In the present day, the technology has developed dramatically during pandemic, especially the communication between consumers and the brands. As a result, customers initially encounter the Chatbot when they contact call center services in the present day.
Problem Statement
LITERATURE REVIEW
Chatbot: Definition
Keywords/Variables
- Satisfaction
- Information Quality
- Perceived Usefulness
- Perceived Risk
- Perceived Ease of Use
- Anthropomorphism
Satisfaction is a concept for measuring the degree/level towards certain products or services based on the customer's expectations in the business field (Chung et al., 2020). Poor information quality can affect customer trust and tends to be negative (Mithas et al., 2011).
RESEARCH METHODOLOGY
- Quantitative Approach
- Sampling Plan
- Questionnaire
- Data Collection
- Analysis and interpretation
Respondents could voluntarily decide to participate in the study and could withdraw their participation at any time. This study conducts a questionnaire in two languages; Thai and English, which could reach potential Thai and foreign speakers who experience chatbots, and the questionnaire is distributed to social media platforms (Facebook public groups, Instagram and LINE private groups), the survey is distributed in the type of Google form line to acquire samples in different online communities, the distribution of the questionnaire is between October 14, 2022 and October 20, 2022. The survey collects a total of 158 responses, there are 108 valid responses from the total due to the reason that the eliminated samples in the screening questions do not fit the sample frame.
2 samples are eliminated in the item "Your age is over 15 years old", 6 samples are eliminated in the item "Have you currently used the Thai Telecom service for more than 3 months. The result from the online questionnaire (related to the quantitative approach) will be entered in the step further analysis IQ3: I think the chatbot provides reliable information IQ4: I think the chatbot provides the information that is sufficient volume for my needs.
PE3: I think chatbot technology is easy to use PE4: I think it is easy to learn how to use chatbot PE5: I can find information quickly using chatbot. AT6: I think the chatbot can experience emotions of human satisfaction ST1: I feel satisfied with the performance of this company.
FINDING AND DATA ANALYSIS
Demographic and respondents’ behavior output
According to the result in terms of age range, the data of the 20-29 year group exceeds more than half of all respondents, while the other age range groups seem to be in the minority compared to the 20-29 year group. The data starts with the current use of telecom chatbot services, the result shows the ratio of three chatbot brands (Aunjai; AIS, Mari; TRUE and DTAC). The 2-3 times a month and more than 3 times a month group is dramatically smaller than the once a month group.
According to the chart of the frequency of chatbot use in a month, I regroup from three groups into two groups; once a month remains unchanged, but 2-3 times a month and more than 3 times a month are merged into 2-3 times or more in a month. According to the reliability test, information quality, perceived usefulness, perceived ease of use and anthropomorphism have Cronbach's alpha exceeding 0.8;. Perceived Risk has a Cronbach's Alpha of 0.6, which is still considered an acceptable level of reliability (Ursachi et al., 2015).
According to the study's theoretical framework, there are five independent variables (Information Quality, Perceived Usefulness, Perceived Risk, Perceived Ease of Use and Anthropomorphism) and one dependent variable (Satisfaction). Perceived Risk has the highest mean score with 3.11, which is only one construct with a mean score of more than 3, Anthropomorphism has the lowest mean score with 2.34, and Satisfaction has the mean score of 2.7.
Test of Difference
- T-Test finding
- One-way ANOVA finding
Furthermore, High School of Below has the average score of 2.9 (chatbots have their own opinion) over the Bachelor Degree group and 2.8 (chatbot can experience human emotions) over the Bachelor Degree and Master Degree or higher group. Starting with information quality, Baht's group has an average score of 3, higher than Baht's group in terms of chatbot providing information with few errors. The group of less than 15,000 baht agree on the characteristic that chatbot improves service experience (perceived usefulness) and help chatbot to get information easily (perceived ease of use), compared to the group of more than 45,000 baht. Ultimately, a group of less than 15,000 baht agree on four characteristics (chat behaves humanly, chatbot has consciousness during the conversation, chatbot has its own opinion and chatbot can experience human emotions) in the anthropomorphism, the degree of agreement of less than 15,000 baht is higher compared to baht and more than 45,000 baht groups.
Referring to Table 4.6, there is one significant difference (sig.<0.05 in the column of Post Hoc Tests) in the construct of Information Quality, the group of DTAC users agree with the characteristic of "chatbots provide the complete information" to compare with Mari: Real users. In the aspect of Information Quality, the groups that prefer to use Check Balance and Pay Bills agree that chatbot provides the complete information, sufficient volume for their needs and relevant to their decision making which is in contrast to the group Call for Human Agent by the group payment accounts have the highest grade in the aspect of information quality. Furthermore, a group of accounts payable think that the chatbot is useful for solving telecom issues as opposed to Apply for the Package and Call for the Human Agent and think the chatbot is useful for improving service experiences.
Group Check Balance also agrees that the chatbot is easy to use compared to Call for Human Agent and group Check Balance also agrees that the chatbot helps the users to find information easily. Regarding anthropomorphism, the Pay Bills group agrees that the chatbot behaves humanly, with a difference from the Call for Human Agent and Pay Bills groups. The group also agrees that the chatbot can experience the human emotion, which is different among the groups of Report Internet Problem and call for human agent.
Regression Analysis
Agent and payment accounts correspond to the statement compared to Report internet problems, apply for the packages and Call for human agent. According to the model summary, all independent variables can explain the changes in overall satisfaction by 60.8% (Adjusted R Square: 0.608). The regression analysis indicates that there are three independent variables that have a significant relationship with satisfaction referring to the Sig.
Regarding Figure 4.10, three independent variables (information quality, perceived ease of use, and anthropomorphism) have a positive relationship with satisfaction. Information Quality has a Beta score of 0.239, Perceived Ease of Use has a Beta score of 0.369 and Anthropomorphism has a Beta score of 0.174. As perceived, ease of use has the most significant impact on satisfaction, while anthropomorphism has a lower impact on satisfaction.
CONCLUSIONS & RECOMMENDATIONS
- Conclusions
- Recommendations
- Limitation
- Direction of future research
Respondents agree that chatbots can potentially solve the simple needs of checking balances and paying bills; the users can do it easily and do not have to travel to stores or wait for the response from the human agents as the chatbots can respond immediately to the user's command. I would say from the finding that chatbots still have limitations to respond to customers in more complicated contexts. Chatbots still cannot cover all the problems of the users, especially if the users need more information to help make decisions. Users prefer to know the relevant information for making decisions. They believe that human agents can provide a more complicated service than a chatbot; therefore, customers use the chatbot as another channel to reach the customer service center to meet their specific needs.
In conclusion, chatbots still have limitations to solve complex problems, but chatbots have the strength to provide quality information and serve as a convenient channel for users. Finally, the screening question examined respondents who have never had chatbot experience, the target respondents would know the chatbot feature, but this finding does not include the chatbot feature understanding question scale to measure that how knowledgeable the respondent understands the chatbot feature. . Due to the reason that telecom brands are using different types of chatbots (Menu based Chatbot, Voice Bot, Keyword recognition Chatbot etc.) and users are also using these different types of chatbots across channels. different communication to reach the services.
Regarding the limitations, future research should focus on one specific type of chatbot from the general type of chatbot, which could distinguish differences in chatbot usage that lead to different levels of customer satisfaction with each type of chatbot. Future research could be extended to other industries, not just the telecommunications industry, and could compare this area with other countries other than the focus of Thailand.
Overcoming customer innovation resistance to sustainable adoption of chatbot services: A community-enterprise perspective in Thailand. Retrieved October 8, 2022, from https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/how-covid-19-has-pushed-companies-over-the-technology-tipping - point-and-transformed-business-forever. Perceived quality, perceived risk and customer trust influencing customer loyalty for environmentally friendly electronics products.
NBTC delays planned merger of DTAC/true, clears air for 'fake news' - thai enquirer main Retrieved on October 8, 2022 from https://www.thaienquirer.com/43412/nbtc-delays-planned-merger-of-dtac-true-for-another-month-clears-the-air-for-fake-news / . How can I help you today?” the use of AI chatbots in small family businesses and the moderating role of Customer Affective Commitment.
When do Ai chatbots lead to higher customer satisfaction than human frontline employees in online shopping assistance? Who or what to believe: trust and the differential persuasion of human and anthropomorphic messengers.
APPENDICES
APPENDIX A: Questionnaire
General Questions
Specific Questions
Demographic Questions 1. Please state your gender
