AI evidence extraction

ANAQUALI: DEVELOPMENT OF A GENERATIVE AI-BASED ASSISTANT TO SUPPORT BARDIN'S CONTENT ANALYSIS – AN EXPERIENCE REPORT

Jorge Carlos Menezes Nascimento, Vivaldo Gemaque De Almeida, July Ane Almeida Batalha Rodrigues, Thainá Kássia Lima Rabelo, Edna Ferreira Coelho Galvão, Higson Rodrigues Coelho · Artefactum · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.23900/artefactum.v25i3.2975

Methodology & findings

Study design

Experience report with descriptive approach.

Primary method

Design science research approach combined with iterative refinement methodology. The tool was conceptualized to address practical research needs, structured based on classical content analysis principles, tested in pilot studies with qualitative data, and refined through successive adjustments.

Main result

The study found that AnaQuali demonstrated "potencial para optimizar el tiempo de análisis, una mayor sistematización del proceso y apoyo para la formación de investigadores noveles" (potential to optimize analysis time, greater systematization of the process, and support for training early-stage researchers). Additionally, the experience "permite afirmar que a GPT AnaQuaLi se configura como uma inovação relevante no campo da pesquisa qualitativa em saúde, ao contribuir para a organização, sistematização e análise inicial dos dados" (allows us to affirm that AnaQuaLi configures itself as a relevant innovation in the field of qualitative research in health, by contributing to the organization, systematization, and initial analysis of data).

Research paradigm

pragmatist/design science

Author conclusions

The authors conclude that "a experiência de desenvolvimento da AnaQuali, desde a etapa de sua concepção ao pleno uso em pesquisa, destaca que a incorporação de ferramentas fundamentadas em inteligência artificial generativa, sobretudo no campo da pesquisa qualitativa consiste em um movimento promissor, especialmente quanto à organização, sistematização e apoio às etapas iniciais da análise de conteúdo" (the experience of developing AnaQuali, from its conception stage to full use in research, highlights that the incorporation of tools based on generative artificial intelligence, especially in the field of qualitative research, constitutes a promising movement, especially regarding the organization, systematization, and support for the initial stages of content analysis). They also conclude that "a AnaQuali deve ser compreendida como um recurso complementar, cujo potencial depende diretamente da mediação humana qualificada" (AnaQuali should be understood as a complementary resource whose potential depends directly on qualified human mediation).

Risk of bias

Algorithmic bias in model training; Risk of uncritical automation of interpretation; Potential reproduction of social inequalities through biases embedded in training data; Oversimplification of complex analytical contexts; Inconsistency between linguistic coherence and analytical consistency; Algorithmic bias in model training data; Potential amplification of social inequalities through biased training corpora; Risk of uncritical adoption of tool-generated results by novice researchers; Bias in selection of academic sources for model training (limited to Portuguese-language and health education contexts); Single institutional context for tool development (Federal University of Pará); Algorithmic bias in language models; Potential reproduction of social inequalities through biased training data; Risk of uncritical automation of interpretation without human oversight; Dependence on quality of input data

Limitations

The authors state that "a necessidade de mediação humana e o risco de automatização acrítica da interpretação" (the need for human mediation and the risk of uncritical automation of interpretation) are key limitations
Furthermore, they note that "mesmo com treinamento especializado, tais modelos não substituem a interpretação situada do pesquisador, sendo necessário considerar aspectos históricos, sociais e políticos que extrapolam o texto analisado" (even with specialized training, such models do not replace the situated interpretation of the researcher, making it necessary to consider historical, social, and political aspects that go beyond the analyzed text)
Additional limitations include "a necessidade de refinamento dos comandos, a padronização das categorias temáticas e o risco de simplificação interpretativa" (the need for command refinement, standardization of thematic categories, and the risk of interpretive simplification).

Open questions raised

The authors identify future perspectives including: continuous improvement of the tool through testing in different research contexts to expand its training base and enable more global evaluation of applicability in other qualitative approaches; the role of AI-based tools in researcher training by facilitating access to scientific methods and reducing operational barriers, especially in educational contexts; and the need to establish ethical guidelines and methodological directives to orient the tool's use.
Future perspectives include: (1) continuous improvement of the tool through testing in different research contexts to expand its training base and enable broader evaluation of applicability to other qualitative approaches; (2) development of ethical guidelines and methodological directives to guide tool use; (3) exploration of AI-based tools' role in researcher training by facilitating access to scientific methods and reducing operational barriers, especially in educational contexts; (4) consolidation of technologies like AnaQuali dependent on both technical advances and promotion of ethical and methodological guidelines.
The authors identify future perspectives including: continuous improvement of the tool through testing in different research contexts to broaden its training base and enable more comprehensive evaluation of its applicability to other qualitative approaches; recognition that AI-based tools may play an intrinsic role in researcher training by facilitating access to scientific methods and reducing operational barriers, especially in educational contexts; and the need for development of ethical and methodological guidelines to orient tool use.

Extracted from: pdfAgreement 62%

Explore related topics

Qualitative Coding with AI Design Science