, and chatbots. These systems use NLP techniques, such as semantic analysis, entity extraction, and automatic summarization, to improve the accuracy and accessibility of information (Caseli & Nunes, 2024). Thus, NLP and IS complement each other by combining computational linguistic advances with information organization and retrieval fundamentals, resulting in efficient systems aligned with users' information needs. Each approach meets specific NLP needs and can be combined with other solutions to improve performance and better suit user demands (Yang, 2023).

    4.6. Summary of the literature review

    Upon analyzing all five categories, it became apparent that IS, despite being related to various techniques and processes, receives little attention, even though it is the fundamental element of chatbots. Figure 3 provides an overview of the review based on different aspects related to (1) data on publications and (2) content analysis. The publication data analysis divides the information into location, year of publication, and area of knowledge. For content analysis, five categories were identified using a faceted classification system because some solutions fall into more than one category.

    thumbnail of Figure 3
    Figure 3
    Summary of the literature review

    The literature review sought to understand the techniques and methods used in the development and implementation of chatbots, with an emphasis on the contributions of Information Science (IS). The results reveal that these conversational systems have established themselves as relevant tools for information mediation and retrieval (Patel et al., 2019; Panda; Chakravarty, 2022; Chumkaew, 2023; Moore et al., 2023), especially considering the growing complexity of information (Karri; Kumar, 2020) and the demand for agility, accessibility, and personalization in access to knowledge (Villanueva; Aguilar-Alonso, 2021; Stepanov et al., 2023; Attigeri; Agrawal and Kolekar (2024).

    In this context, an alignment was observed between the classic fundamentals of IS, such as information organization, representation, retrieval, and the techniques used in the construction of chatbots (Rodriguez; Mune, 2022; Elsayed et al., 2023). The approaches identified range from rule-based models to sophisticated deep learning architectures, such as transformers and hybrid systems (Baizal et al., 2023; Foosherian et al., 2023; Wang; Liu; Song, 2022; Whittle; Hall, 2022).

    Despite technological advances, significant gaps remain in the structuring and processing of the information bases that feed these systems (Guimarães, 2022; Buckland, 1991; Lima, 2020). This research focuses on these gaps. Many studies work with scattered documents, such as PDF files, spreadsheets, and content extracted from institutional websites. These documents often do not follow structural standards compatible with automated reading (Lancaster, 2004; Kingchang, Chatwattana, & Wannapiroon, 2024; Santana et al., 2021; Rukhiran & Netinant, 2022). This disorganization directly affects retrieval accuracy and the reliability of responses provided to users.

    Additionally, there is a lack of information curation strategies based on consolidated organizational and representational models. Without these practices, it is difficult to build robust and sustainable systems, particularly in institutional settings like universities, where there is frequent regulatory change and constant updating is necessary (Prakasam et al., 2023; Mishra et al., 2023; Chase, 2024).

    Another recurring issue is the low adoption of semantic structures, such as controlled vocabularies, taxonomies, thesauri, and ontologies. These tools promote disambiguation and terminological standardization and increase the effectiveness of retrieval processes (Allison, 2012; Agus Santoso et al., 2018; Touimi et al., 2020). However, these tools are still rarely incorporated into projects, which compromises the coherence of automated dialogues and the consistency of responses (Hodge, 2000; Soergel, 1999).

    Regarding system maintenance, there was an absence of well-defined procedures for automated updates, version control, and periodic knowledge base validation (Oliveira & Matos, 2023; Mendoza et al., 2022; Singla et al., 2024). This gap is particularly concerning in environments with fast-paced information dynamics, such as public universities.

    Additionally, chatbot evaluation remains focused on traditional computational metrics (e.g., accuracy, F-score), while qualitative informational criteria such as clarity, contextual relevance, user satisfaction, and mediation effectiveness are underaddressed (Feitoza, 2021; Wang, Liu, & Song, 2022).

    Finally, most of the analyzed studies focus predominantly on technical and computational aspects, with little incorporation of the conceptual and methodological dimensions specific to information science (Borko, 1968). These studies present well-structured frameworks, replicable methodologies, and clear guidelines that favor institutional reuse, the consolidation of good practices, and the integrated evolution of systems (Rosenfeld, Morville, & Arango, 2015).

    5 CONCLUSION

    This literature review aimed to investigate methodologies and technologies for developing chatbots, as well as their main applications and relationship with IS. The goal was to create a map that addresses aspects such as methodologies, models, programming languages, platforms, and techniques used in chatbot construction. Initially, 2,040 studies were retrieved. After applying inclusion and exclusion criteria, 55 studies were selected for the bibliographic review. This selection process guaranteed that the chosen studies represented the research's central questions, offering an overview of current chatbot and application trends.

    Four main questions guided the review: What methodologies are used in chatbot development? What are the main technologies and platforms used? What are the main areas of application for chatbots? These questions guided the analysis of the selected studies and the analysis categories, which are: Methodologies, models, languages, platforms, and NLP techniques. These categories were designed to explore the diversity of approaches and techniques in the literature and to understand how different areas use chatbots to optimize processes and services.

    Efficiently organizing and retrieving information through practices such as classification, categorization, and metadata techniques is key to developing effective chatbots that can handle large volumes of data and provide interactive, personalized experiences. In the field of IS, applying these best practices ensures that chatbots continuously evolve as new information enters the system. This allows them to adapt to user needs and improve information retrieval over time. Therefore, the structured organization of data directly contributes to the quality of interactions and the experience provided by chatbots.

    As chatbots become essential tools for organizing and disseminating knowledge, keeping pace with technological developments and the social context is crucial. The success of their implementation depends as much on the language model adopted as on the information organization architecture. The constant evolution of AI technologies and information retrieval methods promises increasingly efficient chatbots that provide users with intuitive, productive experiences accessing, organizing, and retrieving data. In the field of IS, these advances are an important step toward innovative solutions that optimize interaction between users and information. These solutions contribute to more responsive and ethical systems aligned with the demands of the field.

    For future work, we recommend conducting empirical studies that evaluate the implementation and impact of chatbots in informational environments in practice. These studies should consider aspects such as usability, mediation efficiency, and user satisfaction. We also suggest developing methodological frameworks for IS that cover everything from data curation and organization to evaluating results generated by automated systems. Another possibility is constructing information organization models based on the findings of this review to support developing more interoperable solutions aligned with the information field's specific demands.

    Acknowledgments:

    Not applicable

    • Funding:
      Not applicable.
    • Ethical approval:
      Not applicable.
    • Image:
      Taken from the Lattes platform
    • JITA:
      LM. Automatic text retrieval
    • SDG:
      10. Reduced Inequalities

    Availability of data and material:

    Not applicable.

    References

    • ADAMOPOULOU, E.; MOUSSIADES, L. Chatbots: history, technology, and applications. Machine Learning with Applications, Amsterdam, v. 2, n. art. 100006, 2020. Disponível em: https://l1nq.com/T94zz Acesso em: 19 maio 2024.
      » https://l1nq.com/T94zz
    • AGRAWAL, R. et al WASABI contextual BOT. International Journal of Intelligent Systems, Oxford, v. 37, n. 8, p. 5234-5251, 2022. Disponível em: https://ieeexplore.ieee.org/document/9989007 Acesso em: 19 maio 2024
      » https://ieeexplore.ieee.org/document/9989007
    • AGUS SANTOSO, H. et al Dinus Intelligent Assistance (DINA) chatbot for university admission services. In: INTERNATIONAL CONFERENCE ON INFORMATION AND COMMUNICATIONS TECHNOLOGY, 2018, Semarang, Indonesia. Proceedings [..]. Piscataway, NJ: IEEE, 2018. Disponível em: https://ieeexplore.ieee.org/document/8549797 Acesso em: 19 maio 2024.
      » https://ieeexplore.ieee.org/document/8549797
    • ALLISON, D. Chatbots in the library: is it time? Lincoln: UNL Libraries, 2012. Disponível em: https://sl1nk.com/yZ3My . Acesso em: 9 ago. 2025.
      » https://sl1nk.com/yZ3My
    • ALQAIDI, S.; ALHARBI, W.; ALMATRAFI, O. A support system for formal college students: a case study of a community-based app augmented with a chatbot. Educational Technology & Society, Taiwan, v. 24, n. 4, p. 156-171, 2021. Disponível em: https://ieeexplore.ieee.org/document/9759796 Acesso em: 19 maio 2024.
      » https://ieeexplore.ieee.org/document/9759796
    • ALVES, A. P. S. et al Development and evaluation of a chatbot for the Regional Museum of São João del-Rei Chile: Centro Latinoamericano de Estudios en Informática, 2018. Disponível em: https://clei.org/clei2018/docs/SLIS/182920.pdf Acesso em: 8 ago. 2025.
      » https://clei.org/clei2018/docs/SLIS/182920.pdf
    • ATTIGERI, G.; AGRAWAL, A.; KOLEKAR, S. Advanced NLP models for technical university information chatbots: development and comparative analysis. IEEE Access, Piscataway, NJ, v. 12, p. 29633-29647, 2024. Disponível em: https://doi.org/10.1109/ACCESS.2024.3368382 Acesso em: 19 maio 2024.
      » https://doi.org/10.1109/ACCESS.2024.3368382
    • BAEZA-YATES, R.; RIBEIRO-NETO, B. Recuperação de informação: conceitos e tecnologia das máquinas de busca. Porto Alegre: Bookman, 2013.
    • BAIZAL, Z. A. et al Movie recommender chatbot based on Dialogflow. Entertainment Computing, London, v. 13, n. 1, p. 936-947, 2023. Disponível em: https://ijece.iaescore.com/index.php/IJECE/article/view/26761 Acesso em: 19 maio 2024.
      » https://ijece.iaescore.com/index.php/IJECE/article/view/26761
    • BARDIN, L. Análise de conteúdo São Paulo: Edições 70, 2011.
    • BARRETO, A. Transferência da informação para o conhecimento. In: AQUINO, M. A. (org.). O campo da ciência da informação: gênese, conexões e especificidades. João Pessoa: Universitária, 2002.
    • BARRETO, I. et al Development and evaluation of the GISSA mother-baby chatbot application in promoting child health. Ciência & Saúde Coletiva, Rio de Janeiro, RJ, v. 26, n. 5, p. 1679-1690, 2021. Disponível em: https://sl1nk.com/UGNZ8 Acesso em: 19 maio 2024.
      » https://sl1nk.com/UGNZ8
    • BATLLORI, E. Applications of chatbots in education. In: GHAYOOMI, M. Advances in web technologies and engineering book series London: IGI Global Scientific Publishing; Information Science, 2023. p. 80-118, DOI: 10.4018/978-1-6684-6234-8.ch004 Acesso em: 19 maio 2024.
      » https://doi.org/10.4018/978-1-6684-6234-8.ch004
    • BECK, K. et al Manifesto ágil para o desenvolvimento de software 2001. Disponível em: https://agilemanifesto.org/iso/ptbr/manifesto.html Acesso em: 19 maio 2024.
      » https://agilemanifesto.org/iso/ptbr/manifesto.html
    • BENDOTTI, H. et al Introducing quin: the design and development of a prototype chatbot to support smoking cessation. Digital Health, London, v. 26, n. 5, p. 612-620, 2024. Disponível em: https://academic.oup.com/ntr/article/26/5/612/7343970 Acesso em: 19 maio 2024.
      » https://academic.oup.com/ntr/article/26/5/612/7343970
    • BORKO, H. Information science: what is it? American Documentation, Hoboken, NJ, v. 19, n. 1, p. 3-5, 1968. Disponível em: https://sl1nk.com/M8tDa Acesso em: 19 maio 2024.
      » https://sl1nk.com/M8tDa
    • BUCKLAND, M. K. Information as thing. Journal of the American Society for Information Science, Hoboken, NJ, v. 42, n. 5, p. 351-360, 1991. Disponível em: https://l1nq.com/8TqIZ Acesso em: 19 maio 2024.
      » https://l1nq.com/8TqIZ
    • CARVALHO, R. C. Chatbot aplicado à recuperação de informação: um modelo orientado a metadados. 2022. 214f. Tese (Doutorado) - Faculdade de Filosofia e Ciências, Universidade Estadual Paulista (Unesp), Marília, 2022. Disponível em: https://repositorio.unesp.br/items/4aa1e4a4-aaa1-4651-aeee-77c174cb448a Acesso em: 6 jun. 2025.
      » https://repositorio.unesp.br/items/4aa1e4a4-aaa1-4651-aeee-77c174cb448a
    • CASELI, H. M.; NUNES, M. G. V. (org.). Processamento de linguagem natural: conceitos, técnicas e aplicações em português. 2. ed. São Carlos: BPLN, 2024. Disponível em: https://brasileiraspln.com/livro-pln/2a-edicao Acesso em: 6 jun. 2025.
      » https://brasileiraspln.com/livro-pln/2a-edicao
    • CASTOR, E. C. S. et al Chatbot: impactos no ambiente acadêmico de uma universidade do Rio de Janeiro. P2P & Inovação, Rio de Janeiro, v. 8, n. 1, p. 71-92, 2021. Disponível em: https://revista.ibict.br/p2p/article/download/5760/5372/20086 Acesso em: 15 mar. 2024.
      » https://revista.ibict.br/p2p/article/download/5760/5372/20086
    • CHAN, C. H. et al Developing a chatbot for college student programme advisement. In: INTERNATIONAL SYMPOSIUM ON EDUCATIONAL TECHNOLOGY, 2018, Osaka, Japan, Proceedings [..]. Piscataway, NJ: IEEE, 2018. Disponível em: https://ieeexplore.ieee.org/document/8456189 Acesso em: 6 jun. 2025.
      » https://ieeexplore.ieee.org/document/8456189
    • CHASE, M. Academic libraries can develop AI chatbots for virtual reference services with minimal technical knowledge and limited resources. Evidence Based Library and Information Practice, Edmonton, AB, v. 19, n. 2, p. 136-138, 2024. DOI: https://doi.org/10.18438/eblip30523 Acesso em: 9 ago. 2025.
      » https://doi.org/10.18438/eblip30523
    • CHOW, J. C. L.; SANDERS, L.; LI, K. Design of an educational chatbot using artificial intelligence in radiotherapy. AI, San Jose, CA, v. 4, n. 1, p. 319-332, 2023. Disponível em: https://doi.org/10.3390/ai4010015 Acesso em: 9 ago. 2025.
      » https://doi.org/10.3390/ai4010015
    • CHUMKAEW, S. The development of chatbot provided registration information services for students in distance learning. ABAC Journal, Bangkok, v. 43, n. 4, p. 97-112, 2023. Disponível em: https://sl1nk.com/gAZgV Acesso em: 6 jun. 2025.
      » https://sl1nk.com/gAZgV
    • COMARELLA, R. L.; CAFÉ, L. M. Chatterbot: conceito, características, tipologia e construção. Informação & Sociedade, Campina Grande, PB, v. 18, n. 2, p. 55-67, 2008. Disponível em: https://l1nq.com/3A4cy Acesso em: 6 jun. 2025.
      » https://l1nq.com/3A4cy
    • COMBELLES, A.; EBERT, C.; LUCENA, P. Design thinking. IEEE Software, Piscataway, NJ, v. 37, n. 2, p. 21-24, 2020. DOI: https://doi.org/10.1109/MS.2019.2959328 Acesso em: 6 jun. 2025.
      » https://doi.org/10.1109/MS.2019.2959328
    • DANIEL, G. et al. Multi-platform chatbot modeling and deployment with the Jarvis framework. In: GIORGINI, P.; WEBER, B. (ed.) Advanced information systems engineering Cham, Switzerland: Springer, 2019. p. 177-193. Disponível em: https://link.springer.com/chapter/10.1007/978-3-030-21290-2_12 Acesso em: 26 maio. 2025.
      » https://link.springer.com/chapter/10.1007/978-3-030-21290-2_12
    • ELSAYED, H. S. et al Chatbot as a virtual assistant to retrieve information from datasheets using memory controllers domain knowledge. Computer Systems Science and Engineering, Henderson, NV, v. 46, n. 3, p. 3245-3260, 2023. Disponível em: https://ieeexplore.ieee.org/document/10382771 Acesso em: 9 maio 20.
      » https://ieeexplore.ieee.org/document/10382771
    • EREKATA, O. et al Development of a text and speech enabled conversational agent for students' activities planning using dialog flow. In: SINGH, P. K. et al Futuristic trends in network and communication technologies Gateway East, Singapore: Springer, 2021. p. 486-499. Disponível em: https://l1nq.com/z6vAk Acesso em: 9 maio 2024.
      » https://l1nq.com/z6vAk
    • FEITOSA, W. R. et al Proposal for a chatbot prototype in the Plant Health Department of Brazilian Ministry of Agriculture. In: INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE FOR GOOD (AI4G), 2020, Geneva, Switzerland. Proeedings [..]. Piscataway, NJ: IEEE, 2020. Disponível em: https://ieeexplore.ieee.org/document/9311048 Acesso em: 9 maio 2024.
      » https://ieeexplore.ieee.org/document/9311048
    • FEITOZA, F. R. Cecílio: um chatbot para automação do atendimento aos usuários nas Instituições Federais de Ensino Superior. 2021. 100f. Dissertação (Mestrado em Ciência da Computação) - Centro de Educação, Universidade Federal da Paraíba, João Pessoa, 2021. Disponível em: https://sl1nk.com/CyCs9 Acesso em: 10 maio 2024.
      » https://sl1nk.com/CyCs9
    • FERNEDA, E. Introdução aos modelos computacionais de recuperação de informação. Rio de Janeiro: Ciência Moderna, 2012.
    • FIGUEIREDO, K. et. al Uma abordagem gamificada para o Ensino de Programação Orientada a Objetos. In: WORKSHOP SOBRE EDUCAÇÃO EM COMPUTAÇÃO (WEI), 23, 2015, Recife. Anais... Porto Alegre: Sociedade Brasileira de Computação, 2015. DOI: https://doi.org/10.5753/wei.2015.10248 Acesso em: 10 maio 2024.
      » https://doi.org/10.5753/wei.2015.10248
    • FOOSHERIAN, M. et al. Enhancing pipeline-based conversational agents with large language models. AI Communications, Thousand Oaks, CA, v. 36, n. 2, p. 89-108, 2023. Disponível em: https://arxiv.org/abs/2309.03748 Acesso em: 9 maio 2024.
      » https://arxiv.org/abs/2309.03748
    • GIL, A. C. Como elaborar projetos de pesquisa 4. ed. São Paulo: Atlas, 2002.
    • GOWDA, C. P. M. et al Development of information technology telecom chatbot: an artificial intelligence and machine learning approach. 2021. In: 2nd INTERNATIONAL CONFERENCE ON INTELLIGENT ENGINEERING AND MANAGEMENT (ICIEM), 2021, London. Proceedings [..]. Piscataway, NJ: IEEE, 2021. DOI: 10.1109/ICIEM51511.2021.9445354. Disponível em: https://l1nq.com/egIfV Acesso em: 8 ago. 2025.
      » https://doi.org/10.1109/ICIEM51511.2021.9445354.» https://l1nq.com/egIfV
    • GUIMARÃES, L. J. B. Chatbot em contexto: design de experiência do usuário aplicado à recuperação da informação no catálogo de teses e dissertações da CAPES. 2022. 323f. Tese (Doutorado) - Escola de Ciência da Informação, Universidade Federal de Minas Gerais, Belo Horizonte, 2022.
    • GUIMARÃRES, L. J. B.; ROCHA, E. C. F. Práticas informacionais e design thinking abordando usuários 3.0 na Ciência da Informação. RDBCI: Revista Digital de Biblioteconomia e Ciência da Informação, Campinas, v. 19, n. art. e021029, 2021. DOI: https://doi.org/10.20396/rdbci.v19i00.8666871
      » https://doi.org/10.20396/rdbci.v19i00.8666871
    • HJØRLAND, B. The foundation of the concept of relevance. Journal of the American Society for Information Science and Technology, Hoboken, NJ, v. 61, n. 2, p. 217-237, 2010. Disponível em: https://onlinelibrary.wiley.com/doi/abs/10.1002/asi.21261 Acesso em: 8 ago. 2024.
      » https://onlinelibrary.wiley.com/doi/abs/10.1002/asi.21261
    • HNATUSHENKO, V.; OSTROVSKA, K., NOSOV, V. Development and research of a chatbot using the linguistic core of Amazon Lex V2. In: CONFERENCE INTELLIGENT SYSTEMS WORKSHOP AT COLINS, 2024, Lviv, Ukraine. Proceedings [..]. [S.l. : s.n.]. Disponível em: https://ceur-ws.org/Vol-3688/paper5.pdf Acesso em: 8 ago. 2025
      » https://ceur-ws.org/Vol-3688/paper5.pdf
    • HYRMET, M.; ARBANA, K. The impact of chatbots in driving digital transformation. International Journal of E-services and Mobile Applications, Hershey, PA, v. 13, n. 4, p. 88-104, 2021. Disponível em: https://l1nq.com/1qEFE Acesso em: 30 nov. 2024.
      » https://l1nq.com/1qEFE
    • IPARRAGUIRRE-VILLANUEVA, O. et al Agente inteligente para la gestión de incidencias. RISTI - Revista Ibérica de Sistemas e Tecnologias de Informação, Lousada, n. 51, p. 99-115, 2023. Disponível em: https://l1nq.com/Nc4nA Acesso em: 11 jul. 2025.
      » https://l1nq.com/Nc4nA
    • JURAFSKY, D.; MARTIN, J. H. Speech and language processing 3. ed. Cambridge, MA: MIT Press, 2023. Disponível em: https://web.stanford.edu/~jurafsky/slp3/ Acesso em: 8 ago. 2024.
      » https://web.stanford.edu/~jurafsky/slp3/
    • KARRI, S. P. R.; KUMAR, B. S. Deep learning techniques for implementation of chatbots. In: IEEE INTERNATIONAL CONFERENCE ON COMPUTING, COMMUNICATIONS AND NETWORKING TECHNOLOGIES, 2020, Coimbatore, India. Proceedings [..]. Piscataway, NJ: IEEE, 2020. Disponível em: https://ieeexplore.ieee.org/document/9104143 Acesso em: 8 ago. 2024.
      » https://ieeexplore.ieee.org/document/9104143
    • KINGCHANG, T.; CHATWATTANA, P.; WANNAPIROON, P. Artificial intelligence chatbot platform: AI chatbot platform for educational recommendations in higher education. International Journal of Information and Education Technology, Singapore, v. 14, n. 1, p. 34-41, 2024. Disponível em: https://sl1nk.com/2Miq1 Acesso em: 8 ago. 2024
      » https://sl1nk.com/2Miq1
    • KITCHENHAM, B. Procedures for performing systematic reviews Keele, Staffs: Keele University, 2004. (Technical Report). Disponível em: https://www.inf.ufsc.br/~aldo.vw/kitchenham.pdf Acesso em: 15 mar. 2024.
      » https://www.inf.ufsc.br/~aldo.vw/kitchenham.pdf
    • LANCASTER, F. W. Indexação e resumos: teoria e prática. Brasília: Briquet de Lemos, 2004.
    • LIMA, G. A. Representação do conhecimento e da informação no contexto da web. In: LIMA, G. A. (org.). Organização e representação do conhecimento e da informação para web: práticas metodológicas. Rio de Janeiro: Interciência, 2020. p. 1-20.
    • LÓPEZ, A. et al From process models to chatbots. In: GIORGINI, P.; WEBER, B. (ed.) Advanced information systems engineering Cham, Switzerland: Springer, 2019. p. 383-398. Disponível em: https://sl1nk.com/gEb1B Acesso em: 8 ago. 2024.
      » https://sl1nk.com/gEb1B
    • LYNNYK, Y.; KRESTYANPOL, L.; ROZVOD, E. Development of a natural language chatbot interface for website users. Eastern-European Journal of Enterprise Technologies, Kharkiv, v. 1, n. 2 (127), p. 35-44. Disponível em: https://journals.uran.ua/eejet/article/view/299200 Acesso em: 8 ago. 2024.
      » https://journals.uran.ua/eejet/article/view/299200
    • MANNING, C. D.; RAGHAVAN, P.; SCHÜTZE, H. Introduction to information retrieval Cambridge: Cambridge University Press, 2009. Disponível em: https://nlp.stanford.edu/IR-book/pdf/irbookonlinereading.pdf Acesso em: 8 ago. 2024.
      » https://nlp.stanford.edu/IR-book/pdf/irbookonlinereading.pdf
    • MATTHEWS, P. Knowledge organisation systems for chatbots and conversational agents: a review of approaches and an evaluation of relative value-added for the user. In: LYKKE, M. et al (ed.). Knowledge organization at the interface. Baden-Baden: Ergon Verlag, 2020. p. 274-283. (Proceedings of the Sixteenth International ISKO Conference). Disponível em: https://sl1nk.com/beuDT Acesso em: 15 mar. 2024.
      » https://sl1nk.com/beuDT
    • MENDOZA, S. et al A model to develop chatbots for assisting the teaching and learning process. Sensors, Basel, v. 22, n. 15, n. art. 5532, 2022. Disponível em: https://pubmed.ncbi.nlm.nih.gov/35898035/ Acesso em: 9 maio 2024.
      » https://pubmed.ncbi.nlm.nih.gov/35898035/
    • MESHRAM, S. et al Chatbot de consulta universitária usando rasa framework. In: ASIAN CONFERENCE ON INNOVATION IN TECHNOLOGY (ASIANCON), 2021, India. Proceedings [..]. Piscataway, NJ: IEEE, 2021. Disponível em: https://doi.org/10.1109/ASIANCON51346.2021.9544650 Acesso em 10 out. 2024.
      » https://doi.org/10.1109/ASIANCON51346.2021.9544650
    • MISHRA, K. et al Advancements in conversational AI: building mental health chatbot with BERT model. In: INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE FOR INNOVATIONS IN HEALTHCARE INDUSTRIES (ICAIIHI), 2023, India. Proceedings [..]. Piscataway, NJ: IEEE, 2023. Disponível em: https://ieeexplore.ieee.org/document/10489148 Acesso em: 8 ago. 2024.
      » https://ieeexplore.ieee.org/document/10489148
    • MITCHELL, J. Functions and foundations. In: MITCHELL, J. Concepts in programming languages London: Cambridge University Press, 2002. p. 48-90. Disponível em: https://doi.org/10.1017/CBO9780511804175.005 Acesso em 10 out. 2024.
      » https://doi.org/10.1017/CBO9780511804175.005
    • MOORE, K. et al A comprehensive solution to retrieval-based chatbot construction. Computer Speech and Language, London, v. 83, n. art. 101522, 2023. Disponível em: https://doi.org/10.1016/j.csl.2023.101522 Acesso em: 15 mar. 2024.
      » https://doi.org/10.1016/j.csl.2023.101522
    • MUHYIDIN, A.; SETIAWAN, M. A. F. Developing UNYSA chatbot as information services about Yogyakarta State University. Indonesian Journal of Information Systems, Yogyakarta, v. 4, n. 1, p. 67-82, 2021. Disponível em: https://iopscience.iop.org/article/10.1088/1742-6596/1737/1/012038 Acesso em: 8 ago. 2024.
      » https://iopscience.iop.org/article/10.1088/1742-6596/1737/1/012038
    • MUNERA TORRES, M. T.; ALVAREZ, L. M. S.; OSORIO, A. S. Estudio inicial de un chatbot para estudiantes de la modalidad virtual de la Escuela Interamericana de Bibliotecología. Investigación Bibliotecológica, Ciudad de México, v. 36, n. 90, p. 13-30, 2022. Disponível em: https://l1nq.com/ei7dl Acesso em: 09 ago. 2025.
      » https://l1nq.com/ei7dl
    • NADKARNI, P. M.; OHNO-MACHADO, L.; CHAPMAN, W. W. Natural language processing: an introduction. Journal of the American Medical Informatics Association, Oxford, v. 18, n. 5, p. 544-551, 2011. Disponível em: https://pubmed.ncbi.nlm.nih.gov/21846786/ Acesso em: 8 ago. 2024.
      » https://pubmed.ncbi.nlm.nih.gov/21846786/
    • NASHARUDDIN, N. A. et al Designing an educational chatbot: a case study of CikguAIBot. In: INTERNATIONAL CONFERENCE ON INFORMATION RETRIEVAL AND KNOWLEDGE MANAGEMENT (CAMP), 5., 2021, Malaysia. Proceedings [..]. Piscataway, NJ: IEEE, 2021. Disponível em: https://iopscience.iop.org/article/10.1088/1742-6596/1737/1/012038 Acesso em: 9 maio 2024.
      » https://iopscience.iop.org/article/10.1088/1742-6596/1737/1/012038
    • OLIVEIRA, P. F.; MATOS, P. Introducing a chatbot to the web portal of a higher education institution to enhance student interaction. Engineering Proceedings, Basel, v. 56, n. 128, 2023. Disponível em: https://doi.org/10.3390/ASEC2023-16621 Acesso em: 8 ago. 2025.
      » https://doi.org/10.3390/ASEC2023-16621
    • PATEL, N. P. et al. AI and web-based human-like interactive university chatbot (UNIBOT). In: INTERNATIONAL CONFERENCE ON ELECTRONICS, COMMUNICATION AND AEROSPACE TECHNOLOGY (ICECA), 3., 2019, India. Proceedings [..]. Piscataway, NJ: IEEE, 2019. Disponível em: https://ieeexplore.ieee.org/document/8822176 Acesso em: 6 jun. 2024.
      » https://ieeexplore.ieee.org/document/8822176
    • PRAKASAM, S. et al Design and development of AI-Powered healthcare whatsapp chatbot. 2023. In: INTERNATIONAL CONFERENCE ON VISION TOWARDS EMERGING TRENDS IN COMMUNICATION AND NETWORKING TECHNOLOGIES (VITECON), 2., 2023, Vellore, India, 2023. Proceedings [..]. Piscataway, NJ: IEEE, 2023. Disponível em: https://ieeexplore.ieee.org/document/10157423 Acesso em: 8 ago. 2025.
      » https://ieeexplore.ieee.org/document/10157423
    • RAMALAKSHMI, K. et al. Using artificial intelligence methods to create a chatbot for university questions and answers. Journal of Educational Computing Research, Thousand Oaks, CA, v. 61, n. 4, p. 789-812, 2023. Disponível em: https://www.mdpi.com/2673-4591/59/1/16 Acesso em: 24 jun. 2025.
      » https://www.mdpi.com/2673-4591/59/1/16
    • RODRIGUEZ, S.; MUNE, C. Uncoding library chatbots: deploying a new virtual reference tool at the San Jose State University library. Reference Services Review, Bingley, v. 50, n. 1, p. 34-47, 2022. Disponível em: https://scholarworks.sjsu.edu/faculty_rsca/3506/ Acesso em: 24 jun. 2025.
      » https://scholarworks.sjsu.edu/faculty_rsca/3506/
    • ROSENFELD, L.; MORVILLE, P.; ARANGO, J. Information architecture for the web and beyond 5. ed. Sebastopol: O'Reilly, 2015.
    • RUDKOWSKY, E. et al More than bags of words: sentiment analysis with word embeddings. Communication Methods and Measures, Philadelphia, PA, v. 12, n. 2-3, p. 140-157, 2018. Disponível em: https://doi.org/10.1080/19312458.2018.1455817 Acesso em: 24 jun. 2025.
      » https://doi.org/10.1080/19312458.2018.1455817
    • RUKHIRAN, M.; NETINANT, P. Automated information retrieval and services of graduate school using chatbot system. International Journal of Information Management, Oxford, v. 12, n. 5, p. 102-118, 2022. Disponível em: https://ijece.iaescore.com/index.php/IJECE/article/view/27818 Acesso em: 24 jun. 2025.
      » https://ijece.iaescore.com/index.php/IJECE/article/view/27818
    • SAFI, Z. et al Technical aspects of developing chatbots for medical applications: scoping review. Journal of Medical Internet Research, Toronto, v. 22, n. 12, n. art. e19127, 2020. Disponível em: https://doi.org/10.2196/19127 Acesso em: 9 ago. 2025.
      » https://doi.org/10.2196/19127
    • SALTON, G. Automatic information organization and retrieval New York: McGraw-Hill, 1968.
    • SANJAYA, A.; SANTOSO, H. Virtual assistant prototype for property sales. In: INTERNATIONAL CONFERENCE OF COMPUTER AND INFORMATICS ENGINEERING (IC2IE), 4., 2021, Depok, Indonesia. Anais [..]. Piscataway, NJ: IEEE, 2021. Disponível em: https://ieeexplore.ieee.org/document/9649238 Acesso em: 5 set. 2024.
      » https://ieeexplore.ieee.org/document/9649238
    • SANTANA, R. et al A chatbot to support basic students questions. In: LATIN AMERICAN CONFERENCE ON LEARNING ANALYTICS, 4., 2021, Arequipa, Perú. Anais [..]. [S. l. : s. n.], 2021. Disponível em: https://ceur-ws.org/Vol-3059/paper6.pdf Acesso em: 24 jun. 2025.
      » https://ceur-ws.org/Vol-3059/paper6.pdf
    • SARACEVIC, T. A natureza interdisciplinar da ciência da informação. Ciência da Informação, Brasília, DF, v. 24, n. 1, 1995. Disponível em: https://revista.ibict.br/ciinf/article/view/608 Acesso em: 27 nov. 2024.
      » https://revista.ibict.br/ciinf/article/view/608
    • SCHWABER, K.; SUTHERLAND, J. The scrum guide. [S. l. : s. n.], 2020. Disponível em: https://l1nq.com/v1iua Acesso em: 27 nov. 2024.
      » https://l1nq.com/v1iua
    • SEJNOWSKI, T. J. Large language models and the reverse turing test. Neural Computation, Cambridge, MA, v. 35, n. 3, p. 309-342, 2023. Disponível em: https://ieeexplore.ieee.org/abstract/document/10887133 Acesso em: 5 set. 2024
      » https://ieeexplore.ieee.org/abstract/document/10887133
    • SILVA, P. N. Recuperação de Informação na Ciência da Informação: produção acadêmico-científica brasileira (2012-2021). Transinformação, Campinas, v. 35, p. 1-17, 2023. Disponível em: https://sl1nk.com/lgtdG Acesso em: 5 ago. 2025.
      » https://sl1nk.com/lgtdG
    • SILVA, P. N. Métodos ágeis na Ciência da Informação: ensino do Scrum. In: WORKSHOP DE INFORMAÇÃO, DADOS E TECNOLOGIA, 5., 2022, Vitória. Anais [..]. 2022. Disponível em: https://sl1nk.com/WebUo Acesso em: 9 maio 2024.
      » https://sl1nk.com/WebUo
    • SINGLA, A. et al Developing a chatbot to support individuals with neurodevelopmental disorders: tutorial. JMIR Medical Informatics, Toronto, ON, v. 26, n. art. e50182, 2024. Disponível em: https://pmc.ncbi.nlm.nih.gov/articles/PMC11220430/ Acesso em: 24 jun. 2025.
      » https://pmc.ncbi.nlm.nih.gov/articles/PMC11220430/
    • SOARES, J. R.; SILVA, P. N. Panorama da pesquisa sobre chatbots no Brasil. BIBLOS - Revista do Instituto de Ciências Humanas e da Informação, Rio Grande, v. 38, n. 1, p. 199-218, 2024. DOI: https://doi.org/10.63595/biblos.v38i1.17837 Acesso em: 5 ago. 2025.
      » https://doi.org/10.63595/biblos.v38i1.17837
    • SOMMERVILLE, I. Software engineering 10. ed. Boston: Pearson, 2016.
    • SRIVASTAVA, S.; PRABHAKAR, T. Características desejáveis de uma plataforma de construção de chatbot. In: IEEE INTERNATIONAL CONFERENCE ON HUMANIZED COMPUTING AND COMMUNICATION WITH ARTIFICIAL INTELLIGENCE (HCCAI), 2020, Irvine, CA. Proceedings [..]. Piscataway, NJ: IEEE, 2020. Disponível em: https://doi.org/10.1109/HCCAI49649.2020.00016 Acesso em: 9 ago. 2025.
      » https://doi.org/10.1109/HCCAI49649.2020.00016
    • STEPANOV, M. S. et al The automation of client servicing in university and college admission office. In: SYSTEMS OF SIGNALS GENERATING AND PROCESSING IN THE FIELD OF ON BOARD COMMUNICATIONS, 2023, Moscow. Proceedings [..]. Piscataway, NJ: IEEE, 2023. Disponível em: https://ieeexplore.ieee.org/document/10092103 Acesso em: 26 jul. 2024.
      » https://ieeexplore.ieee.org/document/10092103
    • SUTSKEVER, I., VINYALS, O., LE, Q. V. Sequence to sequence learning with neural networks. NeurIPS, 2014. Disponível em: https://arxiv.org/abs/1409.3215 Acesso em: 26 jul. 2024.
      » https://arxiv.org/abs/1409.3215
    • TANWAR, P. et al AI based chatbot for healthcare using machine learning. In: INTERNATIONAL CONFERENCE ON ADVANCES IN COMPUTATION, COMMUNICATION AND INFORMATION TECHNOLOGY (ICAICCIT), 2023, Faridabad, India. Proceedings [..]. Piscataway, NJ: IEEE, 2023. Disponível em: ,https://ieeexplore.ieee.org/document/10465728/ Acesso em: 26 jul. 2024.
      » https://ieeexplore.ieee.org/document/10465728/
    • THAIPRASERT, B. et al Development of a class materials search system using LINE chatbot. In: INTERNATIONAL CONFERENCE ON ELECTRICAL ENGINEERING/ELECTRONICS, COMPUTER, TELECOMMUNICATIONS AND INFORMATION TECHNOLOGY (ECTI-CON), 2023, Nakhon Phanom, Thailand. Proceedings [..]. Piscataway, NJ: IEEE, 2023. Disponível em: https://ieeexplore.ieee.org/document/10153145 Acesso em: 26 jul. 2024.
      » https://ieeexplore.ieee.org/document/10153145
    • TOMMY, L.; KIRANA, C.; RISKA, L. The combination of natural language processing and entity extraction for academic chatbot. In: INTERNATIONAL CONFERENCE ON CYBER AND IT SERVICE MANAGEMENT (CITSM), 8., 2020, Pangkal, Indonesia. Proceedings [..]. Piscataway, NJ: IEEE, 2020. Disponível em: https://ieeexplore.ieee.org/document/9268851 Acesso em: 26 jul. 2024.
      » https://ieeexplore.ieee.org/document/9268851
    • TOUIMI, Y. B. et al Intelligent chatbot-LDA recommender system. International Journal of Emerging Technologies in Learning, Wien, v. 15, n. 20, p. 4-20, 2020. Disponível em: https://online-journals.org/index.php/i-jet/article/view/15657 Acesso em: 26 jul. 2024.
      » https://online-journals.org/index.php/i-jet/article/view/15657
    • VASILEVICH, A. et al Language knowledge graphs for smarter multilingual chatbots. Language Resources and Evaluation, Dordrecht, v. 56, n. 4, p. 1234-1251, 2022. Disponível em: https://ceur-ws.org/Vol-3235/paper2.pdf Acesso em: 26 jul. 2024.
      » https://ceur-ws.org/Vol-3235/paper2.pdf
    • VILLANUEVA, D. P. P.; AGUILAR-ALONSO, I. A chatbot as a support system for educational institutions. In: INTERNATIONAL SCIENTIFIC CONFERENCE ON INFORMATION TECHNOLOGY AND MANAGEMENT SCIENCE OF RIGA TECHNICAL UNIVERSITY (ITMS), 62., 2021, Riga, Latvia. Proceedings [..]. Piscataway, NJ: IEEE, 2021. Disponível em: https://ieeexplore.ieee.org/document/9615271 Acesso em: 9 ago. 2025.
      » https://ieeexplore.ieee.org/document/9615271
    • WANG, Y.; LIU, S.; SONG, L. Designing an educational chatbot with joint intent classification and slot filling. In: IEEE INTERNATIONAL CONFERENCE ON TEACHING, ASSESSMENT AND LEARNING FOR ENGINEERING (TALE), 2022, Hung Hom, Hong Kong. Proceedings [..]. Piscataway, NJ: IEEE, 2022. Disponível em: https://ieeexplore.ieee.org/document/10148291 Acesso em: 20 maio 2024.
      » https://ieeexplore.ieee.org/document/10148291
    • WHITTLE, D.; HALL L. Applying cognitive computing to legal services. In: INTERNATIONAL BCS HUMAN-COMPUTER INTERACTION CONFERENCE (HCI2022), 35., 2022, Keele. Proceedings [..]. [S. l. : s. n.], 2022. Disponível em: https://www.scienceopen.com/hosted-document?doi=10.14236/ewic/HCI2022.35 Acesso em: 8 ago. 2025
      » https://www.scienceopen.com/hosted-document?doi=10.14236/ewic/HCI2022.35
    • YANG, X. Deep learning-based intelligent chatbot in a digital campus. In: INTERNATIONAL CONFERENCE ON DATA ANALYTICS, COMPUTING AND ARTIFICIAL INTELLIGENCE (ICDACAI), 2., Zakopane, Poland. Proceedings [..]. Piscataway, NJ: IEEE, 2023. Disponível em: https://ieeexplore.ieee.org/document/10361205 Acesso em: 20 maio 2024.
      » https://ieeexplore.ieee.org/document/10361205
    • YOSSY, E. H.; BUDIHARTO, W. Knowledge-Based chatbot for humanoid robot in restaurant for question and answering system. ICIC Express Letters, Part B: Applications, Kumamoto, v. 13, n. 3, p. 315-320, 2022. Disponivel em: http://www.icicelb.org/ellb/contents/2022/3/elb-13-03-11.pdf Acesso em: 8 ago. 2025
      » http://www.icicelb.org/ellb/contents/2022/3/elb-13-03-11.pdf
    • ZAMBIASI, S. P.; RABELO, R. J. Arisa Nest - A cloud-based platform for development of virtual assistant. Revista de Informática Teórica e Aplicada, Porto Alegre, v. 27, n. 2, p. 116-126, 2020. Disponível em: https://seer.ufrgs.br/rita/article/view/RITA_VOL27_NR2_116 Acesso em: 8 ago. 2025.
      » https://seer.ufrgs.br/rita/article/view/RITA_VOL27_NR2_116

    Edited by

    Publication Dates

    • Publication in this collection
      31 Oct 2025
    • Date of issue
      2025

    History

    • Received
      04 Mar 2025
    • Accepted
      01 Aug 2025
    • Published
      14 Aug 2025
    Creative Common - by 4.0
    Este é um artigo publicado em acesso aberto sob uma licença Creative Commons.
    vertical_align_top show_chart
    location_on
    Universidade Estadual de Campinas Rua Sérgio Buarque de Holanda, 421 - 1º andar Biblioteca Central César Lattes - Cidade Universitária Zeferino Vaz - CEP: 13083-859 , Tel: +55 19 3521-6729 - Campinas - SP - Brazil
    E-mail: rdbci@unicamp.br
    rss_feed Acompanhe os números deste periódico no seu leitor de RSS
    Reportar erro