The Impact of Chatbots on the Relationship between Integrated Marketing Communication and Online Purchasing Behavior in The Frontier Market

Artificial Intelligence (AI), applied in many fields, is the core of the fourth technological revolution. In business, AI is used for customer relationship management as applied in the autoresponder systems, i.e., chatbot. Chatbots were an essential tool in the marketing relationship as many companies applied this function to their website; hence, this study analyzed the influence of chatbots on the enterprise's integrated marketing communication (IMC) activities, resulting in impulse purchase behavior and repurchase intention behavior. The mixed research method was used, particularly the in-depth interview and the survey with 886 online consumers, who shop from the online websites with chatbots system in Vietnam as Tiki, Lazada, Sendo, excetera. The research results showed that the perceived usefulness and ease of use of chatbots have positively affected the attitude of online consumers to the IMC activities of businesses. Simultaneously, IMC leads to impulse buying as well as the repurchase intention behavior of customers. The study proposed some managerial implications for an online business to enhance the chatbot functions to consumer behaviors in the website.


Introduction
Chatbots are heavily used globally in education (Kane, 2016;Smutny & Schreiberova, 2020); healthcare service, or medical purposes (Chukhleb, 2020;Viswanath Prakash & Das, 2020); banking (Li et al., 2020). However, the research into the application of chatbots in integrated marketing communication in the electronic commerce industry, the central customer concept (Schultz, 2016), has not been paid attention to by the scientific community. Although expected as a new communication channel to make a difference between online businesses, chatbots are only used as a temporary tool to replace customer care staff when they can not answer customers directly (Van den Broeck et al., 2019). Another critical point is that IMC analysis research is still concentrated in developed countries or managers' points, although the customer-centric view or the emerging and frontier markets are potential for behavior science or economic growth; therefore, it is crucial to consider that IMC is significant in frontier markets Wagner Mainardes et al., 2019). Another consideration is that the enterprise IMC impact studies focus on two dimensions, i.e., customer-oriented and technology-oriented, which are the generic concepts of a technology (Butkouskaya et al., 2020;Peltier et al., 2003). Based on the Technology Acceptance Model (TAM), the perceived usefulness can be considered customeroriented, and the perceived ease of use is related to the technology-oriented system (Davis, 1989). Finally, the business cares about the customers' behavior, which is influenced by its communications activities. Advertising and public relation aim to 22  P-ISSN: 2086-1559, E-ISSN: 2527-2810 Jurnal The Messenger, Vol. 13, No. 1 (2021), pp. 19-32 customer list for e-commerce sites is not too difficult. In some studies, impulse purchase behavior has been known to increase consumers' lives simply because they want to experience the feeling of shopping, relieving the mind when done the act of buying something that they like (Wood, 2005). The impulse purchase behavior is associated with the decision to buy quickly, more emotional than rational (Rook, 1987;Rook & Hoch, 1985), and does not include buying a product to fulfill a predetermined purpose (Beatty & Ferrell, 1998).
However, the critical issue that determines a business's existence is whether customers continue shopping on e-commerce sites or not? The customer's repurchase behavior expresses the answer to that problem after the first purchase due to improvisation. Repurchase behavior is an expression of customer loyalty (Khoa, 2020a). Customers are willing to invest in loyalty when they provide superior value to competitors' services (Reichheld & Schefter, 2000). The intention to repurchase refers to the possibility of using a service provider again in the future (Fornell, 1992). Boonlertvanich (2011) defined repurchasing intent as the ability to use a brand again in the future.

Methods
Perceived usefulness of technology was that consumers find that using technology in purchases saves them time, effort, and they can make purchases, thereby increasing acceptance and intention to use it in the future. Perceived ease of use, which is an essential factor in the TAM, is the level at which a person believes that using a particular system will require no effort (Davis, 1989;Khoa, Nguyen, Tran, et al., 2020). Integrated communication emphasizes the relationship between the provided content and customers' information needs (Bruhn & Schnebelen, 2017;Khoa, Huynh, et al., 2020). The user-oriented dimension of chatbots will better satisfy the information needs of customers; chatbots will answer the question "what is beneficial for purchasing?" and "how to buy?" (Narver & Slater, 1990). Thereby, it will build a two-way relationship between customers and businesses. Therefore, chatbots will be able to influence customers' experience of business communication positively. It is impossible to ignore the ease of use in technology adoption (Khoa et al., 2021). In the context of the technological revolution, technology's rapid change creates enormous pressure on users, such as customers (Khoa, 2020c;Schultz, 2016). A typical activity of mobile commerce services is that consumers do not interact directly with service providers. Therefore, ease of learning and using is crucial for chatbots regardless of whether consumers are technically proficient (Dai & Palvi, 2009). Therefore, the use of easy chatbots will positively affect the perception of the value of communication activities (Ndubisi et al., 2008) and the usefulness of chatbots themselves (Davis, 1989).
Communication activities are an essential activity of all businesses; however, it is a complex activity that requires the active participation of both businesses and customers. When there is a need for information, customers often use many IMC tools as advertising, personal sales, or email marketing from different sources, i.e., companies, competitors, communes (Finne & Grönroos, 2009). Therefore, if the information from the communication activities is incorrect, it will lead to false perception values and affect IMC activities (Lee & Park, 2007). Chatbots are used as a direct response and interactive marketing tool to enable businesses and customers to provide two-way information online. Consumers can receive instant advice such as selling and purchasing impulse purchasing (Villarreal, 2010). Currently, businesses The Impact of Chatbots on ... (Bui Thanh Khoa) focus primarily on web-based marketing communications to influence impulse purchasing (Al Khattab et al., 2015). Besides, repurchase behavior is essential in customer behavior research (Khoa, 2020b;Schultz, 2016). Customer repurchase behavior depends on various factors, including evaluating business communication (Lee & Park, 2007). Communication content inconsistent with consumers' views can negatively affect consumer trust and long-term relationships (Bruhn & Schnebelen, 2017;Moon et al., 2017). Therefore, IMC activity will affect the impulse purchase behavior and the intention of repeating when buying online. Moreover, impulse buying had an impact on repurchase intention (Lee, 2011). Hence, it can be inferred that consumer experience through impulse in the purchasing process directly affects customer satisfaction leading to post-purchase behavior such as intent to buy back and word of mouth (Hamza, 2013). The study proposed hypotheses: H1: The perceived usefulness of chatbots positively affects customer-centered IMC in e-commerce H2: The perceived ease of use of chatbots positively affects customer-centered IMC in e-commerce H3: The perceived ease of use of chatbots positively affects the perceived usefulness in e-commerce H4: Customer-centered IMC positively affects impulse purchase behavior in ecommerce H5: Customer-centered IMC positively affects repurchase intention in e-commerce H6: The impulse purchase behavior positively affects repurchase intention in ecommerce This study focused on online companies, which lead the online market in Vietnam as Tiki, Lazada, and also Sendo. The quantitative method was applied to explore this study's new ideas related to the research problem (Khoa, Hung, et al., 2022). The results of in-depth interviews showed that the scales of perceived ease of use (PEOU) and perceived useful (PU) was initially adopted from the study of Khoa, Nguyen, Tran, et al. (2020) to fit with the use of chatbots in e-commerce, including four items of PEOU construct and four items of PU construct. Customer-centered IMC (CIMC) consists of 5 items, and repurchase intention behavior (RIB) consists of 3 items, adapted from the research results of Butkouskaya et al. (2020). Impulsive buying behavior (IPB) was measured with three items referenced and corrected from Liu et al. (2013). In total, there are five constructs in this research model, including 19 items designed in the form of a 5-point Likert scale with 1: Strongly disagree, 2: Disagree, 3: Neutral, 4: Agree, 5: Strongly agree.
The participants are consumers living and working in Vietnam, a country with a frontier economy according to the statistics of Morgan Stanley Capital International (2020), and using chatbots during the online purchase process through electronic devices (i.e., smartphones, tablets, and laptops). Analytical data was collected using non-probability sampling by a purposive sampling method, suitable for differences in consumer behavior (Zikmund et al., 2013). The study surveyed 900 consumers after 886 remaining screening questions were appropriate for further analysis. The information about the participants presents in

Results
Chatbot has become an essential means for the business to communicate to online customers who shop from the website. The research results showed that the perceived usefulness and ease of use of chatbots have positively affected the attitude of online consumers to the IMC activities of businesses. Simultaneously, IMC leads to impulse buying as well as the repurchase intention behavior of customers In the first part, the research will assess the reliability of the scale via Cronbach's Alpha (CA); the convergent validity via the Composite Reliability (CR), the Average Variance Extracted (AVE), and outer loading; and the discriminant validity via the Heterotrait-Monotrait Ratio (HTMT). The assessment thresholds are in Table 2.  (2016) The construct reliability and validity assessment are present in Table 3. The minimum Cronbach's Alpha coefficient is 0.735, which is more than 0.7; therefore, all constructs in the research are reliable. Furthermore, the lowest Composite Reliability of all constructs in this research is 0.85, more than 0.7. The minimum of Average Variance Extracted is 0.653, which is higher than 0.5. Moreover, the minimum outer loading index of each construct in the research is more than 0.708. Hence, all constructs in the research got convergent validity.  The discriminant validity between the two constructs ensures the difference, and there is no correlation between measuring factors. This study used HTMT coefficients to confirm the discriminant validity (Joseph F. . According to Table 4, we have the value of the HTMT among the constructs is less than 0.85, showing the guarantee of discriminant validity.  -SEM) was verified by five steps process, which has the bootstrap procedure with 5000 random subsamples, including five steps below in Table 5 (Joseph F.  as (1) Collinearity is assessed via the VIF value, which is less than 3; (2) The path coefficient is assessed p-value, which is less than 0.05; (3) Coefficient of Determination (R² value) is 0.75, 0.50, or 0.25, respectively, described as the substantial, moderate, or weak; (4) Effect Size (f 2 value) is 0.02, 0.15, and 0.35, respectively, has small, medium, and significant effects; finally, Predictive Relevance (Q² value) is more significant than 0. Table 5 showed the result of VIF, R 2 , f 2 , and Q 2 . The maximum VIF in table 6 is 1.39, and all the VIF values are smaller than 3; therefore, the constructs in the research do not 26  P- ISSN: 2086-1559, E-ISSN: 2527-2810 Jurnal The Messenger, Vol. 13, No. 1 (2021), pp. 19-32 have collinearity with the others. Moreover, the CIMC, IPB, PU, and RIB have been explained by 34.7%, 28.1%, 21.5%, and 29.6% by the change of independent variables. Although the R 2 is less than 50%, they can accept that due to the R 2 value, it can be allowed with 20% in behavioral science (Joseph F Hair et al., 2017). The CIMC has a large effect on the IPB (f 2 value is 0.39); PEOU has a large effect on the CIMC (f 2 value is 0.369). Furthermore, PEOU has a medium effect size on the PU, of which f 2 value is 0.169; CIMC, IPB has a small effect size on the RIB with f 2 value, respectively, are 0.067, 0.048; PU has the small effect size on the CIMC with f 2 value is 0.059. Table 5 pointed out all Q 2 value are more than 0 (Q 2 CIMC = 0.222, Q 2 IPB = 0.180; Q 2 PU= 0.081, Q 2 RIB= 0.143).  Table 6 archived significantly with the 99% confidence level because the p-value was less than 0.001. Hence, all the hypotheses were accepted.

Discussion
The study was conducted in a frontier economy like Vietnam, different from economic and social conditions than developed or emerging economies. Chatbots is a new technology that helps businesses integrate marketing communications activities when customers use mobile devices to access business websites. The research has applied previous studies to build and reaffirm the relationships between structures such as perceived usefulness, perceived ease of use, customer-centered IMC, impulse purchase behavior, and repurchase intention.
First of all, the first three hypotheses are accepted with a 99% confidence level. The perceived usefulness of chatbots positively affects customer-centered IMC in ecommerce (Beta = 0.208, p-value = 0.000); the perceived ease of use of chatbots positively affects customer-centered IMC in e-commerce (Beta = 0.485, p-value = 0.000); and the perceived ease of use of chatbots positively affects the perceived usefulness in e-commerce (Beta = 0.338, p-value = 0.000). In particular, customers in the frontier economy are more interested in ease of use than useful. What stands out in this research is that the customer-centric IMC concept is considered in terms of customer attitudes. Therefore, the research results are in line with Davis's technology acceptance The Impact of Chatbots on ... (Bui Thanh Khoa) theory (Davis, 1989) and the empirical studies of other researchers related to technology (Lee, 2009;Luan & Teo, 2009) and chatbots (Rietz et al., 2019;Winkler & Söllner, 2018). If compared with the research of Butkouskaya et al. (2020), this result also creates a certain consensus if considering perceived usefulness as customer-oriented decision-making and perceived ease of use is technology-oriented decision-making. Therefore, the application of chatbots needs to be invested by businesses in both technical aspects and content. In the frontier market, mainly in developing countries, the technical level of customers is not high; therefore, the use of an automated communication system can cause difficulties for customers. Therefore, businesses need to use reputable chatbots systems, which should avoid technical errors that make customers confused in use. Besides, there should be a user guide located on the website to teach customers how to comunicate through chatbots. Usefulness also affect the perception of business communication. Hence, an online business needs to research questions frequently to complement the database of chatbots. In addition, it is possible to integrate the feature of introducing promotion programs during the display of chatbots to create more information for customer.
Note: ***p < 0.001 Figure 1. Hypotheses testing (source: Author's analysis) Secondly, this study also pointed the customer-centered IMC activities, which related the chatbots, will positively impact on the impulse purchase behavior (Beta = 0.530, p-value = 0.000); repurchase intention (Beta = 0.273, p-value = 0.000); and the impulse purchase behavior has the positive effect on repurchase intention (Beta = 0.232, p-value = 0.000). In particular, improvising purchases is strongly influenced by IMC activities through chatbots of the business on the website. This study shows that AI technology, i.e., chatbots are a factor that influences consumer emotions and leads to impulse purchasing behavior, and may also lead to customers' repurchase intention behavior on websites (Lim & Kim, 2020;Liu et al., 2013). Besides, the study reinforces a positive view of buy-back after the first impulse purchase (Lee, 2011). Thereby, the business should promote the effective use of chatbots in IMC activities on the website through several solutions such as enhancing the building of information in chatbots that are up-to-date and suitable to customers' needs. Because customers access business information through multiple communication channels, another important thing is that there is information consistency between chatbots and corporate policies.