The Role of Artificial Intelligence in Forming Favorable Customer Attitudes on E-Commerce Websites

I.INTRODUCTION

The development of information and communication technology, which is strengthened by the emergence of artificial intelligence (AI), is being used by many companies to provide services to make things easier for their customers [1,2]. One industry extensively uses this technology in providing services and products to its customers is the e-commerce industry. AI can interpret external data and learn to develop adaptive strategies [3]. Examples of implementing AI in e-commerce are product search (using images, keywords, or voice), automatic recommendations, personalized dynamic pricing, and chatbots that speed up responses to customer questions [4]. However, the rapid development of information technology that competitors quickly adopt requires companies to be able to adapt their business models to customer needs [5]. Therefore, it is essential to determine a strategy for using AI to increase positive customer attitudes that benefit the company.

AI technology is creating a revolutionary transformational impact on service companies, changing how companies interact with their customers [6]. The success of interactions through technology can be measured by how involved customers are in using the technology [7]. Therefore, many studies on technology behaviour examine variables related to factors that influence technology engagement [8,9], and there is still minimal research measuring technology engagement in e-commerce. Understanding customer involvement in technology adoption is essential to direct relevant product development and increase market acceptance of these innovations, especially in the e-commerce retail market.

Although previous research confirms that AI can increase technology engagement and attitudes toward firms using technology, it has implications for increasing purchase intentions [7,10]. However, they ignore the vital role of perceived effectiveness of e-commerce institutional mechanisms (PEEIM), which has been proven to be able to increase purchase intentions by protecting consumers from the risks of e-commerce transactions [11,12]. PEEIM will manage customer perceptions of the uncertainty of the online trading environment that underlies repeated online transactions by ensuring consumers are free from any adverse risks [13]. When PEEIM is high, it will reduce uncertainty in online transactions because customers feel protected [14]. Therefore, this research will look at the role of PEEIM in moderating the variables of technology engagement and attitude toward the firm as an impact generated by AI by increasing repurchase intention.

This study is the only one that explores the influence of AI by looking at each of its constituent dimensions that can increase technology engagement and consumer attitudes toward websites. PEEIM moderation role will also strengthen the latest research results in e-commerce. Therefore, this research was completed in several stages. First, we formulate the problem by looking at conceptual and practical gaps based on previous literature to obtain research updates. Second, we determine the method for solving problems using a quantitative approach. Third, the data should be analyzed by interpreting it into research results. Fourth, the research results will be discussed, focusing on theoretical and practical implications.

This study is designed with a systematic flow that begins with identifying the problem and the novelty of the research as a strong research basis. The second section reviews related literature, builds a conceptual foundation through theory, and develops hypotheses that support the research. Furthermore, the third section explains the research methodology, including the research design, data collection techniques, and analysis methods used to ensure the validity and reliability of the results. The fourth section presents the findings and their discussion, with interpretations linked to theory and previous research. Finally, the conclusion summarizes the main findings, presents the research implications, and provides recommendations for further research.

II.LITERATURE REVIEW

AI is a technology that can provide services to consumers as humans do. AI can understand, think, respond, act, and make decisions just like humans simulated on machines based on the data obtained [15]. Some systems that can be used to optimize the use of AI are prediction and recommendation. The prediction system uses Graph Convolutional Networks (GCN), which aims to understand how the relationship between users and their interactions affects shopping preferences, and the user representation fusion mechanism that integrates various aspects of behavior, such as purchase history and interaction with reviews, to form a more accurate user profile [16]. The recommendation system uses the Temporal Interest Attention Network (TIAN), which groups items viewed in close time into the same interest bucket, allowing for more accurate item representation modelling. In addition, the Time-Enhanced Session Graph (TES-Graph) was developed to capture changes in user interests with adaptive weights on each side of the graph [17]. This method overcomes the interference caused by accidental clicks, improving the accuracy of e-commerce recommendations.

The advantages of this prediction and recommendation system not only increase the accuracy in understanding user preferences but also have a significant impact on the marketing and sales areas [18]. With the ability to offer more customized services and project consumer behaviour patterns more precisely, this technology is further strengthening the role of AI in modern marketing strategies [19,20]. Some of these marketing activities can be evaluated based on customer perceptions of the usefulness of AI.

AI marketing activities involve four dimensions, namely interaction, information, accessibility, and customisation [21]. Interaction refers to the communication process carried out between AI and customers. For example, interactions occur via chat platforms, voice calls, or visual interactions. AI can also be used to provide product information to its customers. For example, recommending products based on purchasing history or customer preferences and making it easier for customers to search for products based on uploaded images. Accessibility refers to the response to information obtained from customers. AI acts as a customer service that can provide services in real-time quickly and accurately. Customisation relates to services that can be personalized. The buyer’s advantage is that he gets personalized service in the form of an offer that has beneficial value with the potential to reduce search costs, while the seller’s advantage is that he can sell at a high price to a party with a high assessment [22].

When AI can provide good service to its customers, it will have implications for increasing customer engagement using this technology [23]. We chose the technology engagement variable to be specified in this study because it consists of items that can capture consumers’ perceptions regarding technology use, and respondents may have different views on their level of involvement. Technology engagement is defined as the cognitive and affective commitment of a relationship when using technology [24]. Consumers who have become familiar with AI can form positive or negative attitudes, which will determine how involved they are with the technology [7]. In this paper, we attempt to validate the following hypotheses:

H1: AI Marketing has a positive and significant effect on increasing technology engagement

One of the critical issues discussed in looking at the results of technology creation is measuring attitudes toward using technology. Attitude is defined as a person’s feelings that can give positive or adverse reactions to other objects [25]. In the technological context, it is defined as the behavior of assessing likes or dislikes of the use of technology [26]. When using technology, Companies can predict interactions through their optimism and insecurity [27]. This attitude can be influenced by the use of artificial intelligence [7,28]. However, they ignore information about the factors shaping marketing AI that could improve attitudes towards technology. Therefore, we investigate the importance of existing service features in AI for improving customer attitudes. In this paper, we attempt to validate the following hypotheses:

H2: AI Marketing has a positive and significant effect on increasing attitudes towards firms that use technology

Online consumer behavior can be understood in two stages. The first is to encourage new people to buy online, and the second is to enable customers to buy again [29]. The fact is that getting new customers will take more time and energy than retaining existing customers [30]. Companies will spend many resources if they allocate a lot to finding new customers. Therefore, we use the repurchase intention variable to see indicators of the success of using AI on e-commerce websites.

Repurchase intention is defined as the probability that a customer will continue to buy products from the same seller [31]. We measure the extent to which online shoppers via e-commerce are willing to choose the platform as their first choice in the future to show loyalty and willingness to recommend it to others. Several factors that can form high repurchase intention are increasing customer engagement using technology [32,33] and attitude towards firms that use technology [34,35]. Therefore, marketers must know whether engagement and attitude can increase repurchase intention. In this paper, we attempt to validate the following hypotheses:

H3: Technology engagement has a positive and significant effect on increasing repurchase intentions

H4: Attitude toward firm has a positive and significant effect on increasing repurchase intentions

Online shopping has several risks that customers may experience. Such as financial risk, information risk, and privacy risk, which can influence purchasing decisions [36]. This high risk influences companies to develop services that can guarantee consumer safety in all dangers through third parties. This method can facilitate increased trust when shopping online [37]. For example, they guarantee that ordered goods will arrive at the customer without damage or errors and that transactions will be managed financially [13]. PEEIM is consumers’ perception that third parties can protect and mitigate risks during online transactions. Obtaining positive customer perceptions using e-commerce websites is important to increasing repeat purchases.

Previous research investigated the critical role of PEEIM in moderating by increasing buyers’ interpersonal trust through online vendors to increase repurchase intention to purchase products in e-commerce [37,38]. However, more is needed to assume that the moderating role will be the same under different conditions and contexts. [37] There is no unique guarantee that vendors will always gain customers’ trust when using e-commerce websites. [39] also found that the influence of perceived benefits becomes insignificant with online trust because, at the post-purchase stage, users have had experience using the product. In contrast to customer attitudes and involvement, they will be able to improve their performance when they influence PEEIM because it can trigger internal cognitive and emotional behavioural responses [12]. Therefore, this gap will be answered by testing the moderating effect on the relationship between technology engagement, attitude toward the firm, and repurchase intention. In this paper, we attempt to validate the following hypotheses:

H5: PEEIM positively moderates the relationship between technology engagement and repurchase intention so that when PEEIM is higher, the relationship between these two variables is also more robust.

H6: PEEIM positively moderates the relationship between attitude toward the firm and repurchase intention so that when PEEIM is higher, the relationship between these two variables is more robust. Considering the importance of issues related to the use of artificial intelligence in e-commerce, we created a research model (Figure 1) that focuses on optimizing the use of artificial intelligence in the marketing context to increase technology engagement.

Fig. 1. Research model.

III.RESEARCH METHODOLOGY

IV.RESULTS AND DISCUSSION

V.DISCUSSION OF FINDINGS

The main objective of this research is to examine the influence of artificial intelligence on increased technology engagement and attitude toward the firm, which has implications for repurchase intention. This research also investigates the support for the contribution of the PEEIM moderating variable between these two variables and repurchase intention. This research has several significant research findings. First, the influence of artificial intelligence from interaction, information, accessibility, and customisation can have a considerable positive value on technology involvement and attitude toward the firm. These results support previous research, which explains that AI can increase customer engagement using technology through consumer perceptions of new experiences and help companies process consumer feedback to adjust the services provided [10,47]. Other results also confirm that AI can improve attitudes towards company websites. These results support previous research, which explains that the result of using AI is a positive attitude that feels the benefits when using technology [48]. The human-like intelligence possessed by AI is a unique characteristic for this technology to continue to be remembered by customers. Even in some cases, gender differences between women and men have different attitudes regarding accepting technology. Women have more negative attitudes than men toward receiving technology services [49]. However, this did not happen in Indonesia’s representative sample of men and women.

This research also confirms that high customer involvement can increase repurchase intentions. These results align with the research [32]. Customer involvement with technology is crucial to ensure the usefulness and success of creating technology in business. In this context, it is also important to note that our study is the first to use a testing methodology that looks at customer engagement based on the influence of artificial intelligence on repurchase intention in the case of e-commerce. Previous research only analyzed AI variables that focused on the stimulus organism response (SOR) theory [23,50], which had weaknesses in capturing the complexity of behavior produced by AI. In contrast, AI may struggle to capture the proper nuance and complexity of emotions, leading to inaccurate interpretations of customer engagement. Therefore, this research looks at customer engagement resulting from the influence of AI by looking at each of its constituent dimensions to predict the likelihood of customers making repeat purchases.

Repurchase intention measurement refers to the Theory of Reasoned Action (TRA) and developed into the theory of planned behaviour, which predicts a person’s behaviour based on tendencies or relative weights of attitudes, norms, and behavioural control based on the stimulus they receive [25,51]. This theory is widely cited by research related to consumer behaviour in digital media or e-commerce to predict user behaviour measured from intention to action [52,53]. Therefore, it is appropriate to use it to measure how likely someone is to make repeat purchases due to AI technology’s influence on e-commerce.

Another different result is that attitudes towards e-commerce websites have little influence on increasing repurchase intention. The high customer attitude resulting from the impact of AI is unable to increase repurchase intention. These results are different from previous research, which states that this attitude will influence someone to repurchase products in e-commerce [54]. This difference can lie in the ease with which a person uses technology and the complexity resulting in risks during online shopping [55,56]. Other supporting factors must be present for this positive attitude to benefit the company. Therefore, our study provides a new conceptualization regarding positive customer attitudes always being associated with company profits. This research explains that this is only sometimes true in different contexts.

We expanded the concept of consumer behaviour theory and acceptance model theory, especially on customer attitudes on e-commerce websites. If we look at it from the perspective that repeat purchases in e-commerce are primarily driven by pleasure and excitement during shopping [57]. The question is whether or not the AI used by customers can provide that feeling. It is limited to making the purchasing process more accessible, which this research confirms and does not benefit the company. Another fact is that customer perceptions of AI devices tend to lead to negative attitudes towards these devices due to the perception of threats to personal identity, so they do not provide any benefits related to task performance and service quality expected from AI devices [48].

This research also considers the moderating role formed from PEEIM in the relationship between technology engagement and attitude toward the firm with repurchase intention. The results show that PEEIM positively moderates the relationship between these three variables. This means that a high PEEIM value will strengthen the influence of technology engagement and attitudes towards the website on increasing repurchase intention. Consumers who believe in the effectiveness of e-commerce institutional mechanisms tend to be more confident and comfortable in transactions, thereby increasing the influence of technology involvement on their purchasing decisions. Previous research explains that the moderating role of PEEIM greatly influences someone to make repeat purchases [37,38]. This implies that PEEIM can be an essential source of customer attitudes and engagement on e-commerce websites, whether PEEIM is low or high. More specifically, when PEEIM is high, people tend to assume that e-commerce managers will comply with the regulations and rules provided by the government to protect consumers. If not, they will be punished and sanctioned by the government or regulator, so they don’t need to remember past experiences when shopping on e-commerce websites. On the other hand, when PEEIM is low, consumers will judge that existing mechanisms in e-commerce are not fully effective or do not guarantee security, comfort, and fairness in online transactions. However, they still rely on their first-hand knowledge to re-evaluate the platform’s feasibility. This shows that PEEIM has a strong influence on repurchase intention.

This research enriches the literature on digital consumer behaviour by investigating the factors and reasons why consumers make repeat purchases on e-commerce websites. This research provides a theoretical contribution by emphasizing the need to include relevant theories that have the potential for future research. We use a combination of the acceptance model and the Theory of Relationship Marketing theory related to commitment to show that a conceptual model can be proposed to further increase repurchase intention in e-commerce. This research is the starting point for further study because it is the first to show how the influence of AI technology can specifically increase repurchase intentions on e-commerce websites by paying attention to the moderating role of PEEIM.

VI.CONCLUSIONS

E-commerce companies worldwide are currently using artificial intelligence to respond to technological advances. The aim of utilizing this technology is to provide the best service by making it easier for consumers to shop for products online. The usefulness of AI has attracted the attention of academics and practitioners in identifying essential impacts, such as what can increase consumer purchasing intentions in e-commerce from this technology. Finally, our results showed that leveraging AI should ensure that customers buy products in e-commerce that were easy, safe, and user-friendly. Therefore, it was recommended that management adjust AI service features by focusing on the factors that shape customer engagement using this technology. We recommend that companies implement responsive chatbots to help answer questions and resolve problems quickly. The company continues using AI-based data analysis to understand customer preferences and behavior to provide more relevant and personalized product recommendations. It further optimizes the user experience with an AI-powered search system, ensuring customers can find the products they seek more easily. Another way is to offer personalized promotions and discounts based on a customer’s purchase history and interests. With these strategies, e-commerce companies can create a more exciting shopping experience, ultimately increasing consumers’ desire to make repeat purchases.