Customers’ acceptance for use of Chatbot: Case study at Vietnam prospectus joint stock commercial bank Hue branch

Chia sẻ: _ _ | Ngày: | Loại File: PDF | Số trang:17

Thêm vào BST

Báo xấu

5
lượt xem 2
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

Nghiên cứu nhằm mục đích khám phá sự chấp nhận sử dụng ứng dụng chatbot của khách hàng trong lĩnh vực ngân hàng. Tác giả khảo sát 125 khách hàng đã trải nghiệm sử dụng chatbot của ngân hàng Việt Nam Thịnh vượng chi nhánh Huế bằng phương pháp chọn mẫu thuận tiện. Thống kê mô tả, phân tích nhân tố và hồi quy đa biến được sử dụng để phân tích dữ liệu. Mời các bạn cùng tham khảo!

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Customers’ acceptance for use of Chatbot: Case study at Vietnam prospectus joint stock commercial bank Hue branch

CUSTOMERS’ ACCEPTANCE FOR USE OF CHATBOT: CASE STUDY AT VIETNAM PROSPECTUS JOINT STOCK COMMERCIAL BANK HUE BRANCH Nguyen Thi Thuy Dat, Nguyen Van Phat*, Le Thi Thuy, Ngo Viet Minh Quang University of Economics, Hue University Email: nvphat@hce.edu.vn Abstract: This study aims to investigate customers' acceptance of using a chatbot in the banking sector. The author collected opinions of 125 customers who experienced chatbots at Vietnam Prosperity Commercial Joint Stock Bank, Hue branch by the convenient sampling. Descriptive statistics, factor analysis, and multivariate regression were applied to analyze the collected data. The results show that perceived usefulness and attractiveness positively influence the acceptance of chatbot while perception of risk has a negative effect. Based on the results, the authors propose some implications to stimulate customers to use chatbot applications at banks. Keywords: Acceptance, Chatbot, banking SỰ CHẤP NHẬN CỦA KHÁCH HÀNG ĐỐI VỚI CHATBOT: TRƯỜNG HỢP NGHIÊN CỨU TẠI NGÂN HÀNG THƯƠNG MẠI CỔ PHẨN VIỆT NAM THỊNH VƯỢNG CHI NHÁNH HUẾ Tóm tắt: Nghiên cứu nhằm mục đích khám phá sự chấp nhận sử dụng ứng dụng chatbot của khách hàng trong lĩnh vực ngân hàng. Tác giả khảo sát 125 khách hàng đã trải nghiệm sử dụng chatbot của ngân hàng Việt Nam Thịnh vượng chi nhánh Huế bằng phương pháp chọn mẫu thuận tiện. Thống kê mô tả, phân tích nhân tố và hồi quy đa biến được sử dụng để phân tích dữ liệu. Kết quả cho thấy nhận thức sự hữu ích, sự hấp dẫn có tác động tích cực đến sự chấp nhận sử dụng chatbot trong khi đó nhận thức rủi ro có mối quan hệ ngược chiều đối với sự chấp nhận sử dụng. Từ đó, các hàm ý chính sách được đề xuất nhằm thúc đẩy sự chấp nhận sử dụng chatbot trong ngân hàng. Từ khóa: Sự chấp nhận, chatbot, ngân hàng 1. Introduction The development of science and technology has had a remarkable impact on consumer behavior. According to We Are Social (2022), the global population is 7.91 billion people, of which more than 67% have mobile phones, more than 60% connect to the Internet, and 58.4% have social media accounts. In addition, consumers increasingly prefer quick responses (Forbes, 2016; Mero, 2018) and anytime, anywhere (Chung et al., 2018). In addition, customers pay attention to highly personalized messages and service solutions (Taylor, 2020). The chatbot has risen since 1950 (Turning, 1950). Chatbot is defined as a computer program that can mimic human communication using natural language and computer 909
learning programs (Araujo, 2018). ). The first known chatbot was Eliza (Weizenbaum, 1966). Chatbots have had continuous innovation over the years with different technologies, and today chatbots are virtual assistants, such as Siri of Apple) or Alexa of Amazon (Adamopoulou & Moussiades, 2020). Chatbots have brought benefits to businesses as well as users. As for the company, chatbots provide 24/7 service, increase touch points, collect data, personalize data, interact one-one, and reduce service and support costs (Zumstein & Hundertmark, 2018). In addition, customers can use the service at all times, save time and cost in communication, and obtain relevant information (Zumstein & Hundertmark, 2018). Chatbots are seen as "services, powered by rules and sometimes artificial intelligence that you interact with through a chat interface" (Schlicht, 2016). Chatbots are used in a variety of industries, including healthcare, e-commerce, retail, insurance, and customer service.etc. The studies on chatbots are also conducted such as on “the interaction of Chatbots with consumers in the banking sector” (Mogaji et al., 2021), “the level of consumer acceptance of Chatbots for consumers in the retail industry” (Rese et al., 2020) or “customer experience using Chatbots has an impact on brand love” (Trivedi, 2019). In Vietnam, the Bank industry has undergone many changes in the application of technology to optimize the experience and meet customers' needs. Currently, chatbots are developed and used in some banks such as Vietinbank, Vietcombank, and Vietnam Prosperity Bank (VPBank). VPBank, established in 1995, is one of the largest joint stock commercial banks in Vietnam. VPBank owns a large transaction network with 230 transaction points, 61 branches, and 166 transaction offices. To grasp the general development trend of the banking industry, Vietnam Prosperity Joint Stock Commercial Bank is one of the pioneer banks to build a Chatbot system on Fanpage to optimize work and reply quickly to customers' inquiries. It results in the improvement of bank service quality. However, using chatbots is not popular when few customers know and hesitate to use them. Therefore, this article aims to examine customers' acceptance of the Chatbot of VPBank, Hue Branch. The results might help the bank to improve the adoption of chatbot usage among customers at the bank. The paper consists of 5 parts, after the introduction and theoretical overview is the research method section. Part 4 is the Research results and discussion, and finally, the Conclusion and implications. 2. Theoretical overview 2.1. The concept of technology acceptance For a few decades, scientists have had efforts to explain consumers’ behavior through models such as the Theory of Reasonable Action (TRA) (Fishbein & Ajzen, 1975), the Theory of Planned Behavior (TPB) (Ajzen, 1991) or the Technology Acceptance Model (TAM) (Davis, 1986). TRA uses Attitudes and Subjective Norms to predict human action in general (Fishbein & Ajzen, 1975). According to TPB, the Perceived Behavioral Control (PBC) factor is added to the TRA model to form a new model and determine behavior intention and actual behavior adoption. In the TAM model, 910
Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) are the set of beliefs that may impact the attitude and actual usage. Then TAM model was revised by Davis, Bogozzi & Warshaw (1989), Venkatesh & Davis (1996), Venkatesh & Davis (2000), and Venkatesh & Bala (2008). Figure 1: Technology Acceptance Model Source: Davis, Bogozzi & Warshaw (1989) 2.2. Overview of Chatbots 2.2.1. Chatbot definition Chatbot is an "online conversation system between humans and computers in natural language" (Jia, 2003). A chatbot is a computer program capable of responding like an intelligent entity when conversing through text or voice and understanding one or more human languages using "natural language processing" (NLP) (Khanna, 2015). On a technical level, a chatbot is a computer program that simulates a human conversation to resolve customer queries. When a customer or potential customer contacts through any channel, the chatbot will be available to greet them and solve their problem. They can also help customers submit service requests, email, or connect with human agents if needed (Shweta, Kelly, 2022). It is one of the most basic and popular applications of intelligent human-computer interaction (Bansal & Khan, 2018). 2.2.2. Benefits and disadvantages of Chatbots Chatbots can bring advantages to companies and customers. Thanks to chatbots, businesses can supply 24/7 services (Zumstein & Hundertmark, 2018). It is undeniable that chatbots can reply to all requests of consumers anywhere, anytime during the day, on any day in the week, especially on holidays that are day-offs of humans. Hence, it may increase customers’ satisfaction with immediate answers and reduce personnel costs. Furthermore, using chatbots can support companies in conducting the one to one communication with chatters. Moreover, chatbots can collect information in the time of conversations with users. This data source may be useful for personalized offers or marketing strategies to target customers (Zumstein & Hundertmark, 2018). On the other hand, chatbots have some drawbacks and threats. Firstly, private information such as personal or financial data may leak when consumers do contact chatbots (Zumstein & Hundertmark, 2018). It may result in users’ fears of chatbot usage. Therefore, chatbot owners and creators should advance the technology for data protection. 911
The second is the acceptance of users. Chatbot has changed the method of communication between companies and consumers. Instead of approaching information from websites, TV, emails, or apps, customers can use chatbots. However, most of them are not familiar with this new communication channel. Thus, it will be costly and time-consuming for businesses to promote chatbots and get chatbot adoption of users (Zumstein & Hundertmark, 2018). 2.2.3. The fields of the chatbot application Chatbot applications are applied in many different fields and successfully in many industries. In customer care, the chatbot may save time and improve customer experience (IBM, 2017; Accenture, 2018). Chatbots are also widely used in the financial and banking sectors to assist in looking up information about assets, securities, currencies, and paying bills (Forbes, 2017). In the medical field, chatbots are useful to schedule medical appointments and give health advice and information about drugs and diseases (Healthcare IT News, 2017). In education, chatbots can support career counseling and learning orientation for students (EdTech Magazine, 2018). The tourism industry is known for using Chatbots to provide information about landmarks, book flights, book hotels, and advise on travel schedules (Travolution, 2017). Chatbots can also be used in competitions and entertainment, organizing, providing information, and serving the audience in contests, TV games, and entertainment events (Chatbots Magazine, 2017). In addition, Chatbots are also used in the technology field to assist customers in using products, troubleshooting, and updating product information (HubSport, 2018). 2.3. Developing research hypothesis In the TAM model, users accept technology based on perceptions that lead to intention and behavior (Davis, 1989). Three main components in the TAM model include Perceived ease of use, perceived usefulness, and attitude towards use. Therefore, the author uses the factors in the TAM model in combination with the perceived attractiveness factor (Dowling, 1994; Vander Heijden, 2004) in accepting the use of technology. In addition, empirical studies have shown that consumers perform risk and benefit analyses of all factors involved in disclosure situations. And the level of consumer privacy concerns is a prerequisite for risk assessment (Hann et al., 2008; Kehr et al., 2015). From there, the author proposes criteria to measure customer acceptance of Chatbot of Vietnam Prosperity Commercial Joint Stock Bank, including five factors that are (1) Perceived usefulness; (2) Perceived ease of use; (3) Attitude towards use; (4) Perceived attractiveness; (5) Perceived risk. Perceived usefulness Perceived usefulness is the degree to which a technology model is believed to improve job performance. (Davis, 1989). In agreement, Wu and Wang (2005) argue that perceived usefulness is the degree to which users believe that using certain technology will help them achieve their goals more effectively. In the TAM model, perceived usefulness is identified as a factor that directly affects the intention to use technology products including mobile applications. When users find applications useful, they will have a positive attitude 912
and tend to use the application more (Choi et al., 2011). Therefore, the study proposes the following hypothesis: H1: Perceived usefulness has a positive relationship with users' acceptance of using chatbots Perceived ease of use According to the technology acceptance model of Davis et al. (1989), perceived ease of use is a crucial factor in determining consumer acceptance of new technology. Perceived ease of use is the degree to which an individual believes that using a particular system will be effortless (Davis et al., 1989). Using a particular technology does not require too much effort in terms of time and effort. In the banking industry, chatbots can provide customer service and customer account support to improve customer experience and increase consumer acceptance of this technology. The second hypothesis is: H2: Perceived ease of use has a positive relationship with users' acceptance of using chatbots Perceived attractiveness Perceived attractiveness is defined as the user's preference for new technology, new technology features, and perceived value of technology (Venkatesh & Bala, 2008). Consumers will be interested in technology and willing to experience and use it (Li et al., 2012). If chatbots are judged to be pleased, they will be considered attractive and accepted by consumers. Perceived attractiveness can play an important role in determining consumer acceptance of new technology. Therefore, the study proposes the following hypothesis: H3: Perceived attractiveness has a positive relationship with users' acceptance of using chatbots Risk perception Perceived risk is the ability to represent a loss in achieving desired outcomes when using an e-service (Yang et al., 2015). Shin (2010) defines risk perception as the user's understanding of the possible negative effects of technology use and the possibilities that users can minimize or accept those risks. These risks can be identity, security, financial or health threats. Therefore, a risk assessment may determine consumer acceptance of new technology. If the consumer feels safe using the technology, then the acceptability of the technology is high, so the fourth hypothesis is: H4: Perceived risk is negatively related to user acceptance of chatbots Attitude towards use Attitude definition refers to an individual's use of positive or negative feelings about the performance of the target behavior (Davis et al., 1989). Attitudes toward the use of technology are closely related to users' acceptance of that technology. If consumers have a positive attitude towards technology, they are more likely to accept and use it. Conversely, if consumers' attitudes are unfavorable towards the use of technology, they may refuse to use or switch to another technology, so the last hypothesis is as follows: H5: Attitude has a positive relationship with users' acceptance of using Chatbots 913
3. Research Methods The scales used in the customer acceptance of Chatbot measure 6 research concepts are all multivariate scales and use 5-point Likert scale, with 1 being strongly disagree and 5 as strongly agree. The scales are first built and developed from the theoretical basis, inherited from previous studies, and then adjusted via on qualitative research. More specifically, the scale of Perceived usefulness; Perceived Ease of use; Attitude are inherited from Davis et al. (1989), Dobholkar & Bagozzi (2002); the scale of perceived risk adopted from Trivedi's (2019); the scale of the attractiveness variable inherited the research of Dobholkar and Bagozzi (2002). Finally, the acceptability scale is inherited from the research of Richad, Vivensius, Sfenrianto & Emil (2019). The questionnaire sampled 15 customers to ensure that the questions were suitable for the research topic and that the customers understood the same. Quantitative research was then conducted with a questionnaire consisting of 3 parts, Part 1 is for exclusion questions, Part 2 is for the main content, and Part 3 is for demographic factors. Table 1: Concepts and items of scales Concepts and items of scales Authors Perceived usefullness Davis et al. (1989), Using chatbots would improve my performance. Dobholkar & Bagozzi Using chatbots would increase my productivity. (2002) Using chatbots would enhance my effectiveness. I would find chatbots useful. Perceived ease of use Davis et al. (1989), Learning to operate chatbots would be easy for me. Dobholkar & Bagozzi I would find it easy to get to do what I want to do with (2002) chatbots. It would be easy for me to become skillful at using chatbots. I would find easy to use chatbots. Attitude Davis et al. (1989), Receiving information and transaction in chatbots is not a Dobholkar & Bagozzi good idea. (2002) I am willing to use chatbots. I like the idea of using chatbots to facilitate getting information and doing transactions. Using chatbots would be a good idea. Perceived risk Trivedi (2019) I perceived that chatbots service was risky. I perceived that while using chatbots, there was a chance that something could go wrong in the outcome. I perceived that chatbots service outcome and effect were difficult to predict. 914
Perceived attractiveness Dobholkar & Bagozzi I perceived that chatbots were not interesting. (2002) Interacting to chatbots would be entertaining. I perceived that chatbots were not be fun. I perceived that chatbots would be attractive. I perceived that chatbots would be enjoyable Acceptance Richad,Vivensius, I choose to use chatbots for getting information and doing Sfenrianto và Emil (2019) transaction. There is possibility that I will use chatbots. I will not recommend anyone to use chatbots. I expect to always be able to use chatbots. The study used the method of determining sample size according to the formula of Hair et al., (2019) corresponding to n = 115 (n = 5 * 23 (expected number of observed variables)). The number of questionnaires issued was 130, the number of valid collected was 125, and the response rate was 96.15%. Data were processed by SPSS 20.0 software. Descriptive statistics by frequency and percentage, scale testing by Cronbach's Alpha, factor analysis (EFA), and multiple linear regression were used to describe and analyze research data. 4. Research results and discussion 4.1. Descriptive statistics test of the scale 4..11.Respondents’ demographic profile The characteristics of the study sample are shown below: Table 2: Study sample characteristics Criteria Number Percentage (N) (%) Gender Male 82 65.6 Female 43 34.4 Total 125 100.0 Age Under 18 years old 7 5.6 18-25 years old 75 60.0 From 25 to 35 years old 35 29.6 Over 35 years old 6 4.8 Total 125 100.0 Job Student 59 47.2 Officer 47 37.6 Other 19 15.2 Total 125 100.0 915
Income Under 3 million VND 54 43.2 per month From 3-5 million VND 17 13.6 From 5-7 million VND 16 12.8 Over 7 million VND 38 30.4 Total 125 100.0 (Source: Author's data processing) According to Table 1, 65.6% of respondents are male, and rest are female (34.4%). As for age, just under 90% of respondents are young and from 18 to 35 years old. Furthermore, about 50% of customers are students, and 37.5% is the officer. Regarding income, most respondents have income under 3 million VND are the greatest, accounting for 43.2%, followed by those with over 7 million VND (30.4%). Groups of 3-5 million VND and 5-7 million VND are 13.6% and 12.8 respectively. 4.1.2. Scale reliability The study used a scale with the following main components: "Perceived Usefulness - PU" with 5 observed variables, "Perceived Ease of use -PEOE" with 5 observed variables, and "Attitude - AT" with 4 important variables. The survey, “Perceived risk - PR” with 4 observed variables, and “Perceived Attractiveness – PA” with 5 observed variables. Performing the Cronbach's Alpha reliability test for the research components shows that Cronbach's Alpha coefficient of most concepts is greater than 0.6. The lowest is the conceptual scale “Perceived Usefulness” – 0.721 and the highest is the conceptual scale “Perceived risk” – 0.92. Which, the fifth observed variable of the scale "Perceived Usefulness - PU5" and "Perceived Attractiveness - PA5" has a correlation coefficient of less than 0.3, so removing these two observed variables and Cronbach's Alpha coefficient of "Perceived Usefulness” remained the same, “Perceived Attractiveness” increased to 0.758. This shows that the scale of independent variable concepts can be used well and is suitable for performing EFA exploratory factor analysis. The results of testing the reliability of the scale using Cronbach's Alpha coefficient for the acceptance scale (AC) show that Cronbach's Alpha coefficient is equal to 0.774 (>0.6) and the total correlation coefficient is greater than 0.3 so the scale used meets the requirements. 4.2. Customer usage behavior for Chatbot of VPBank 4.2.1. The source of information that customers know about VP Bank's Chatbot Table 3: Sources of information that customers know about VPBank Chatbot Information sources Number (N) Percentage (%) Teller 88 70.4 Friend advice 52 41.6 Bank media 40 32.0 Self-discovery 26 20.8 (Source: Author's data processing) 916
The table 2 shows that most customers know the Chatbot Vietnam Prosperity Bank thanks to the bank staff with 88 people (70.4%). It shows that bank staff does well in introducing customers to use the Chatbot of VPBank. Just over 40% of consumers approach the chatbot after receiving advice from their friends. Moreover, bank media increased customers’ awareness about chatbots (32.0%). Thus, the bank should make efforts to promote chatbots to consumers. Lastly, one-fifth of people do self-discovery about chatbots. 4.2.3. Customer platform using Chatbot of VPBank Table 4: Customer platform using Chatbot of VPBank Communication Number (N) Percentage (%) Facebook 109 87.2 VP Bank NEO App 112 89.6 (Source: Author's data processing) The survey result shows that about 90% of customers accessed chatbots of VPBank on VPBank NEO app, accounting for 87.2%, and on Facebook (89.6%). 4.2.3. Purpose of customers when using Chatbot of VPBank Table 5: Customer's purpose when using VPBank Chatbot Criteria Number (N) Percentage (%) Register to open more cards 72 57.6 View information about savings interest rates on loans 64 51.2 Seek information 45 36.0 Meet a counselor 34 27.2 Other 7 5.6 (Source: Author's data processing) According to Table 5, 72 customers use the chatbot of VPBank to open more cards, equivalent to 57.6% of people. Customers can open cards directly with Chatbot without going through the bank office. All operations are solved online so that it is time-saving for users. Moreover, 64 customers (51.2%) approached chatbot of VPBank to view information about savings and loan interest rates, and 45 customers (36.0%) used chatbot of VPBank to seek information (27.2%). In contrast, only seven customers (5.6%) accessed chatbots to meet counselors. 917
4.2.4. Benefits of using Chatbot VPBank Table 6: Benefits of using VPBank Chatbot Criteria Number (N) Percentage (%) Quick response 87 69.9 Easy and simple communication 76 60.8 Obtaining the desired correct answers 37 29.6 No waiting time 19 15.2 (Source: Author's data processing) The research result in Table 6 shows that 87 (69.6%) customers found that they received quick responses from chatbots. Based on the pre-programmed questions and answers, the chatbot will reply immediately when the user chooses the inquiry. Next, 76 customers (60.8%) can easily communicate with chatbots. In addition, 29.6% of customers received the desired answers. It proves that the chatbots partially meet the user's requirements. However, when the customer makes requests not in the available scenarios, the customer will not receive the answers. Therefore, the bank should solve this problem by adding circumstances and choices corresponding to the generated questions. Lastly, 19 respondents (15.2%) choose chatbot because it is time-saving. 4.2.5. Disappointment when using Chatbot of VPBank Table 7: Disappointment when using Chatbot of VPBank Criteria Number Percentage (%) (N) Receiving undetailed answers 89 71.2 Complaints cannot be resolved 62 49.6 Limited feedback 37 29.6 Untrustworthy tree 14 11.2 Interacting with Chatbots is not the 25 20.0 same as interacting with humans (Source: Author's data processing) The survey outcome showed that 89 customers (71.2%) said that they did not receive detailed answers. The chatbot is pre-programmed, so the answers can not go into detail in some circumstances. The second biggest disappointment is that chatbots could not manage complaints (49.6%). The VPBanks chatbot is a simple type with prepared answers. Therefore, banks often combine chatbots with staff to solve these situations. Next, 37 customers (29.6%) think that chatbot feedbacks are limited. Lastly, 20% of customers stated that interacting with a chatbot was not similar to interacting with a human because a 918
chatbot works mechanically with unnatural language. Thus, it confused the user during the conversation. 4.3. Customer acceptance of VPBank Chatbot 4.3.1. Exploratory factor analysis for independent and dependent variables Exploratory factor analysis is applied to reduce and summarize variables for research into concepts. The relationships among many variables are shown, and the factors that represent the observed variables are found by factor analysis. Exploratory factor analysis should be based on specific and reliable criteria and should satisfy the following conditions: (1) According to Kaiser (1975), the KMO coefficient (Kaiser - Meyer - Olkin) in the range from 0.5 to 1 (0.5 ≤ KMO ≤ 1) is used as the index to consider the appropriateness of data when analyzing factors; (2) Barlett’s test is used to evaluate the correlation between the observed variables in the population and has statistical significance when Sig. 1 will be kept in the analytical model; (4) Total Variance Explained shows how much of the extracted factors are condensed and the EFA model is considered appropriate when the Total Variance extracted is greater than or equal to 50%; (5) Factor Loading is the value indicating the correlation between the observed variable with the factors and the value of the factor loading coefficient is greater than 0.5, the observed variable is statistically significant good (Hair et al., 2010). Table 8: KMO and Bartlett's Test for the independent variable Evaluation factor Independent variables KMO. Coefficient 0.74 Sig value. in Bartlett's test 0.000 (Source: Author's data processing) For the independent variable, based on the above table, we see that the coefficient KMO = 0.74 (satisfying the condition 0.5< 0.74 1, there are 5 factors with the Total Variance extracted. Total Variance Explained 62.79 % (>50%) indicates that these 5 factors will explain 62.79 % of the variation of data. After analyzing EFA factors for 21 observed variables, all observed variables meet the conditions well to conduct the analysis. Table 9: Rotation matrix of independent variables Factor Observed variables 1 2 3 4 5 PR1 0.883 PR4 0.883 PR2 0.876 919
PR3 0.876 AT1 0.809 AT3 0.792 AT4 0.788 AT2 0.745 PEOE5 0.804 PEOE3 0.780 PEOE2 0.731 PEOE1 0.637 PEOE 4 0.508 PU4 0.770 PU3 0.722 PU1 0.674 PU2 0.649 PA1 0.790 PA3 0.680 PA4 0.586 PA2 0.575 Eigenvalue coefficient 1.321 Extracted Variance (%) 62.79 (Source: Author's data processing) For the dependent variable, based on the exploratory factor analysis results, the dependent variable consists of 3 observed variables, showing that the significance level of Bartlett's Test is less than 0.05; KMO value is 0.742 (>0.5), total variance extracted is 65.025% and Factor Loading factor of all observed variables is more than 0.5, proving that factor analysis is appropriate. Table 10: Results of factor analysis for dependent variable Element Value KMO Coefficient 0.742 Sig value – Bartlett. Test 0.000 920
Total variance extracted 59,706 Eigenvalue coefficient 2,388 AC3 0.833 AC4 0.828 AC2 0.795 AC1 0.615 (Source: Author's data processing) 4.3.2. Correlation analysis and regression results The correlation coefficient test aims to assess the linear relationship between the independent variables and the dependent variable. If the variables are highly correlated, the problem of multicollinearity must be taken into account after the regression analysis. According to the correlation coefficient matrix, the dependent variable has a linear correlation with the three independent variables. The Pearson correlation coefficient test table reveals that the correlation coefficient between the dependent variable AC and independent variables that have a significance value of Sig.< 0.05 are PU, PR, and PA. The results of the regression are shown below: Table 11: Regression results Variable Regression coefficient Test value t Coefficient - - PU 0, 102 *** 2.414 PR -0.93 ** -3.972 AT 0.802 *** 17,516 R 2 = 0.829 R 2 adjusted= 0.824 F = 195,063 *** Durbin Watson coefficient = 1.902 (Source: Author's data processing) The result indicates that the adjusted R2 is 0.829. It means that 82.9% of the variation of the dependent variable is explained by the linear relationship of the independent variables. It demonstrates a relatively high level of model fit. The Durbin- Watson score for the model is 1.902 (1 < D < 3), indicating that the model does not have autocorrelation. However, the fit is only valid for the sample data. To test whether the model can be applied to the real population or not, the ANOVA test is performed. The results demonstrate that the F-test's significance level in the model is small (
indicating that the model is suitable for the dataset and can be generalized to the entire population or that there is a significant relationship between the independent variables and the dependent variable. The coefficients of the independent variables' VIF (covariance) are all less than 10, indicating that multicollinearity in the regression model is within the acceptable threshold. Based on the Sig values of the t-test of independent variables, such as PU having a Sig value of 0.0017, PR with a Sig value of 0.000, and PA with a Sig value of 0.000, it can conclude that the regression model is appropriate and exhibits the relationship between acceptance and the three factors. According to the normalization coefficient, the regression equation takes the following form: AC = 0.783PA + 0.102PU – 0.163*PR From the normalized regression equation Beta, the positive sign of the regression coefficient indicates that the factors in the regression model have a proportional influence on the attitude towards accepting the use of Chatbot. It means that when the customer's evaluation level for the Incremental models of these factors increases, the level of acceptance of using Chatbot will increase. Conversely, the negative sign of the regression coefficient indicates that those factors have a negative relationship with chatbot acceptance. Looking at the regression model, it can be seen that two factors have a positive influence, and one factor negatively affects the acceptance of Chatbot use. The coefficient β1 = 0.783 shows that when the attractive variable (PA) changes by one unit while the other variables remain unchanged, the acceptance (AC) moves in the same direction by 0.783 units. The coefficient β2 = 0.102 shows that when the utility variable (PU) changes by one unit while the other variables remain unchanged, the acceptance (AC) moves in the same direction by 0.102 units. The coefficient β3= -0.163 indicates that when the risk perception variable (PR) changes by one unit while the other variables remain unchanged, the acceptance (AC) moves in the opposite direction by - 0.163 units. The research results indicate that three factors impact consumers' acceptance of chatbots in the banking industry. The perceived attractiveness influences most significantly consumers' willingness to use a chatbot because the target audience is young and open-minded. Therefore, they are more willing to experience and use technology than other age groups. Moreover, perceived usefulness has a positive connection with the intent to use the chatbot. The usefulness of chatbots may be quick response, solving problems effectively, or saving time and cost. However, users' acceptance is negatively impacted by the perceived risk. Chatbots can collect data from customers during the conversation. Hence, consumers have the fear of losing private information or financial data. The research results are considered a crucial contribution to the research thesis on accepting the use of new technology for the TAM model. 5. Conclusion and implications In the era of technology and the changes in consumer behavior, companies have applied chatbots to optimize service operations. The study outcomes showed that customers of VPBank have experienced the bank chatbot on Facebook and NEO app. 922
Moreover, users found that using the chatbot has both advantages and disadvantages. In addition, three factors impact customers' perception of using chatbots at VPBank, Hue Branch including perceived attractiveness, perceived usefulness, and perceived risk. From the results, some implications are figured out: Firstly, as for the attractiveness of using chatbots, banks should build a chatbot system with many different scenarios to best respond to customers' needs. Simple things such as addressing the customer by name when texting or inserting customer's social network account name into the message can help the business reach more customers and encourage them to use the bank's services. Thoughtfully preparing chatbot scripts will improve customers' experience and make them feel like they are being talked to, listened to, and consulted by the business. Furthermore, a more customer-friendly chatbot interface can make customers excited to use it. Secondly, to address the risk perception factor when using the chatbot, especially in the banking sector where privacy and data security are crucial, banks must ensure that customer data is owned and can only be accessed by the bank. The bank should have more solutions to protect customer information and commit to securely protecting customer information and not misusing it. Building trust in user protection when using the chatbot in banks is crucial. Thirdly, it is necessary to build diverse questions that provide more features so that customers have more choices. Banks should offer more information about interest rates and exchange rates directly in the conversation without customers having to go to a link provided by the chatbot. It can make customers feel that chatbots can provide many benefits. Updating more features for chatbots and combining chatbots and human service when customers make complex requests that the chatbot cannot fulfill can help avoid customer frustration. Chatbots have the potential to positively impact customer loyalty and brand engagement while helping to increase revenue and reduce costs. However, this will only happen if they deliver the experience customers want. Choosing the right development platform that offers intelligent communication, scalability, and control is critical to success. However, this study has some limitations. Firstly, the sample size of 125 may only apply to the current set of customers of VPBank. Therefore, in the future, the research should be applied to a bigger sample size from different banks using chatbots. Secondly, there was no difference in adoption among groups. Hence, researchers in the future should make a comparison among categories so that suggested marketing plans for each segment will work effectively. REFERENCES 1. Ajzen, I., & Fishbein, M. (1980). Understanding and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall 2. Ajzen, I., & Fishbein, M. (1975). A Bayesian analysis of attribute processes. Psychological bulletin, 82(2), 261. 3. Accenture. (2018). Why Chatbots Are Key to the Future of Customer Experience. 923
https://www.accenture.com/us-en/insight-digital-customer-service-Chatbots, accessed 20 April 2023 4. Bansal, H., & Khan, R. (2018). A review paper on human computer interaction. Int. J. Adv. Res. Comput. Sci. Softw. Eng, 8(4), 53. 5. Choi, J., Lee, H., & Kim, C. (2011). The effects of IT innovation on service quality and perceived value in the hotel industry. International Journal of Hospitality Management, 30(2), 335-344. 6. Chatbots Magazine. (2017). How Chatbots are Revolutionizing the Entertainment Industry. https://Chatbotsmagazine.com/Chatbots-revolutionizing-entertainment- industry-acd0a74825ae, accessed 20 April 2023 7. Davis, FD 1989. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly 13 (3):319-340. 8. Dobhalkar & Bagozzi (2002). An Attitudinal Model of Technology-Based Self- Service: Moderating Effects of Consumer Traits and Situational Factors 9. Dwivedi, YK, Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., ... & Williams, MD (2021). Artificial Intelligence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice and policy. International Journal of Information Management, 57, 101994. 10. EdTech Magazine. (2018). How Chatbots Are Enhancing the Learning Experience. https://edtechmagazine.com/k12/article/2018/03/how-Chatbots-are-enhancing- learning-experience-perfcon, accessed 20 April 2023 11. Forbes. (2017). How Chatbots are Changing the Banking and Finance Industries. https://www.forbes.com/sites/tomgroenfeldt/2017/02/21/how-Chatbots-are-changing- the-banking-and-finance-industries/?sh=1255ca5c5b5b, accessed 20 April 2023 12. Healthcare IT News. (2017). AI-Chatbot Applications in Healthcare Gain Traction. https://www.healthcareitnews.com/news/ai-Chatbot-applications-healthcare-gain- traction, accessed 20 April 2023 13. HubSpot. (2018). How to Use Chatbots for Customer Service: 12 Examples of Companies Getting It Right. https://blog.hubspot.com/service/Chatbot-examples- companies, accessed 20 April 2023 14. Hair, JF, Ringle, CM, & Sarstedt, M. (2011). PLS-SEM: Indeed a silver bullet. Journal of Marketing theory and Practice, 19(2), 139-152. 15. IBM. (2017). Chatbots in Industry: The Rise of Conversational Commerce. https://www.ibm.com/blogs/watson/2017/10/Chatbots-industry-rise-conversational- commerce/, accessed 20 April 2023 16. Jia, J. (2003). The study of the application of a keywords-based Chatbot system on the teaching of foreign languages. arXiv preprint cs/0310018. 17. Khanna, A. (2015). Pandorabots Chatbot Hosting Platform. SARANG Bot. 18. Kehr, F., Kowatsch, T., Wentzel, D., & Fleisch, E. (2015). Blissfully ignorant: the effects of general privacy concerns, general institutional trust, and affect in the privacy 924
calculus. Information Systems Journal, 25(6), 607-635. 19. Li, X., Hess, TJ, Valacich, JS, & Yi, Y. (2012). Exploring the potential dark side effects of social media: A tentative model. In Proceedings of the 33rd International Conference on Information Systems. 20. Oh, SH, Lee, SY, & Han, C. (2021). The effects of social media use on preventive behaviors during infectious disease outbreaks: The mediating role of self-relevant emotions and public risk perception. Health communication, 36(8), 972-981. 21. Rese, A., Ganster, L., & Baier, D. (2020). Chatbots in retailers' customer communication: How to measure their acceptance?. Journal of Retailing and Consumer Services, 56, 102176. 22. Richad Richad, Vivensius Vivensius, Sfenrianto Sfenrianto and Emil R. Kaburuan (2019) Anlysis of factors influencing millennial's technology accetance of Chatbot in the Bankking Industry in Indonesia 23. Shin, DH (2010). User acceptance of mobile Internet: Implications for convergence technologies and business. Journal of Business Research, 63(6), 578-586. doi:10.1016/j.jbusres.2009.01.024 24. Shweta, K., & Kelly, L. (2022). Delivering great customer service with Chatbots. 25. Sarwar, M., & Soomro, TR (2013). Impact of smartphone's on society. European journal of scientific research, 98(2), 216-226. 26. Schlicht, M. (2016, April 20). The Complete Beginner's Guide To Chatbots - Chatbots Magazine. 27. Trivedi, J. (2019). Examining the customer experience of using banking Chatbots and its impact on brand love: The moderating role of perceived risk. Journal of internet Commerce, 18(1), 91-111. 28. Travolution. (2017). Chatbots Drive Change in the Travel Industry https://www.travolution.com/articles/102695/Chatbots-drive-change-in-the-travel- industry, accessed 20 April 2023 29. Van der Heijden, H. (2004). User acceptance of hedonic information systems. MIS quarterly, 695-704. 30. Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda on interventions. Decision Sciences, 39(2), 273-315. doi: 10.1111/j.1540- 5915.2008.00192. 31. Wu, JH, & Wang, SC (2005). What drives mobile commerce?: An empirical evaluation of the revised technology model. Information & management, 42(5), 719- 729. 32. Yang, KC, Hsieh, CT, & Yang, CH (2015). Investigating information security policy compliance in online banking: A self-determination theory perspective. International Journal of Information Management, 35(1), 67-77. doi:10.1016/j.ijinfomgt.2014.09.006 925