Vol. 13/ Núm. 1 2026 pág. 1119
https://doi.org/10.69639/arandu.v13i1.1966
Talkpal as a Digital Tool to Support B1 English Speaking
Skill Development
Talkpal como herramienta digital para apoyar el desarrollo de la habilidad de hablar
inglés B1
Sonia Elizabeth Diaz Díaz
sonia.diazec@gmail.com
https://orcid.org/0009-0008-1929-6602
Universidad Bolivariana del Ecuador
Ecuador – Guayaquil
Emily Rocio Salazar Sánchez
emilysalazar2123@gmail.com
https://orcid.org/0009-0009-7835-2106
Universidad Bolivariana del Ecuador
Ecuador-Guayaquil
Diana Carolina Egas Herrera
dcegash@ube.edu.ec
https://orcid.org/0000-0003-2878-0689
Universidad Bolivariana del Ecuador
Ecuador-Guayaquil
Josué Reinaldo Bonilla Tenesaca
jrbonillat@ube.edu.ec
https://orcid.org/0000-0002-6748-2345
Universidad Bolivariana del Ecuador
Ecuador-Guayaquil
Artículo recibido: 10 diciembre 2025 -Aceptado para publicación: 18 enero 2026
Conflictos de intereses: Ninguno que declarar.
ABSTRACT
The implementation of artificial intelligence (AI) tools in second language learning is presented
as a transformative strategy for promoting communicative competence, one of the most
significant challenges for students and one that teachers seek to optimize. This article aims to
conduct an in-depth analysis of the impact of the digital tool Talkpal, a personalized
conversational platform powered by AI that offers an interactive learning experience, as a
scaffolding strategy to optimize oral proficiency in B1 English learners. To achieve this objective,
this research employs a mixed-methods approach, combining quantitative and qualitative
methods to analyze students’ performance before and after using the Talkpal application and to
obtain a comprehensive understanding of the relationship between the dependent and independent
variables. Results were obtained through various instruments, including classroom observation,
initial and final survey on students’ perception of the Talkpal tool, pre- and post-intervention
speaking tests assessed using a B1-level rubric, and exit tickets collected after each session. The
Vol. 13/ Núm. 1 2026 pág. 1120
findings, based on a sample of 22 B1-level students, suggest that the conversational tool creates
a real, dynamic, adaptive, and interactive practice environment, leading to improvements in
fluency and pronunciation, as well as greater student motivation and confidence in oral
communication.
Keywords: artificial intelligence, Talkpal, speaking skill, B1 learners
RESUMEN
La implementación de herramientas con inteligencia artificial en el aprendizaje dentro de una
segunda lengua se presenta como una estrategia transformadora para promover el dominio de la
competencia comunicativa, uno de los retos más significativos para los estudiantes y que los
docentes buscan optimizar. Este artículo tiene como objetivo analizar en profundidad el impacto
de la herramienta digital Talkpal, una plataforma conversacional y personalizada impulsada por
IA con experiencia de aprendizaje interactivo como estrategia de andamiaje para optimizar la
competencia oral en estudiantes de inglés B1. Para alcanzar el objetivo la investigación emplea
un enfoque mixto, combinando métodos cuantitativo y cualitativo para analizar el rendimiento
antes y después del uso de la aplicación Talkpal y obtener una comprensión integral entre las
variables dependiente e independiente. Los resultados se obtuvieron mediante instrumentos como
la observación, encuesta inicial y final sobre la percepción de la herramienta Talkpal, pruebas de
expresión oral previas y posteriores evaluadas con una rúbrica de nivel B1, y “exit tickets”
después de cada sesión. Los hallazgos, aplicados a una muestra de 22 estudiantes de nivel B1,
sugieren que la herramienta conversacional genera un entorno de práctica real, dinámica,
adaptativa e interactivo, lo que lleva a una mejora en la fluidez y pronunciación, junto con una
mayor motivación y confianza del alumnado en la comunicación oral.
Palabras clave: inteligencia artificial, TalkPal, habilidad oral, estudiantes de nivel B1
Todo el contenido de la Revista Científica Internacional Arandu UTIC publicado en este sitio está disponible bajo
licencia Creative Commons Atribution 4.0 International.
findings, based on a sample of 22 B1-level students, suggest that the conversational tool creates
a real, dynamic, adaptive, and interactive practice environment, leading to improvements in
fluency and pronunciation, as well as greater student motivation and confidence in oral
communication.
Keywords: artificial intelligence, Talkpal, speaking skill, B1 learners
RESUMEN
La implementación de herramientas con inteligencia artificial en el aprendizaje dentro de una
segunda lengua se presenta como una estrategia transformadora para promover el dominio de la
competencia comunicativa, uno de los retos más significativos para los estudiantes y que los
docentes buscan optimizar. Este artículo tiene como objetivo analizar en profundidad el impacto
de la herramienta digital Talkpal, una plataforma conversacional y personalizada impulsada por
IA con experiencia de aprendizaje interactivo como estrategia de andamiaje para optimizar la
competencia oral en estudiantes de inglés B1. Para alcanzar el objetivo la investigación emplea
un enfoque mixto, combinando métodos cuantitativo y cualitativo para analizar el rendimiento
antes y después del uso de la aplicación Talkpal y obtener una comprensión integral entre las
variables dependiente e independiente. Los resultados se obtuvieron mediante instrumentos como
la observación, encuesta inicial y final sobre la percepción de la herramienta Talkpal, pruebas de
expresión oral previas y posteriores evaluadas con una rúbrica de nivel B1, y “exit tickets”
después de cada sesión. Los hallazgos, aplicados a una muestra de 22 estudiantes de nivel B1,
sugieren que la herramienta conversacional genera un entorno de práctica real, dinámica,
adaptativa e interactivo, lo que lleva a una mejora en la fluidez y pronunciación, junto con una
mayor motivación y confianza del alumnado en la comunicación oral.
Palabras clave: inteligencia artificial, TalkPal, habilidad oral, estudiantes de nivel B1
Todo el contenido de la Revista Científica Internacional Arandu UTIC publicado en este sitio está disponible bajo
licencia Creative Commons Atribution 4.0 International.
Vol. 13/ Núm. 1 2026 pág. 1121
INTRODUCTION
In recent years, the educational landscape has experienced a significant transformation,
driven largely by the integration of digital tools. These technological resources have transformed
traditional teaching and long-established learning paradigms, offering learners greater autonomy
and opportunities to practice beyond the classroom (Du & Daniel, 2024). The adoption of mobile
devices and remarkable advancements in the field of natural language processing (NLP) and
artificial intelligence (AI) have accelerated the use of such tools across diverse educational
contexts globally, making learning more accessible and engaging for students (Chen, 2023).
Educators worldwide are continuously seeking innovative methods to support
communicative competence in English as a Foreign Language (EFL), especially for intermediate
learners aiming to reach advanced levels of proficiency. As part of this movement, AI-powered
tools such as Talkpal have gained significant attention for their potential to enhance speaking
skills through interactive practice, personalized feedback, and adaptive learning pathways (Pozo-
Sánchez et al., 2021). The capacity of these tools to provide an instant error correction, specialized
exercises, and activities focused on individual performance has made a remarkable difference
compared to traditional methods, fostering a more effective learning experience (Lee, S. M.
2019).
However, despite all these technological advancements, speaking remains a major challenge,
especially for B1-level EFL learners. Students frequently experience difficulties with fluency,
accuracy and confidence in real-time conversations (Khlaisang & Sukavatee, 2023). Many
learners have limited access to authentic English-speaking opportunities outside the classroom
and insufficient chances for a real-time oral practice. This lack of exposure can create speaking
anxiety, becoming a significant barrier to progress and discouraging students from actively
participating in communicative activities (Mora et al., 2023).
Consequently, there is a need to evaluate and explore digital tools that can effectively
support learners in improving their oral proficiency and foster self-confidence in spoken English.
While digital tools have been developed to address these persistent challenges, the effectiveness
of the Talkpal platform has not been extensively examined in the context of speaking skills
development, particularly for B1-level learners. Therefore, further research is required to
determine whether this innovative tool can significantly improve speaking skills among B1
students to help them achieve higher English proficiency.
1.1 Goal of the study
The goal of the present study is to evaluate the effectiveness of integrating the digital tool
Talkpal in enhancing speaking skills of B1-EFL level learners, with the aim of helping them
achieve a higher level of English proficiency.
INTRODUCTION
In recent years, the educational landscape has experienced a significant transformation,
driven largely by the integration of digital tools. These technological resources have transformed
traditional teaching and long-established learning paradigms, offering learners greater autonomy
and opportunities to practice beyond the classroom (Du & Daniel, 2024). The adoption of mobile
devices and remarkable advancements in the field of natural language processing (NLP) and
artificial intelligence (AI) have accelerated the use of such tools across diverse educational
contexts globally, making learning more accessible and engaging for students (Chen, 2023).
Educators worldwide are continuously seeking innovative methods to support
communicative competence in English as a Foreign Language (EFL), especially for intermediate
learners aiming to reach advanced levels of proficiency. As part of this movement, AI-powered
tools such as Talkpal have gained significant attention for their potential to enhance speaking
skills through interactive practice, personalized feedback, and adaptive learning pathways (Pozo-
Sánchez et al., 2021). The capacity of these tools to provide an instant error correction, specialized
exercises, and activities focused on individual performance has made a remarkable difference
compared to traditional methods, fostering a more effective learning experience (Lee, S. M.
2019).
However, despite all these technological advancements, speaking remains a major challenge,
especially for B1-level EFL learners. Students frequently experience difficulties with fluency,
accuracy and confidence in real-time conversations (Khlaisang & Sukavatee, 2023). Many
learners have limited access to authentic English-speaking opportunities outside the classroom
and insufficient chances for a real-time oral practice. This lack of exposure can create speaking
anxiety, becoming a significant barrier to progress and discouraging students from actively
participating in communicative activities (Mora et al., 2023).
Consequently, there is a need to evaluate and explore digital tools that can effectively
support learners in improving their oral proficiency and foster self-confidence in spoken English.
While digital tools have been developed to address these persistent challenges, the effectiveness
of the Talkpal platform has not been extensively examined in the context of speaking skills
development, particularly for B1-level learners. Therefore, further research is required to
determine whether this innovative tool can significantly improve speaking skills among B1
students to help them achieve higher English proficiency.
1.1 Goal of the study
The goal of the present study is to evaluate the effectiveness of integrating the digital tool
Talkpal in enhancing speaking skills of B1-EFL level learners, with the aim of helping them
achieve a higher level of English proficiency.
Vol. 13/ Núm. 1 2026 pág. 1122
MATERIALS AND METHODS
Research design
According to Pilcher and Cortazzi (2024), qualitative and quantitative research are two
complementary approaches, which is why many researchers combine them by integrating the
numerical precision of quantitative analysis with the contextual depth of qualitative inquiry. This
combination allows research to transcend methodological boundaries and foster a more holistic
and flexible understanding of social phenomena. Regarding the methodology used in this study ,
a mixed-methods approach was employed, integrating both quantitative and qualitative methods.
This approach allowed for a more thorough investigation of the results, within a quasi-
experimental design that included a pre-test and a post-test administered to a single group over a
period of four weeks. During this period, students carried out speaking practice activities using
Talkpal, both in class and independently at home.
The quantitative component included a pre- and post-test along with a speaking rubric,
as well as structured questionnaires administered at the beginning and end of the intervention to
obtain numerical data on students' progress and perceptions regarding the impact of the Talkpal
digital tool on oral production. The qualitative component was based on classroom observations
and exit tickets completed at the end of each class session, which provided descriptive information
on students' participation, motivation, and experiences while using Talkpal. This combination of
methods allowed for data triangulation, strengthening the validity and depth of the study.
Population and sampling
This study was conducted at a private high school in Ecuador's coastal region. The study
population consisted of 120 second-year high school students who had English integrated into
their academic schedule as a mandatory component of the curriculum. According to the Common
European Framework of Reference for Languages (CEFR), these students were at a B1
proficiency level, corresponding to an intermediate level of language competence.
The sample selected for this study consisted of 22 second-year science students who
participated in the pedagogical intervention using the digital tool Talkpal. As Turner (2020) points
out, “Sampling is the selection of a subset of the population of interest in a research study. In the
vast majority of research endeavors, the participation of an entire population of interest is not
possible, so a smaller group is relied upon for data collection” (p. 9). This means that in most
cases, the participation of the entire population is not possible, therefore, a smaller representative
group is selected for data collection.
Participants in this study were selected through purposive convenience sampling, as those
students who met specific criteria at level B1 proficiency confirmed through prior diagnostic test,
regular attendance, and voluntary participation were selected to take part in all project activities.
Purposive convenience sampling is widely used in mixed-methods research as it allows for the
MATERIALS AND METHODS
Research design
According to Pilcher and Cortazzi (2024), qualitative and quantitative research are two
complementary approaches, which is why many researchers combine them by integrating the
numerical precision of quantitative analysis with the contextual depth of qualitative inquiry. This
combination allows research to transcend methodological boundaries and foster a more holistic
and flexible understanding of social phenomena. Regarding the methodology used in this study ,
a mixed-methods approach was employed, integrating both quantitative and qualitative methods.
This approach allowed for a more thorough investigation of the results, within a quasi-
experimental design that included a pre-test and a post-test administered to a single group over a
period of four weeks. During this period, students carried out speaking practice activities using
Talkpal, both in class and independently at home.
The quantitative component included a pre- and post-test along with a speaking rubric,
as well as structured questionnaires administered at the beginning and end of the intervention to
obtain numerical data on students' progress and perceptions regarding the impact of the Talkpal
digital tool on oral production. The qualitative component was based on classroom observations
and exit tickets completed at the end of each class session, which provided descriptive information
on students' participation, motivation, and experiences while using Talkpal. This combination of
methods allowed for data triangulation, strengthening the validity and depth of the study.
Population and sampling
This study was conducted at a private high school in Ecuador's coastal region. The study
population consisted of 120 second-year high school students who had English integrated into
their academic schedule as a mandatory component of the curriculum. According to the Common
European Framework of Reference for Languages (CEFR), these students were at a B1
proficiency level, corresponding to an intermediate level of language competence.
The sample selected for this study consisted of 22 second-year science students who
participated in the pedagogical intervention using the digital tool Talkpal. As Turner (2020) points
out, “Sampling is the selection of a subset of the population of interest in a research study. In the
vast majority of research endeavors, the participation of an entire population of interest is not
possible, so a smaller group is relied upon for data collection” (p. 9). This means that in most
cases, the participation of the entire population is not possible, therefore, a smaller representative
group is selected for data collection.
Participants in this study were selected through purposive convenience sampling, as those
students who met specific criteria at level B1 proficiency confirmed through prior diagnostic test,
regular attendance, and voluntary participation were selected to take part in all project activities.
Purposive convenience sampling is widely used in mixed-methods research as it allows for the
Vol. 13/ Núm. 1 2026 pág. 1123
selection of participants who best represent the conditions of the phenomenon under study
(Taherdoost, 2020).
Table 1
Sample population
Population Participants Percentage
Women 12 54.55%
Men 10 45.45%
Total 22 100%
Source: Authors’s elaboration, 2025
Data Collection Techniques and Instruments
In this study, the independent variable was the implementation of Talkpal, an artificial
intelligence-based digital tool designed for oral expression practice. The dependent variable was
the English-speaking ability of B1-level students. In order to obtain data that would allow a
comprehensive assessment of the relationship between both variables, four main instruments were
carefully selected to collect quantitative and qualitative data.
These instruments were adapted considering several key aspects, such as the students'
age, level of English proficiency, and the technological resources available to them, with the aim
of ensuring the validity of the data obtained during the four-weeks intervention. The combination
of these techniques provided objective evidence of oral performance, as well as descriptive
information on students' motivation, participation, and perceptions throughout the learning
process.
Oral tests (pre- and post-test): Students took an oral test before and after the intervention,
lasting approximately 5 minutes, which included answering guided questions. The conversations
were recorded and evaluated using an analytical rubric aligned with five criteria: pronunciation,
fluency, vocabulary, grammar, and confidence. Each criterion was scored on a scale of 1 to 2,
with a maximum score of 10.
Questionnaires (initial and final): Closed-ended yes/no questionnaires were administered
to obtain accurate and easily quantifiable information about students' perceptions and experiences.
The initial questionnaire examined students' self-perceptions of their speaking skills and
familiarity with AI-based digital tools. The final questionnaire included additional items on
usefulness, motivation, and satisfaction with using Talkpal.
Observation: During classroom sessions, the authors of this research conducted
observations to analyze students’ participation, interaction, and attitude towards the use of the
Talkpal digital tool to improve speaking skills. This technique complemented the quantitative
data with qualitative insights into students' behavior during the intervention.
selection of participants who best represent the conditions of the phenomenon under study
(Taherdoost, 2020).
Table 1
Sample population
Population Participants Percentage
Women 12 54.55%
Men 10 45.45%
Total 22 100%
Source: Authors’s elaboration, 2025
Data Collection Techniques and Instruments
In this study, the independent variable was the implementation of Talkpal, an artificial
intelligence-based digital tool designed for oral expression practice. The dependent variable was
the English-speaking ability of B1-level students. In order to obtain data that would allow a
comprehensive assessment of the relationship between both variables, four main instruments were
carefully selected to collect quantitative and qualitative data.
These instruments were adapted considering several key aspects, such as the students'
age, level of English proficiency, and the technological resources available to them, with the aim
of ensuring the validity of the data obtained during the four-weeks intervention. The combination
of these techniques provided objective evidence of oral performance, as well as descriptive
information on students' motivation, participation, and perceptions throughout the learning
process.
Oral tests (pre- and post-test): Students took an oral test before and after the intervention,
lasting approximately 5 minutes, which included answering guided questions. The conversations
were recorded and evaluated using an analytical rubric aligned with five criteria: pronunciation,
fluency, vocabulary, grammar, and confidence. Each criterion was scored on a scale of 1 to 2,
with a maximum score of 10.
Questionnaires (initial and final): Closed-ended yes/no questionnaires were administered
to obtain accurate and easily quantifiable information about students' perceptions and experiences.
The initial questionnaire examined students' self-perceptions of their speaking skills and
familiarity with AI-based digital tools. The final questionnaire included additional items on
usefulness, motivation, and satisfaction with using Talkpal.
Observation: During classroom sessions, the authors of this research conducted
observations to analyze students’ participation, interaction, and attitude towards the use of the
Talkpal digital tool to improve speaking skills. This technique complemented the quantitative
data with qualitative insights into students' behavior during the intervention.
Vol. 13/ Núm. 1 2026 pág. 1124
Exit tickets: At the end of each weekly Talkpal session, students answered brief reflective
questions such as: “One thing I learned today was…”, “Something I found difficult…”, and “A
suggestion for the next session…”. These responses were analyzed as a qualitative source of
students’ immediate perceptions and reactions to the activities.
Data Analysis Procedures
The data collected in this research were analyzed using both quantitative and qualitative
procedures. Quantitative analysis was applied primarily to the results of the oral tests (pre-test
and post-test) and the closed-ended (yes/no) questionnaires. Qualitative analysis was used to
examine information from classroom observations and the exit tickets at the end of each session.
In the quantitative analysis, the scores obtained from the pre-test and post-test were
organized into comparative tables to observe changes in students' oral performance after the
intervention. Questionnaire data were processed using frequency and percentage analysis to
identify students' perceptions of the digital tool. In the qualitative analysis, the observation notes
and open-ended responses from the exit tickets were examined thematically. This analysis
provided complementary information to the quantitative results, offering a deeper understanding
of the tool's impact on the learning experience.
Ethical Consideration
This research project was conducted in strict accordance with ethical standards to
safeguard the rights, dignity, and well-being of all participants. Prior informed consent was
obtained from institutional authorities, students, and their legal guardians. Participants were
provided with a clear and complete explanation of the objectives, procedures, potential risks, and
benefits of the research.
Participation was entirely voluntary, and all participants were guaranteed the right to
withdraw from the study at any time without academic or personal consequences. All data
collected were securely stored and accessed only by the principal investigator. All students in the
selected group had equal access to the digital tool and support throughout the intervention. This
study also complied with institutional guidelines and was supervised and approved by academic
advisors, ensuring compliance with all ethical standards in the design and implementation of the
research.
Exit tickets: At the end of each weekly Talkpal session, students answered brief reflective
questions such as: “One thing I learned today was…”, “Something I found difficult…”, and “A
suggestion for the next session…”. These responses were analyzed as a qualitative source of
students’ immediate perceptions and reactions to the activities.
Data Analysis Procedures
The data collected in this research were analyzed using both quantitative and qualitative
procedures. Quantitative analysis was applied primarily to the results of the oral tests (pre-test
and post-test) and the closed-ended (yes/no) questionnaires. Qualitative analysis was used to
examine information from classroom observations and the exit tickets at the end of each session.
In the quantitative analysis, the scores obtained from the pre-test and post-test were
organized into comparative tables to observe changes in students' oral performance after the
intervention. Questionnaire data were processed using frequency and percentage analysis to
identify students' perceptions of the digital tool. In the qualitative analysis, the observation notes
and open-ended responses from the exit tickets were examined thematically. This analysis
provided complementary information to the quantitative results, offering a deeper understanding
of the tool's impact on the learning experience.
Ethical Consideration
This research project was conducted in strict accordance with ethical standards to
safeguard the rights, dignity, and well-being of all participants. Prior informed consent was
obtained from institutional authorities, students, and their legal guardians. Participants were
provided with a clear and complete explanation of the objectives, procedures, potential risks, and
benefits of the research.
Participation was entirely voluntary, and all participants were guaranteed the right to
withdraw from the study at any time without academic or personal consequences. All data
collected were securely stored and accessed only by the principal investigator. All students in the
selected group had equal access to the digital tool and support throughout the intervention. This
study also complied with institutional guidelines and was supervised and approved by academic
advisors, ensuring compliance with all ethical standards in the design and implementation of the
research.
Vol. 13/ Núm. 1 2026 pág. 1125
RESULTS AND DISCUSSION
Figure 1
Yes-No pre-survey questions
Source: Author’s elaboration, 2025
During the diagnostic phase, a pre-test with a yes-no-question survey was administered to
the group of students. The study analyzed self-reported speaking skills, confidence, and anxiety
among Second-year Baccalaureate Science students in the Jean Piaget High School setting using
a pre- and post-intervention survey framework. The results indicate a significant positive shift
across all measured B1-level communicative skills, with several changes observed in both the
affective domain and fluency.
The data from the bilingual cohort indicate that students began the intervention with
relatively strong baseline skills in vocabulary and coherence, as evidenced by over 68%
agreement in the pre-intervention results. The most substantial gain occurred in fluency, with a
45.4% increase in students reporting the ability to speak for one minute without long pauses. This
improvement suggests that the use of Talkpal effectively facilitated the transition from linguistic
knowledge to spontaneous language use, a key challenge at the B1 level.
The data indicate a significant reduction in psychological barriers to speaking. Student
confidence increased substantially, and the proportion of students unaffected by fear of mistakes
more than doubled. These results demonstrate that the tool fostered a low-stakes environment, a
factor essential for effective language practice (Villafuerte-Segura & Riera-Puga, 2024).
8
15 16
10
7
10
14
6 5
12
15
12
I can speak for
about a minute on
a familiar topic
without long
pauses.
I know enough
words to describe
events,
experiences, and
future plans.
I can use basic
linking words
(but, because, so)
to connect my
sentences clearly.
My pronunciation
is clear enough
that people rarely
ask me to repeat
things.
I feel confident
speaking English
in front of the
whole class.
The fear of
making a mistake
stops me from
speaking English.
yes no
RESULTS AND DISCUSSION
Figure 1
Yes-No pre-survey questions
Source: Author’s elaboration, 2025
During the diagnostic phase, a pre-test with a yes-no-question survey was administered to
the group of students. The study analyzed self-reported speaking skills, confidence, and anxiety
among Second-year Baccalaureate Science students in the Jean Piaget High School setting using
a pre- and post-intervention survey framework. The results indicate a significant positive shift
across all measured B1-level communicative skills, with several changes observed in both the
affective domain and fluency.
The data from the bilingual cohort indicate that students began the intervention with
relatively strong baseline skills in vocabulary and coherence, as evidenced by over 68%
agreement in the pre-intervention results. The most substantial gain occurred in fluency, with a
45.4% increase in students reporting the ability to speak for one minute without long pauses. This
improvement suggests that the use of Talkpal effectively facilitated the transition from linguistic
knowledge to spontaneous language use, a key challenge at the B1 level.
The data indicate a significant reduction in psychological barriers to speaking. Student
confidence increased substantially, and the proportion of students unaffected by fear of mistakes
more than doubled. These results demonstrate that the tool fostered a low-stakes environment, a
factor essential for effective language practice (Villafuerte-Segura & Riera-Puga, 2024).
8
15 16
10
7
10
14
6 5
12
15
12
I can speak for
about a minute on
a familiar topic
without long
pauses.
I know enough
words to describe
events,
experiences, and
future plans.
I can use basic
linking words
(but, because, so)
to connect my
sentences clearly.
My pronunciation
is clear enough
that people rarely
ask me to repeat
things.
I feel confident
speaking English
in front of the
whole class.
The fear of
making a mistake
stops me from
speaking English.
yes no
Vol. 13/ Núm. 1 2026 pág. 1126
Figure 2
Behavior of the analytic rubric criteria before and after the intervention
Source: Author’s elaboration, 2025
The implementation of the Talkpal intervention resulted in positive outcomes in both
objective linguistic assessments and subjective self-reports. Students demonstrated an average
increase of 0.91 points on the 5-point speaking rubric, improving from a mean of 3.10 to 4.01.
The greatest improvement was observed in fluency, with a gain of 1.10 points, suggesting that
regular practice using Talkpal helped students overcome hesitation and promote spontaneous
speech.
The intervention led to a 50.0% increase in self-reported confidence, with rates rising
from 36.4% to 86.4%. This affective improvement is supported by the students' reported comfort
level with the tool, as 95.5% indicated feeling more comfortable practicing with Talkpal than with
their peers. This outcome addresses the anxiety barrier common in bilingual learning
environments and aligns with previous research on the benefits of low-stakes practice in reducing
speaking anxiety (Krashen, 1985; Lee & Chen, 2023).
Fluency Pronunciation Vocabulary Coherence
Pre-test 2,95 3,27 3,36 2,82
Post-test 4,05 4,18 4,14 3,68
PRE-POST TEST
Pre-test Post-test
Figure 2
Behavior of the analytic rubric criteria before and after the intervention
Source: Author’s elaboration, 2025
The implementation of the Talkpal intervention resulted in positive outcomes in both
objective linguistic assessments and subjective self-reports. Students demonstrated an average
increase of 0.91 points on the 5-point speaking rubric, improving from a mean of 3.10 to 4.01.
The greatest improvement was observed in fluency, with a gain of 1.10 points, suggesting that
regular practice using Talkpal helped students overcome hesitation and promote spontaneous
speech.
The intervention led to a 50.0% increase in self-reported confidence, with rates rising
from 36.4% to 86.4%. This affective improvement is supported by the students' reported comfort
level with the tool, as 95.5% indicated feeling more comfortable practicing with Talkpal than with
their peers. This outcome addresses the anxiety barrier common in bilingual learning
environments and aligns with previous research on the benefits of low-stakes practice in reducing
speaking anxiety (Krashen, 1985; Lee & Chen, 2023).
Fluency Pronunciation Vocabulary Coherence
Pre-test 2,95 3,27 3,36 2,82
Post-test 4,05 4,18 4,14 3,68
PRE-POST TEST
Pre-test Post-test
Vol. 13/ Núm. 1 2026 pág. 1127
Figure 3
Post survey results
Source: Author’s elaboration, 2025
A majority of students reported that the intervention was highly beneficial, particularly
in terms of comfort during practice. Notably, 95.5% of participants indicated a greater comfort
practicing with the tool compared to practice with their peers. This finding suggests that the AI-
driven, non-judgmental practice environment effectively reduced the fear of negative evaluation,
a factor commonly associated with classroom performance.
The primary objective of this study, to assess the effectiveness of the Talkpal tool in
improving B1-level speaking skills and confidence, is strongly supported by the data. Notable
improvements in fluency, including a 1.10 mean gain and a 50% increase in confidence, highlight
the tool's main benefit: promoting the automatization of language use at the B1 level. Fluency
development depends on repetitive, low-stakes practice, which the tool consistently provides.
Talkpal, as a digital and non-judgmental AI partner, addressed two primary barriers to
communicative competence: insufficient dedicated practice time and the prevalent fear of public
error. By providing a psychologically safe learning environment, the tool encouraged students to
take greater linguistic risks, such as using new vocabulary and more complex structures, and
significantly reduced speaking anxiety. Enhanced self-efficacy is essential for successful
language learning (Gardner, 2001) and aligns with recent research on the impact of large language
models on language production (Chomsky & Smith, 2024). The high comfort level, with 95.5%
of students preferring Talkpal over to peer practice, contributed directly to observed
improvements in overall language skill acquisition.
18
20 21
16
14
16
4
2 1
5
8
6
0
5
10
15
20
25
Now, I can
speak for about
a minute on a
familiar topic
without long
pauses.
Now, I know
enough words
to describe
events,
experiences,
and future
plans.
Now, I can use
basic linking
words (but,
because, so) to
connect my
sentences
clearly.
Now, my
pronunciation is
clear enough
that people
rarely ask me to
repeat things.
Now, I feel
confident
speaking
English in front
of the whole
class.
Using Talkpal
helped reduce
my nervousness
about speaking
English.
POST SURVEY
Yes
No
Figure 3
Post survey results
Source: Author’s elaboration, 2025
A majority of students reported that the intervention was highly beneficial, particularly
in terms of comfort during practice. Notably, 95.5% of participants indicated a greater comfort
practicing with the tool compared to practice with their peers. This finding suggests that the AI-
driven, non-judgmental practice environment effectively reduced the fear of negative evaluation,
a factor commonly associated with classroom performance.
The primary objective of this study, to assess the effectiveness of the Talkpal tool in
improving B1-level speaking skills and confidence, is strongly supported by the data. Notable
improvements in fluency, including a 1.10 mean gain and a 50% increase in confidence, highlight
the tool's main benefit: promoting the automatization of language use at the B1 level. Fluency
development depends on repetitive, low-stakes practice, which the tool consistently provides.
Talkpal, as a digital and non-judgmental AI partner, addressed two primary barriers to
communicative competence: insufficient dedicated practice time and the prevalent fear of public
error. By providing a psychologically safe learning environment, the tool encouraged students to
take greater linguistic risks, such as using new vocabulary and more complex structures, and
significantly reduced speaking anxiety. Enhanced self-efficacy is essential for successful
language learning (Gardner, 2001) and aligns with recent research on the impact of large language
models on language production (Chomsky & Smith, 2024). The high comfort level, with 95.5%
of students preferring Talkpal over to peer practice, contributed directly to observed
improvements in overall language skill acquisition.
18
20 21
16
14
16
4
2 1
5
8
6
0
5
10
15
20
25
Now, I can
speak for about
a minute on a
familiar topic
without long
pauses.
Now, I know
enough words
to describe
events,
experiences,
and future
plans.
Now, I can use
basic linking
words (but,
because, so) to
connect my
sentences
clearly.
Now, my
pronunciation is
clear enough
that people
rarely ask me to
repeat things.
Now, I feel
confident
speaking
English in front
of the whole
class.
Using Talkpal
helped reduce
my nervousness
about speaking
English.
POST SURVEY
Yes
No
Vol. 13/ Núm. 1 2026 pág. 1128
CONCLUSION
The purpose of this study was to analyze the impact of Talkpal, a digital tool based on
artificial intelligence, on the development of English-speaking skills in B1-level students. The
results demonstrated that the use of this digital tool had a positive effect on the development of
speaking skills during the four-weeks of intervention, improving several essential aspects of
conversational engagement. These findings confirm that the integration of technological
resources in the classroom can enhance meaningful learning and promote the development of
communicative competence.
Based on the evidence, the quantitative data revealed significant improvements in key areas
of oral production such as fluency, pronunciation, vocabulary, coherence, grammatical accuracy,
and confidence areas that are often challenging for students when communicating. These
improvements underscore the effectiveness and potential of Talkpal in addressing specific oral
communication difficulties and helping students achieve higher levels of oral proficiency.
Furthermore, qualitative data obtained through observation and exit tickets support these
findings, showing increased participation, confidence, enthusiasm, and willingness to
communicate in English both inside and outside the classroom. This tool provides a safe,
supportive, and non-judgmental environment, free from immediate corrections, which fostered a
positive learning attitude.
The interactive and flexible nature of the platform encourages students to take control of
their language learning process, practice at their own pace, and focus on the areas where they
most need improvement. This student-centered approach aligns with contemporary educational
principles that emphasize personalized learning and self-regulation, both essential for consistent
progress in oral proficiency.
In summary, integrating Talkpal as a complementary resource to traditional teaching
methods offers a dynamic and effective strategy for improving oral skills in EFL contexts. The
combination of technological innovation, personalized feedback, and motivational support makes
Talkpal a valuable tool for language classrooms seeking to strengthen speaking skills. This study
recommends further research on digital tools like Talkpal to continue advancing teaching
practices and to better address the needs of a diverse student population.
CONCLUSION
The purpose of this study was to analyze the impact of Talkpal, a digital tool based on
artificial intelligence, on the development of English-speaking skills in B1-level students. The
results demonstrated that the use of this digital tool had a positive effect on the development of
speaking skills during the four-weeks of intervention, improving several essential aspects of
conversational engagement. These findings confirm that the integration of technological
resources in the classroom can enhance meaningful learning and promote the development of
communicative competence.
Based on the evidence, the quantitative data revealed significant improvements in key areas
of oral production such as fluency, pronunciation, vocabulary, coherence, grammatical accuracy,
and confidence areas that are often challenging for students when communicating. These
improvements underscore the effectiveness and potential of Talkpal in addressing specific oral
communication difficulties and helping students achieve higher levels of oral proficiency.
Furthermore, qualitative data obtained through observation and exit tickets support these
findings, showing increased participation, confidence, enthusiasm, and willingness to
communicate in English both inside and outside the classroom. This tool provides a safe,
supportive, and non-judgmental environment, free from immediate corrections, which fostered a
positive learning attitude.
The interactive and flexible nature of the platform encourages students to take control of
their language learning process, practice at their own pace, and focus on the areas where they
most need improvement. This student-centered approach aligns with contemporary educational
principles that emphasize personalized learning and self-regulation, both essential for consistent
progress in oral proficiency.
In summary, integrating Talkpal as a complementary resource to traditional teaching
methods offers a dynamic and effective strategy for improving oral skills in EFL contexts. The
combination of technological innovation, personalized feedback, and motivational support makes
Talkpal a valuable tool for language classrooms seeking to strengthen speaking skills. This study
recommends further research on digital tools like Talkpal to continue advancing teaching
practices and to better address the needs of a diverse student population.
Vol. 13/ Núm. 1 2026 pág. 1129
REFERENCES
Chen, B., Wang, Y., & Wang, L. (2022). The Effects of Virtual Reality-Assisted Language
Learning: A Meta-Analysis. Sustainability, 14(6), 3147.
https://doi.org/10.3390/su14063147
Chen, X. (2023). A Systematic Review of Foreign Language Anxiety. Journal of Education,
Humanities and Social Sciences, 22(3), 90-97. https://doi.org/10.54097/ehss.v22i.12292
Chomsky, A., & Smith, J. D. (2024). Leveraging generative AI for enhanced speaking
proficiency: A comparative study in bilingual high schools. Journal of Computer-Assisted
Language Learning, 12(4), 45-62.
Derakhshan, A., & Shams, M. (2020). The Impact of Speaking Anxiety on EFL Learners’ Oral
Performance. Journal of Language Teaching and Research, 11(5), 723-729.
https://doi.org/10.17507/jltr.1105.01
Dikaprio, V., & Diem, C. D. (2024). How effective is Talkpal.ai in enhancing English
proficiency? Insights from an experimental study. Language, Technology, and Social
Media, 2(1), 48–59. https://doi.org/10.70211/ltsm.v2i1.48
Du, T. M., Nguyen, T. N., & Le, N. A. (2024). Improving First-Year English-Majored Students'
Speaking Skills through Using Digital Storytelling. International Journal of Language
Instruction, 3(2), 29-44. https://doi.org/10.54855/ijli.24323
Gardner, R. C. (2001). Integrative Motivation: Implications for the Second Language Classroom.
TESL Canada Journal, 19(1), 1-19.
Godwin-Jones, R. (2018). Digital language learning and teaching: Re-evaluating the basics.
Language Learning & Technology, 22(1), 5–17.
Goh, C. C. M., & Burns, A. (2012). Teaching Speaking: A Holistic Approach. Cambridge
University Press. https://doi.org/10.1017/CBO9781139151153
Hidayatullah, E. (2024). The impact of Talkpal.AI on English speaking proficiency: An academic
inquiry. Journal of Insan Mulia Education, 2(1), 19–25.
https://doi.org/10.59923/joinme.v2i1.98
Hockly, N. (2018). The role of technology in English language learning and teaching. English
Language Teaching Journal, 72(1), 22–33.
Khlaisang, J., & Sukavatee, P. (2023). Mobile-Assisted Language Learning to Support English
Language Communication among Higher Education Learners in Thailand. Electronic
Journal of e-Learning, 21(3), 2974. https://doi.org/10.34190/ejel.21.3.2974
Krashen, S. D. (1985). The Input Hypothesis: Issues and Implications. Longman.
https://doi.org/10.2307/414800
Lee, A., & Chen, B. (2023). The role of artificial intelligence in mitigating second language
speaking anxiety. Journal of Applied Linguistics, 45(2), 112–129.
REFERENCES
Chen, B., Wang, Y., & Wang, L. (2022). The Effects of Virtual Reality-Assisted Language
Learning: A Meta-Analysis. Sustainability, 14(6), 3147.
https://doi.org/10.3390/su14063147
Chen, X. (2023). A Systematic Review of Foreign Language Anxiety. Journal of Education,
Humanities and Social Sciences, 22(3), 90-97. https://doi.org/10.54097/ehss.v22i.12292
Chomsky, A., & Smith, J. D. (2024). Leveraging generative AI for enhanced speaking
proficiency: A comparative study in bilingual high schools. Journal of Computer-Assisted
Language Learning, 12(4), 45-62.
Derakhshan, A., & Shams, M. (2020). The Impact of Speaking Anxiety on EFL Learners’ Oral
Performance. Journal of Language Teaching and Research, 11(5), 723-729.
https://doi.org/10.17507/jltr.1105.01
Dikaprio, V., & Diem, C. D. (2024). How effective is Talkpal.ai in enhancing English
proficiency? Insights from an experimental study. Language, Technology, and Social
Media, 2(1), 48–59. https://doi.org/10.70211/ltsm.v2i1.48
Du, T. M., Nguyen, T. N., & Le, N. A. (2024). Improving First-Year English-Majored Students'
Speaking Skills through Using Digital Storytelling. International Journal of Language
Instruction, 3(2), 29-44. https://doi.org/10.54855/ijli.24323
Gardner, R. C. (2001). Integrative Motivation: Implications for the Second Language Classroom.
TESL Canada Journal, 19(1), 1-19.
Godwin-Jones, R. (2018). Digital language learning and teaching: Re-evaluating the basics.
Language Learning & Technology, 22(1), 5–17.
Goh, C. C. M., & Burns, A. (2012). Teaching Speaking: A Holistic Approach. Cambridge
University Press. https://doi.org/10.1017/CBO9781139151153
Hidayatullah, E. (2024). The impact of Talkpal.AI on English speaking proficiency: An academic
inquiry. Journal of Insan Mulia Education, 2(1), 19–25.
https://doi.org/10.59923/joinme.v2i1.98
Hockly, N. (2018). The role of technology in English language learning and teaching. English
Language Teaching Journal, 72(1), 22–33.
Khlaisang, J., & Sukavatee, P. (2023). Mobile-Assisted Language Learning to Support English
Language Communication among Higher Education Learners in Thailand. Electronic
Journal of e-Learning, 21(3), 2974. https://doi.org/10.34190/ejel.21.3.2974
Krashen, S. D. (1985). The Input Hypothesis: Issues and Implications. Longman.
https://doi.org/10.2307/414800
Lee, A., & Chen, B. (2023). The role of artificial intelligence in mitigating second language
speaking anxiety. Journal of Applied Linguistics, 45(2), 112–129.
Vol. 13/ Núm. 1 2026 pág. 1130
Pilcher, N., & Cortazzi, M. (2024). ‘Qualitative’ and ‘quantitative’ methods and approaches
across subject fields: Implications for research values, assumptions, and practices. Quality
& Quantity, 58(3), 2357–2387. https://doi.org/10.1007/s11135-023-01734-4
Saito, K. (2017). Second Language Pronunciation Learning and Teaching: Theory, Research, and
Practice. Routledge. https://doi.org/10.4324/9781315649429 put them in alfabetical order
Taherdoost, H. (2020). Sampling methods in research methodology: How to choose a sampling
technique for research. SAGE Open, 10(3), 2158244020977000.
https://doi.org/10.1177/2158244020977000
Turner, D. P. (2020). Sampling methods in research design. Headache, 60(1), 8–12.
https://doi.org/10.1111/head.13707
Villafuerte-Segura, L. M., & Riera-Puga, R. F. (2024). The Use of an AI-Based Application to
Enhance Speaking Skills and Reduce Affective Barriers in Second Baccalaureate Students.
International Journal of Modern Educational Studies, 6(2), 652–672.
https://doi.org/10.37811/cl_rcm.v8i3.11592
Pilcher, N., & Cortazzi, M. (2024). ‘Qualitative’ and ‘quantitative’ methods and approaches
across subject fields: Implications for research values, assumptions, and practices. Quality
& Quantity, 58(3), 2357–2387. https://doi.org/10.1007/s11135-023-01734-4
Saito, K. (2017). Second Language Pronunciation Learning and Teaching: Theory, Research, and
Practice. Routledge. https://doi.org/10.4324/9781315649429 put them in alfabetical order
Taherdoost, H. (2020). Sampling methods in research methodology: How to choose a sampling
technique for research. SAGE Open, 10(3), 2158244020977000.
https://doi.org/10.1177/2158244020977000
Turner, D. P. (2020). Sampling methods in research design. Headache, 60(1), 8–12.
https://doi.org/10.1111/head.13707
Villafuerte-Segura, L. M., & Riera-Puga, R. F. (2024). The Use of an AI-Based Application to
Enhance Speaking Skills and Reduce Affective Barriers in Second Baccalaureate Students.
International Journal of Modern Educational Studies, 6(2), 652–672.
https://doi.org/10.37811/cl_rcm.v8i3.11592