The Impact of Web-Based Augmented Reality on Continuance Intention: A Serial Mediation Roles of Cognitive and Affective Responses

Abstract

The aim of this study is to investigate how consumers’ cognitive and affective responses to web-based augmented reality affect their intention to continue to use augmented reality. The novelty of this study is the integration of the Stimulus–Organism–Response model with Technology Continuance Theory, allowing for an investigation of the relationships among the following critical variables: augmented reality (AR), utilitarian value, perceived risk, user satisfaction, attitude toward AR, and continuance intention. The study sample consisted of 452 participants. Data were analyzed using the Partial Least Squares–Structural Equation Modeling (PLS-SEM) approach. The results indicate significant direct relationships between all variables. Furthermore, this study demonstrated an indirect relationship between AR and continuance intention, mediated sequentially by cognitive responses, namely, utilitarian value and perceived risk, and affective responses, including user satisfaction and attitude toward AR. Consequently, it was revealed that all indirect relationships were significant, except for the pathways from AR to continuance intention involving perceived risk. This study presents key insights for online retailers, demonstrating how the integration of AR technology into conventional online shopping platforms can optimize user experiences by enhancing the cognitive and affective responses of customers. This, in turn, strengthens their intention to continue using AR technology, fostering sustained engagement and the long-term adoption of AR technology.

1. Introduction

The COVID-19 pandemic has accelerated the pace of digital transformation across various sectors, including the retail industry. This rapid shift has driven the adoption of emerging technologies, which enable multi-directional communication between buyers and sellers, fostering proactive engagement, interactivity, and value co-creation [1]. Augmented reality (AR) integrates virtual elements—such as images, text, and sound—into real-world environments, allowing users to visualize digital objects within their physical surroundings [2,3]. This integration relies on real-time object recognition, image processing, location services, and micro-devices such as cameras and sensors [4]. AR technology has been widely adopted across various industries, including tourism [5,6]; education [7,8]; gaming [9]; and online shopping [10,11]. This study specifically examines the application of AR in the context of online shopping, with a particular focus on eyewear products. By leveraging AR-based innovations such as virtual try-on, retailers can recreate the conventional shopping experience, enabling customers to interact with products virtually, demonstrating how the products will appear and fit in real-life scenarios [12,13]. As a result, AR is becoming increasingly prevalent in online retail environments, as its dynamic product visualization helps decrease the gap between online and offline shopping experiences [14]. Although retailers continuously strive to provide sensory information for product evaluation, such efforts often remain insufficient, preventing consumers from making confident decisions due to the absence of direct inspection or physical product trials [15]. Therefore, from a retailer’s perspective, AR enhances customer decision-making by providing a deeper understanding of products before purchasing [3]. From a consumer perspective, AR offers key perceived benefits, including informational and emotional advantages [16]. Informational benefits stem from realistic computer-generated product representations, which improves visualization and evaluation [17,18]; meanwhile, emotional benefits foster engagement through the creation of a more enjoyable and immersive shopping experience [12]. So, through AR consumers can engage in virtual try-on experiences, while retailers benefit from improved conversion rates [15]. Therefore, the implementation of AR in online retailing has garnered significant academic interest, particularly regarding its potential to enhance consumer experiences and behavioral intentions [19]. For instance, in the furniture industry, companies such as Amazon and IKEA have used AR to help consumers visualize how furniture might look when integrated with other households items [3]. Similarly, in the beauty sector, L’Oréal and Sephora have leveraged AR technology to allow consumers to examine beauty products before purchase [20]. Additionally, various leading brands—including Louis Vuitton, Gucci, Burberry, Levi’s, The Home Depot, Target, Converse, and Adidas—have adopted AR technologies to enrich customer experiences, strengthen engagement, increase profitability, and foster brand loyalty [21]. AR applications in retail include web-based platforms, in-store experiences, and mobile apps [22]. While numerous studies have explored the application of mobile AR apps [11,21,23], in-store virtual AR mirrors [24], and somatosensory augmented reality [25], there remains a limited body of research examining web-based AR technologies and their impact on consumer behavior [20,26]. Therefore, this study focuses on web-based AR, which allows users to access AR content directly through a webpage without requiring a separate app, thus enabling seamless interaction via their device’s camera.

Research on AR and behavioral intentions has indicated that AR enhances consumers’ affective aspects, including satisfaction [23], engagement [19], and attitude [27], as well as cognitive aspects such as interactivity [28], usefulness [29], and ease of use [13]. This study examines the impacts of AR on consumers’ cognitive responses, including utilitarian value and perceived risk, as well as affective responses, such as user satisfaction and attitude. While previous research has extensively explored the impact of hedonic [30] and utilitarian values [31] on user satisfaction, the role of perceived risk in web-based AR shopping experiences remains insufficiently examined. Although some studies have investigated the relationship between AR and perceived risk [13,26], concerns about product quality and fit persist due to the absence of physical interaction before purchase [20]. While AR’s immersive 3D visualization has the potential to reduce product uncertainty [14], its effectiveness in lowering perceived product risk and enhancing user satisfaction is still unclear. Furthermore, the mediating role of perceived risk in the relationship between AR shopping experiences, satisfaction, and continuance intention has yet to be thoroughly explored. Moreover, Jayaswal and Parida [32] demonstrated that research on continuance intention with respect to AR technology is still scarce, and only a few studies investigated continuance intention as a response to AR usage [33,34]. Therefore, to address this gap, this study aims to extend existing knowledge, investigating whether web-based AR can influence consumers’ cognitive responses by increasing utilitarian value and decreasing perceived product risk. These cognitive shifts are expected to indirectly impact affective responses such as user satisfaction and attitudes toward AR technology. Furthermore, this serial mediation effect is hypothesized to examine the extent to which cognitive and affective responses affect continuance intention, which is a post-adoption phase that occurs after actually having a satisfied shopping experience using AR in the initial phase.

Scholars have proposed various theories to understand the use of different AR apps for online shopping, such as the Technology Acceptance Model (TAM) [5,8,35,36,37], uses and gratifications theory (UGT) [19,38], media richness theory (MRT) [15,39], motivational model [31] and Cognition–Affect–Conation (C-A-C) [18], and situated cognition theory (SCT) [40]. This study integrates the Stimulus–Organism–Response (S-O-R) model with Technology Continuance Theory (TCT) and Perceived Risk Theory (PRT) in order to investigate how consumers’ initial adoption of AR influences their post-adoption behavior, specifically their intention to continue to use AR technology.

2. Theoretical Background and Hypotheses

2.1. S-O-R Model

The Stimulus–Organism–Response (S-O-R) model, proposed by Mehrabian and Russell [41], was originally designed for general environmental psychology; however, it has also proven successful in retail environments. In particular, prior studies have confirmed that the S-O-R model is critical in forecasting how the retail environment affects consumer decision-making [12]. The model describes sequential events, starting with the environmental factors (Stimuli) that cause changes in the consumer’s internal psychological states (cognitive and affective; Organism), which then lead to behavioral responses, such as acceptance or rejection (Response) [34]. Therefore, according to the S-O-R model, an individual’s “inner organism” changes in response to external stimuli, resulting in behavioral responses [42]. Several studies have utilized the S-O-R model to examine the effects of new retail technologies on the emotional and behavioral responses of consumers during online shopping [27]. The model has demonstrated exceptional effectiveness in understanding consumer behaviors within AR environments across various business domains, including interior design [4], cosmetics [20,34,43,44], furniture [23,45], advertising [2], tourism [6,30], and wearable products (e.g., sunglasses, clothes, accessories, shoes) [27,46]. As the S-O-R framework encompasses the technological features of AR applications and user experiences during interactions with AR, it was considered the most suitable framework to employ in this study, as the researchers aimed to explore how the integration of the AR technology into the online shopping process affects the psychological behaviors (cognitive and affective) of consumers and, sequentially, their continuance intention regarding AR technology.

2.2. Technology Continuance Theory (TCT)

Technology Continuance Theory (TCT) was formulated by Liao et al. [47], with the aim of anticipating and elucidating the adoption of information systems and users’ continuous usage intention. TCT was developed by merging three well-known IS models—namely, the Technology Acceptance Model (TAM), Expectation Confirmation Model (ECM), and Cognitive Model (COG)—into a single framework to understand and interpret user behaviors in relation to continuance in the use of technology [47,48]. In comparison to the other models, TCT signifies a substantial advancement, illustrating superior practicality and explanatory capability. The integration of attitude and satisfaction into a single model for continuance is a key theoretical contribution of TCT [47]. TCT was developed to illustrate that satisfaction is a key factor in predicting the behaviors of initial users who may not fully comprehend the technology’s performance, while attitude is more important for long-term users [19]. Thus, the TCT is applicable to users at various stages of the adoption life cycle, including initial, short-term, and long-term users [47]. Given that this study focuses on online shopping using AR technology, which has not yet been widely adopted, its primary emphasis is on satisfaction and the formation of attitudes, in order to examine the initial adoption of AR and its influence on post-adoption behaviors; particularly continuance intention.

2.3. Integrating the S-O-R Model with TCT

Building on the above discussion of technology adoption models, the S-O-R model stands out in terms of emphasizing external stimuli that shape emotional and cognitive states, thus driving behaviors. Unlike the TAM—which focuses on perceived usefulness and ease of use—the S-O-R model integrates both rational and emotional decision making. While UGT assumes that users actively seek media for specific needs, S-O-R considers external factors to unconsciously influence emotions and behaviors. MRT explains the effect of media richness in communication, but fails to consider psychological mediating effects, while SCT focuses on real-world learning without detailing how stimuli trigger internal responses. The S-O-R model is thus well-suited for understanding how AR shopping features influence consumer engagement, emotions, and purchase decisions. In addition, TCT differs from TAM, UGT, MRT, and SCT in that it specifically focuses on the long-term use of technology and post-adoption behavior, rather than just initial acceptance, use of media, or cognitive interactions. Considering the aforementioned aspects, the S-O-R and TCT frameworks were identified as the most suitable theoretical foundations for the development of the conceptual framework outlined in this study.

The integration of the S-O-R model and TCT provides a comprehensive framework for understanding both the initial adoption and long-term use of AR technology in online shopping. S-O-R explains how external stimuli (e.g., AR features) influence cognitive and affective responses, such as utilitarian value, perceived risk, satisfaction, and attitude, which in turn drive behavioral intentions, including both initial adoption and continuance. TCT extends this by emphasizing the evolving role of satisfaction and attitude over time, shaping consumers’ continued use of web-based AR shopping technology. Hence, the application of the TCT and PRT in this study establishes a conceptual framework that formulates sequential pathways within the S-O-R model, where AR features (Stimulus) influence cognitive and affective responses (Organism), which ultimately shape continuance intention (Response).

2.4. The Effect of AR on Utilitarian Value and Perceived Risk

Augmentation is a key and distinguishing feature of AR. As defined by Rauschnabel et al. [38], it refers to “the extent to which a user perceives the augmented content as realistic.” The literature identifies three primary types of augmentation; namely, self-augmentation (e.g., YouCam Makeup, Sephora), direct environment augmentation (e.g., IKEA Place), and object augmentation (e.g., Ray-Ban) [22]. It is worth noting that AR is uniquely characterized by augmentation, setting it apart from Virtual Reality and other interactive technologies [49]. Therefore, augmentation was chosen to represent AR as a stimulus factor in this study; henceforth, for sake of brevity, the term “AR” will be used to refer to this factor for the rest of this study.

AR involves overlaying digital objects onto the physical word, creating an interactive immersive shopping experience [49]. AR is an interactive tool that allows consumers to gather information, experience products, and share experiences [1]. AR-based interactions, such as virtual try-ons and AR filters, boost user engagement, fostering pleasure and emotional attachment [50]. For instance, in online apparel shopping, AR enables users to interact with virtual garments by adjusting their view using device motion sensors, allowing them to modify their fit and appearance by moving their arms to the right or left [14]. Therefore, both the cognitive and affective states of consumers are influenced. User engagement with web-based AR is shaped by a combination of utilitarian and hedonic benefits, as users pursue information, communication, recreation, and emotional gratification to satisfy their intrinsic needs [16].

Utilitarian value, as defined by [51], refers to “the extent to which a user evaluates AR as useful, both for making purchases and acquiring essential information for purchase decisions.” In the context of AR shopping, utilitarian value closely aligns with perceived usefulness—a core construct of TAM- [52], which significantly influences individuals’ adoption and utilization of information systems, shaping user behavior and intention [14]. One of the key advantages of AR is its ability to provide enriched and contextually relevant information [1]. Yoo et al. [15] declared that although consumers encounter both visual and verbal information, they often process visual information first, leading to visual and haptic imagery before perceiving verbal information. In this regard, providing detailed product information through AR enables consumers to assess the shopping experience as useful and helpful [1,2], assess product’s suitability, and make well-informed decisions with confidence [28]. Moreover, according to TAM, when web-based AR delivers valuable information, users experience greater enjoyment and satisfaction, which subsequently enhances their perception of the technology’s usefulness [30]. AR focuses on maximizing usefulness and is associated with task-oriented, functional, instrumental, and practical benefits. Consumers driven by utilitarian motives prioritize practicality, efficiency, and accessibility when evaluating a product or service [53]. Therefore, utilitarian value plays a critical role in shaping consumers’ cognitive evaluations, as it directly influences how well a product or service meets expectations, ultimately facilitating informed purchasing decisions [37] and strengthens technology adoption in AR shopping [23].

Perceived risk can be defined as “consumers’ understanding of potential uncertainties and negative outcomes associated with purchasing a product or service” [26]. The Perceived Risk Theory (PRT), initially proposed by Bauer [54], posits that consumers perceive purchasing decisions as inherently risky due to uncertainties regarding potential negative outcomes. Cunningham [55] expanded upon this theory by categorizing perceived risk into various dimensions, including financial risk, performance risk, social risk, psychological risk, physical risk, and time risk. In the 21st century, scholarly attention has increasingly focused on the application of perceived risk within the domain of online shopping [56]. Researchers have highlighted three dimensions of risk in the online shopping context: product risk, financial risk, and information risk [26]. To the best of the researcher’s knowledge, only a limited number of studies have explored perceived product risk [13,26]. Consequently, this study will focus exclusively on examining perceived product risk.

Performance risk—also called Functional Risk—presented by [55] is defined as the potential loss incurred when a product or service fails to perform as anticipated. This perception is based on two factors: uncertainty about the result of the purchase and uncertainty about the potential losses. Many consumers are uncertain about whether the product or service will meet their expectations [28]. This could relate to the product’s functionality, quality, or how effectively it fulfills their needs. While product risk also exists in traditional offline purchase context, it is magnified in online settings particularly for wearable products, because consumers are unable to physically touch or feel the product in a virtual environment [46]. Thus, the lack of physical interaction is a significant contributor to perceived product risk, as it hinders the consumer’s ability to evaluate the product’s features effectively [46,57]. However, the incorporation of AR technology allows consumers to use their smartphones to virtually try the product on, providing more information and presenting products in a 3D view [29]. In this way, consumers can gain a better understanding of the product features, reducing their risk perception and increasing confidence in their online choices [46]. Therefore, this study posits the following:

H1. 

AR has a positive impact on utilitarian value.

H2. 

AR has a negative impact on perceived risk.

2.5. The Effects of Utilitarian Value and Perceived Risk on User Satisfaction

AR technology provide customers with functional, hedonic, social, personal, and economic benefits, enhancing their overall shopping experience. Functionally, AR allows users to virtually interact with products, reducing uncertainty and improving decision making by providing a realistic preview of an item. These features not only increases convenience, but also reduces the cognitive dissonance associated with online purchases. Hedonically, AR tools offer an engaging and immersive shopping experience, transforming routine purchases into enjoyable activities. Socially, AR features integrated with mobile apps and social media platforms enable users to share their experiences with friends, fostering a sense of connection [50]. These multi-dimensional benefits work together to build cumulative satisfaction through consistently meeting and exceeding customer expectations throughout the online shopping experience.

Satisfaction, in this context, is defined as the overall assessment of a consumer’s experience with a service, characterized by a positive, neutral, or negative emotional state [30]. In the context of online shopping, satisfaction represents an affective response, which is particularly influenced by the use of AR technology [58]. This emotional response is reflected in the consumer’s overall evaluation of the shopping experience. When a user’s expectations of an AR shopping experience are met, they are more likely to feel satisfied, which can increase their engagement with the online retailer’s website [51]. Therefore, according to the expectation disconfirmation model, a consumer is satisfied when the online shopping expectations are higher than pre-adoption expectations, in which satisfaction will increase [47]. Consequently, satisfaction serves as a critical driver of long-term growth for online retailers [59]. In other words, success in online shopping depends on user satisfaction. Moreover, this cumulative satisfaction achieved through AR experiences in online shopping has a profound impact on electronic Word-Of-Mouth (eWOM) engagement [60]. When customers find value in AR tools, they are more likely to share their positive experiences on social media platforms, write reviews, and recommend products on their social networks. Satisfied customers, in turn, become advocates, sharing their experiences online, which amplifies the online retailer’s reach and credibility [61].

This advocacy behavior is largely influenced by the perceived value that customers derive from their shopping experience. The Expectation Confirmation Model suggests that perceived usefulness has a significant positive impact on user satisfaction with specific information systems [62]. In this context, utilitarian value is recognized as a fundamental determinant of user satisfaction [17,28]. It plays a critical role in shaping consumer evaluations, as it reflects the extent to which an experience is perceived as useful and informative in aiding purchasing decisions. Moreover, the effectiveness of the shopping experience is contingent not only on its usefulness but also on its user-friendliness, as ease of use and understanding are key factors influencing the overall user experience [53]. Extending this perspective to AR, prior research has also suggested that factors contributing to user satisfaction with AR are perceived usefulness, ease of use, and perceived enjoyment [12]. As such, user satisfaction is determined by consumer perceptions of the practical and hedonic benefits offered by AR content. When consumers perceive the web-based AR shopping experience as useful, easy to use, and enjoyable, they are more likely to feel satisfied with the overall online experience [45]. Therefore, as mentioned in [53], satisfaction is greater when interactions are perceived as both functionally beneficial and intrinsically enjoyable.

Perceived risk was formally defined as the fear of potential losses related to a purchase, serving as a barrier to making a buying decision [59]. AR presents an innovative solution to mitigate perceived product risk [46]. By substituting direct product experiences with virtual interactions, AR enables consumers to better understand product features and facilitates product evaluation [63]. Therefore, as uncertainty reduces, as noted in [3,28], choice confidence, decision comfort, and user satisfaction increases, as demonstrated in [64]. Aligned with Cognitive Dissonance Theory, when users perceive high risk in AR shopping, they may experience psychological discomfort, leading to lower satisfaction. Barta et al. [20] found that when consumers engage in virtual product try-ons and perceive the experience as useful and enjoyable, their psychological discomfort is mitigated, ultimately enhancing overall shopping satisfaction. Consequently, a positive correlation exists between perceived risk and the likelihood of abandoning an online purchase due to a dissatisfied customer [63].

It should be noted that perceived risk negatively influences the perceived utilitarian value of online services and shopping environments [65]. Utilitarian value, which encompasses practical and functional benefits such as convenience, product information, and cost savings, tends to be diminished when users perceive high levels of uncertainty or risk, including concerns related to security, privacy, product performance, or delivery [66]. According to prospect theory, consumers become risk-averse in uncertain situations and prioritize loss prevention over potential functional gains, shifting their focus away from task-oriented benefits [66]. Moreover, according to TAM, when users perceive online shopping as risky or uncertain, they are less likely to consider it useful [67]. This negative relationship implies that reducing perceived risks is essential to enhance user evaluations of a service’s usefulness and encourage their engagement or repeat purchase intentions [68]. According to the findings of Faqih [67], mitigating perceived risk in the online environment is likely to enhance consumers’ perceived utilitarian value in the online shopping domain, thereby positively influencing their intention to engage in online shopping.

In summary, delivering valuable and engaging user experiences plays a crucial role in enhancing user satisfaction within the online shopping environment. Simultaneously, reducing consumers’ perceived risk contributes to increased utilitarian value and overall user satisfaction. Therefore, the following hypotheses are proposed:

H3. 

Utilitarian value has a positive impact on user satisfaction.

H4. 

Perceived risk has a negative impact on user satisfaction.

H5. 

Perceived risk has a negative impact on utilitarian value.

2.6. The Effect of User Satisfaction on Attitude

An attitude toward technology refers to a specific feeling about it, shaped by a particular understanding of the technology, and encompasses a tendency to act in support of or against it [12]. It is viewed as one of the affective responses to AR usage. Thus, an individual’s attitude towards AR technology relies upon their perception and belief that it provides added value and delivers benefits [69]. The TCT was formulated to demonstrate that satisfaction serves as a crucial determinant in forecasting the behavior of initial users who may have limited understanding of the technology’s performance, whereas attitude plays a more dominant role in influencing long-term users [19]. Therefore, a consumer’s expectations are fulfilled, they experience satisfaction, which fosters positive attitudes toward future technology adoption [52]. This, in turn, leads to increased purchase intention, sales, and continuance intention [27]. Although previous research has theoretically examined and empirically confirmed that satisfaction serves as an antecedent of attitude in the online shopping context, this relationship has not been studied in the AR context [48,52]. To the best of the researcher’s knowledge, no studies have empirically investigated user satisfaction and consumer attitudes toward AR technology. Therefore, this study proposes the following hypothesis:

H6. 

User satisfaction has a positive impact on attitude.

2.7. The Effects of User Satisfaction and Attitude on Continuance Intention

In line with TCT, an individual’s attitude and satisfaction act as key drivers for their intention to continue to use a particular technology [47,48]. Continuance intention—also referred to as post-adoptive intention behavior [16,62]—has been defined by Liao et al. [47] as “an individual’s intention to use or reuse a particular system continuously.” In the early stages, users can only assess how well the anticipated performance has been achieved; therefore, their satisfaction arises from the fulfillment of expected outcomes in an AR shopping experience. However, after the initial use, they no longer base their judgment on pre-adoption expectations but, instead, rely on their initial perception of the technology’s usefulness to decide whether to continue using it [70]. As users start to evaluate its actual performance, their attitude toward AR begins to take shape, which then influences their ongoing engagement and intention to adopt the AR technology in the long-term [19]. Thus, in this study, the researchers specifically chose continuance intention to enhance the understanding of consumer behaviors in the post-adoption phase.

Moreover, several studies have theoretically and empirically proven that user satisfaction positively affects the intention to continue using a particular technology in several research areas, such as mobile banking [71], mobile wallet [52], and internet banking [72]. Furthermore, research on IT continuance has indicated that satisfaction plays a critical role as a precursor to continuance intention [58]. Similarly, the Expectation Confirmation Model asserts that satisfaction has the most profound impact on user behaviors in the context of information systems [62]. Therefore, satisfaction serves as a crucial indicator for evaluating the success and effectiveness of an information system. The authors have also stated that an increase in user satisfaction will lead to greater utilization of AR technology, subsequently resulting in a higher intention to continue its use [73].

According to the TAM, attitude toward a technology has a significant influence on behavioral intention during the initial phase of technology adoption. In contrast, the TCT posits that attitude plays a crucial role in shaping continuance intention in the post-adoption phase [47]. Thus, the significance of the influence of attitude on continuance intention has been confirmed in different information system contexts; for example, in [74], it was found that attitude has a higher contribution to the continuance intention of mobile taxi booking application users. Similar findings have been reported in other research settings, such as mobile banking [71,72], mobile bike-sharing [75], and mobile wallets [48,76], as well as mobile payment [77]. Moreover, prior research in interactive marketing has demonstrated that consumers’ intention to continue to use an interactive technology such as AR, significantly enhances their likelihood of revisiting shopping websites while simultaneously diminishing their inclination to switch to competing brands and retail platforms [25].

It should be noted that the integration of generative artificial intelligence (AI), such as ChatGPT, into online shopping using AR can create a transformative and highly personalized shopping experience [78]. The motivational determinants of continuance intention regarding generative AI offer a complementary perspective [79]. Factors like satisfaction, intrinsic motivation, and the perceived values gained from using ChatGPT align closely with the psychological needs that AR addresses in online shopping [80]. For ChatGPT users, practical benefits such as problem-solving and convenience drive perceived usefulness, while personalized and engaging interactions provide enjoyment similar to AR technologies [81]. The commitment of users to AI platforms is strengthened by satisfaction and perceived quality, emphasizing the importance of meeting user expectations and addressing psychological needs, again similar to the crucial roles that these factors play in AR technologies [53]. Together, these findings suggest that fostering both utility and emotional engagement is essential for sustaining long-term user loyalty across innovative digital platforms. By addressing these factors and meeting user expectations, ChatGPT can enhance satisfaction, build commitment, and drive long-term adoption, much like AR technologies in the context of digital innovation [82]. This highlights the importance of addressing the psychological needs of users and ensuring a satisfying experience to promote the sustained use of AR technologies in online shopping.

Therefore, this study proposes the following hypotheses:

H7. 

User satisfaction has a positive impact on continuance intention.

H8. 

Attitude has a positive impact on continuance intention.

2.8. The Serial Mediation Effect

The serial mediation model, grounded in the S-O-R framework and TCT, provides a structured explanation of AR continuance intention by outlining the sequential influence of key psychological mechanisms. The process begins with external stimuli, such as AR features, which serve as antecedents shaping users’ perceptions and beliefs about the expected outcomes of AR usage. These stimuli influence the organism, which represents users’ cognitive and affective states, including their attitude and satisfaction toward the technology. Second, during the initial adoption phase, outcome expectations derived from these stimuli lead to the formation of positive or negative attitudes, which influence initial acceptance behavior. Finally, as users gain experience, the response stage determines whether AR continuance intention is reinforced or diminished. At this stage, satisfaction plays an essential role in sustaining engagement, while attitude becomes increasingly influential as users transition into long-term usage [47].

Conducting a more in-depth examination of serial mediation, Vieira et al. [83] assert that AR provides significant utilitarian benefits, which subsequently enhance user satisfaction and foster positive attitudes toward its adoption. The utilitarian value of AR, particularly in the context of online shopping, underscores its practical benefits. However, ensuring that AR meets user expectations and fosters satisfaction is even more critical to sustain long-term engagement [84]. Empirical research supports this perspective, demonstrating that AR features (e.g., virtual try-ons) enhance the shopping experience by integrating both functional utility and hedonic enjoyment, ultimately strengthening consumer loyalty and increases continuance intention [85]. Thus, these factors collectively exert an indirect influence on consumer behavior, primarily by shaping users’ sustained engagement and reinforcing their intention to continue using AR technology. While perceived risk has a detrimental impact on affective responses, such as attitudes toward technology and user satisfaction, often resulting in less favorable evaluations of products or services [86]. Furthermore, perceived risk has a significant negative effect on behavioral intentions, such as the intention to purchase or continue to use a technology [13]. Beyond these direct effects, perceived risk also serves as a mediator, shaping the extent to which external elements—such as AR augmentation, interactivity, and engagement—influence behavioral intentions [26]. This highlights the critical need to mitigate perceived risk to improve consumer confidence, increase perceptions of utilitarian value, improve satisfaction, and ultimately drive favorable behavioral outcomes [13].

In conclusion, the existing literature acknowledges that attitude and satisfaction are pivotal determinants of behavioral intentions. Building on these foundational relationships, this study proposes that AR not only enhances utilitarian value but also reduces perceived risk in online shopping contexts. This dual impact leads to more favorable attitudes toward AR technology, as users develop positive perceptions of its usefulness and reliability. Additionally, AR fosters greater satisfaction by aligning online shopping experiences with user expectations, creating a balance between anticipated outcomes and actual performance. Within this framework, AR emerges as a tool that not only addresses practical needs but also elevates the overall shopping experience, further driving consumers’ continuance intention. Thus, consumers are willing to continue to use AR technology if they perceive high benefits (i.e., information and emotional benefits) in the shopping experience [16]. Figure 1 illustrates the serial mediation pathways, beginning with AR’s influence on cognitive responses (i.e., utilitarian value and perceived risk). These cognitive responses subsequently impact affective responses (i.e., user satisfaction and attitude) which, in turn, drive the intention to continue to use AR technology. Therefore, the following serial mediation hypotheses are proposed:

H9. 

Utilitarian value and user satisfaction jointly mediate the indirect effect of AR on continuance intention.

H10. 

Utilitarian value, user satisfaction, and attitude jointly mediate the indirect effect of AR on continuance intention.

H11. 

Perceived risk and user satisfaction jointly mediate the indirect effect of AR on continuance intention.

H12. 

Perceived risk, user satisfaction, and attitude jointly mediate the indirect effect of AR on continuance intention.

H13. 

Perceived risk, utilitarian value, and user satisfaction jointly mediate the indirect effect of AR on continuance intention.

H14. 

Perceived risk, utilitarian value, user satisfaction, and attitude jointly mediate the indirect effect of AR on continuance intention.

3. Materials and Methods

3.1. Research Design

The S-O-R model and TCT were applied in this research to explore how augmented reality (web-based AR) stimulates a serial mediation process, transitioning from consumers’ cognitive responses (utilitarian value and perceived risk) to affective responses (user satisfaction and attitude toward AR technology), ultimately influencing their intention to continue to use AR technology. In light of this, a deductive research approach was adopted, focusing on testing the proposed hypotheses derived through the integration of these two frameworks. Consequently, the study is classified as a descriptive correlational study, designed to accurately depict the key variables related to the problem and analyze the relationships between exogenous and endogenous constructs. Therefore, an online self-completion questionnaire was developed to collect quantitative data, in order to address the research gap and achieve the research objectives.

The measures employed in this study were adapted from previously validated instruments and were meticulously revised to align with the research context. The variable AR was measured using four items from the scale [2]. The cognitive responses (utilitarian value and perceived risk) were measured using the four-item scale developed in [19] and three-item scale from [28], respectively. The affective responses (user satisfaction and attitude toward AR technology) were measured using the three-item scales from [30] and [35], respectively. The variable continuance intention was measured using seven-item scales adopted from [34]. All constructs were measured using a seven-point Likert scale, ranging from 1 (“strongly disagree”) to 7 (“strongly agree”).

3.2. Data Collection

This study used a mixed non-probability sampling approach, combining convenience and snowball sampling techniques to recruit participants. Participants were initially recruited via convenience sampling through social media platforms. Then, snowball sampling was later used to expand the sample, ensuring the broader inclusion of online shopping users from different backgrounds. The selection was limited to individuals who were born between 1981 and 2009 (Gen Y to Gen Z) and had purchased clothes online at least once in the previous year. Respondents were initially directed to the official Ray-Ban website (publicly accessible) to utilize the web-based AR feature, ensuring they had a comprehensive understanding of web-based AR before completing the questionnaire. Next, respondents were guided to choose a pair of glasses from the website and align their smartphone cameras with their faces. Subsequently, the smartphone camera generates an overlay, projecting the glasses onto the virtual version of the user’s face. This allows consumers to visualize themselves wearing the selected glasses, providing a preview experience akin to visiting a physical retail store.

First, a pilot study was carried out, resulting in the collection of 47 valid questionnaires. Based on the respondents’ feedback, the wording of certain questionnaire items underwent adjustments, in order to enhance their clarity and fix any errors. Next, the pilot study results were analyzed to confirm the content validity and reliability, after which the formal survey was conducted. Data were collected in a natural, non-contrived setting over a two-month period (April–May 2024), as a cross-sectional study with a minimal interference from the researcher. The questionnaire consisted of two sections. The first section included measures for six variables: AR, utilitarian value, perceived risk, user satisfaction, attitude toward AR technology, and continuance intention; while the second part included demographic questions such as age, educational level, residential area, and years of shopping experience. As such, gender, age, and shopping experience were controlled in the proposed model shown in Figure 2. Finally, the collected data were analyzed using SmartPLS (4.1.0.3).

According to the Ministry of Communication and Information Technology [87], the estimated population of online shoppers in 2024 was 65 million users. Based on the sampling formula of Krejcie and Morgan [88], the minimum required sample size for this population is 388 respondents. Therefore, a total of 452 valid questionnaires were obtained, following the exclusion of 64 invalid responses due to incomplete answers, as well as the exclusion of respondents who had not recently made online purchases. The respondents included in the study comprised individuals who have made at least one online clothing purchase within the past year. According to the data presented in

Table 1. Descriptive statistics of sample.

Item		Samples	Percentage
Gender	Male	196	43.4%
	Female	256	56.6%
Age	15–17	26	5.8%
	18–23	234	51.8%
	24–30	48	10.6%
	31–36	89	19.7%
	37–43	55	12.1%
Residence	Alexandria	26	5.8%
	Cairo	419	92.7%
	Other	7	1.5 %
Educational Level	Student	190	42%
	Bachelor’s degree	175	38.7%
	Master’s degree	49	10.8%
	Ph.D.	38	8.4%
Shopping Experience	<1 year	82	18.15%
	1–2 years	144	31.85%
	3–4 years	105	23.23%
	>4 years	121	26.77%

, approximately 56.6% of participants identified as female, while around 43.4% identified as male. Respondents aged under 18 years old were approximately 5.8%, 51.8% were aged in the range 18–23, 10.6% falling in the 24–30 age bracket, close to 19.7% were in the 31–36 age range, and 12.1% were aged between 37 and 44 years old. Among those surveyed, 5.8% were living in Alexandria, 92.7% in Cairo, and 1.5% in other cities across Egypt. Concerning the educational level, 8.4% of the respondents were Ph.D. holders, 10.8% had completed a Master’s program, 38.7% had obtained a Bachelor’s degree, and 42% were currently pursuing their studies. Respondents who had made online purchases within the past year were also required to indicate the number of shopping experiences they had, in order to assess their familiarity with online shopping. The results revealed that approximately 18.15% had less than one year of experience, while 31.85% had between one and two years of experience. Additionally, 23.23% had engaged in online shopping for three to four years and, finally, 26.77% had more than four years of experience with online shopping.

3.3. Data Analysis

Data were analyzed using various statistical packages. SPSS version 28 was first utilized for data entry, screening, and cleaning, as well as for conducting the descriptive statistics. Next, SmartPLS 4.1.0.3 was used for advanced statistical data analysis, and PLS-SEM was applied to test the research model. Structural Equation Modeling (SEM) was employed in this study to test the previously stated hypotheses and determine the relationships between independent and dependent variables, including testing for mediating effects in the research model. Unlike simple linear regression, this method examines each mediation pathway separately. Moreover, SEM enables the statistical modeling and testing of complex phenomena, making it a preferred method for confirming or rejecting theoretical models in a quantitative study. As the PLS calculation does not generate formal significance test outcomes for each parameter, a bootstrap approach was employed to derive the t-statistics and standard errors. Bootstrapping analysis method was conducted with 5000 re-samples, in order to test the direct and indirect effects for the purposes of this study.

4. Results

4.1. Validation of the Measurement Model

In the initial phase, the reflective measurement model was assessed using PLS-SEM. The reliability of the constructs was evaluated using both Cronbach’s alpha coefficients and composite reliability (CR), while the validity was assessed through the outer loadings of the indicators and the average variance extracted (AVE). Cronbach’s alpha coefficients for each construct ranged from 0.841 to 0.920, exceeding the standard threshold of 0.7 and demonstrating strong reliability [89]. Similarly, the composite reliability ranged from 0.894 to 0.936, indicating a high level of internal consistency for all constructs within the instrument [90]. Subsequently, the outer loading for each item surpassed the standard threshold of 0.708, as recommended in [90], signifying a noteworthy level of measurement validity for each construct. Additionally, the AVE values were above the minimum threshold of 0.50, as suggested in [90,91]. These results affirm the convergent validity of the construct (see

Table 2. Results of measurement model validation.

Construct	Item	Item Loading	Cronbach’s Alpha	Composite Reliability (CR)	Average Variance Extracted (AVE)
Augmented reality (AR)	AR1	0.841	0.842	0.894	0.679
	AR2	0.820
	AR3	0.861
	AR4	0.771
Utilitarian value (UV)	UV1	0.834	0.848	0.898	0.688
	UV2	0.857
	UV3	0.841
	UV4	0.783
Perceived risk (PR)	PR1	0.834	0.853	0.910	0.771
	PR2	0.920
	PR3	0.877
User satisfaction (SAT)	SAT1	0.884	0.841	0.904	0.759
	SAT2	0.885
	SAT3	0.844
Attitude (ATT)	ATT1	0.873	0.848	0.908	0.767
	ATT2	0.893
	ATT3	0.861
Continuance intention (CI)	CI1	0.744	0.920	0.936	0.677
	CI2	0.786
	CI3	0.783
	CI4	0.853
	CI5	0.858
	CI6	0.866
	CI7	0.858

).

Discriminant validity was subsequently evaluated using the Fornell–Larcker criterion, cross-loadings, and the heterotrait–monotrait (HTMT) ratio of correlations. According to the Fornell–Larcker criterion results presented in

Table 3. Fornell–Larcker criterion results.

	AR	ATT	CI	PR	SAT	UV
AR	0.824
ATT	0.536	0.876
CI	0.585	0.686	0.823
PR	−0.111	−0.156	−0.301	0.878
SAT	0.654	0.749	0.735	−0.236	0.871
UV	0.661	0.639	0.627	−0.141	0.683	0.829

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

, the square root of each construct’s AVE was higher than its correlation with any other construct. This indicates that the diagonal values are greater than the off-diagonal values, as outlined by Fornell and Larcker (1981). The cross-loading results in

Table 4. Cross-loading results.

	AR	ATT	CI	PR	SAT	UV
AR1	0.841	0.444	0.498	−0.076	0.551	0.541
AR2	0.820	0.388	0.458	−0.113	0.483	0.512
AR3	0.861	0.477	0.506	−0.080	0.559	0.570
AR4	0.771	0.454	0.463	−0.099	0.558	0.551
ATT1	0.490	0.873	0.587	−0.116	0.691	0.574
ATT2	0.448	0.893	0.609	−0.138	0.661	0.540
ATT3	0.471	0.861	0.606	−0.159	0.613	0.563
CI1	0.471	0.610	0.747	−0.180	0.608	0.548
CI2	0.465	0.506	0.786	−0.279	0.573	0.478
CI3	0.430	0.521	0.783	−0.256	0.575	0.492
CI4	0.462	0.536	0.853	−0.261	0.599	0.447
CI5	0.506	0.590	0.858	−0.277	0.624	0.547
CI6	0.547	0.615	0.866	−0.221	0.644	0.550
CI7	0.476	0.559	0.858	−0.264	0.603	0.541
PR1	−0.087	−0.109	−0.210	0.834	−0.161	−0.087
PR2	−0.125	−0.161	−0.297	0.920	−0.246	−0.150
PR3	−0.073	−0.133	−0.272	0.877	−0.200	−0.123
SAT1	0.635	0.681	0.687	−0.173	0.884	0.668
SAT2	0.547	0.658	0.643	−0.225	0.885	0.578
SAT3	0.520	0.614	0.586	−0.222	0.844	0.531
UV1	0.581	0.517	0.523	−0.188	0.554	0.834
UV2	0.587	0.560	0.516	−0.067	0.587	0.8587
UV3	0.572	0.526	0.523	−0.085	0.560	0.841
UV4	0.441	0.515	0.522	−0.130	0.567	0.783

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

demonstrate that items loaded significantly on the specific constructs they were intended to measure. To overcome any limitations of cross-loadings and the Fornell–Larcker criterion, as recommended in [90], the study also assessed the heterotrait–monotrait (HTMT) ratio of correlations as proposed in [92]. The HTMT values reported in

Table 5. Heterotrait–monotrait ratio (HTMT) results.

	AR	ATT	CI	PR	SAT	UV
AR
ATT	0.633
CI	0.663	0.775
PR	0.128	0.180	0.335
SAT	0.772	0.884	0.833	0.273
UV	0.778	0.753	0.710	0.161	0.806

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

indicate that all constructs met the ${\mathrm{HTMT}}_{0.90}$ criterion. Therefore, the discriminant validity was ascertained.

4.2. Validating the Structural Model

The primary evaluation stage for the structural model involved examining the variance inflation factor (VIF) across different constructs, in order to verify the absence of lateral collinearity issues. The findings shown in

Table 6. Variance inflation factor (VIF) and effect size ($f^2$) results.

	VIF	${\mathit{f}}^2$
AR → UV	1.012	0.754
AR → PR	1.000	0.012
ATT → CI	2.289	0.097
PR → UV	1.012	0.008
PR → SAT	1.020	0.039
SAT → ATT	1.000	1.274
SAT → CI	2.332	0.248
UV → SAT	1.020	0.839

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

reveal that the VIF values were in accordance with the guidelines recommended in [90], confirming that there were no concerns regarding lateral collinearity in this study. The next stage to assess the structural model included calculation of the coefficient of determination ($R^2$), Effect Size ($f^2$), Predictive Relevance ($Q^2$), and path coefficients. The coefficient of determination $R^2$ signifies the explained variance within endogenous constructs and serves as an indicator of a model’s explanatory power [93]. The $R^2$ ranges from 0 to 1, with greater values suggesting stronger explanatory power. The research model utilized explained 58.1% of the variance in continuance intention, 48.4% of the variance in user satisfaction, and 55.9% of the variance in attitude toward AR technology, indicating a moderate level of in-sample prediction. Subsequently, the effect size ($f^2$) was used to illustrate the strength of the impact of an exogenous variable on the explanation of an endogenous variable in terms of the $R^2$ values. According to [94], values of 0.02, 0.15, and 0.35 represent small, medium, and large effects, respectively. As shown in Table 6, all $f^2$ values had at least a small effect (i.e., greater than 0.02), except for the path from AR to perceived risk, for which there was no effect.

Notably, $R^2$ indicates the in-sample explanatory power only and does not capture the out-of-sample predictive performance of a research model [95,96]. Therefore, [97] introduced the PLSpredict value to explain the predictive power of a study. As recommended in [93], PLSpredict was used with 10 folds and 10 repetitions.

Table 7. Predictive performance of the PLS model.

	${\mathit{Q}}^2$	PLS_SEM_MAE	LM_MAE	PL_SEM − LM_MAE
ATT1	0.200	0.776	0.766	0.010
ATT2	0.177	0.781	0.790	−0.009
ATT3	0.188	0.823	0.802	0.021
CI1	0.178	0.843	0.836	0.007
CI2	0.195	1.131	1.073	0.058
CI3	0.166	1.118	1.084	0.034
CI4	0.206	1.133	1.081	0.052
CI5	0.236	1.118	1.043	0.075
CI6	0.260	1.013	0.940	0.073
CI7	0.205	1.032	0.994	0.038
PR1	0.003	1.551	1.531	0.020
PR2	0.010	1.620	1.611	0.009
PR3	0.000	1.668	1.641	0.027
SAT1	0.347	0.771	0.730	0.041
SAT2	0.276	0.794	0.770	0.024
SAT3	0.249	0.801	0.779	0.022
UV1	0.333	0.771	0.779	−0.008
UV2	0.341	0.724	0.735	−0.011
UV3	0.323	0.742	0.747	−0.005
UV4	0.184	0.756	0.756	0.000

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

shows that all indicators obtained $Q^2$ values higher than zero. Subsequently, a closer analysis of the prediction errors was conducted, in order to identify the relevant statistics. Visual assessment of the prediction errors pointed towards a highly non-symmetric distribution. Thus, the MAE was used as a metric to evaluate the predictive power [93]. The data in Table 7 indicate that LM had a higher MAE than PLS-SEM for minority indicators; therefore, it can be inferred that the model’s predictive power is limited, as stated in [93].

The path coefficient was measured to test the hypotheses through the examination of p-values and t-values.

Table 8. Direct effects analysis results.

Hypotheses	Relationship	Path Coefficient	t-Value	p-Value	Decision
H1	AR → UV	0.653	18.724	0.000	Supported
H2	AR → PR	−0.111	2.120	0.017	Supported
H3	UV → SAT	0.663	18.493	0.000	Supported
H4	PR → SAT	−0.142	3.838	0.000	Supported
H5	PR → UV	−0.068	2.004	0.023	Supported
H6	SAT → ATT	0.749	30.575	0.000	Supported
H7	SAT → CI	0.484	9.661	0.000	Supported
H8	ATT → CI	0.300	5.154	0.000	Supported

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

demonstrates the direct relationship results. Regarding the stimulus factor, AR had a significant effect on utilitarian value ($\beta$ = 0.653, t-value = 18.724, $p<$ 0.001, 95% CI = [0.595, 0.709]) and perceived risk ($\beta$ = −0.111, t-value = 2.120, $p<$ 0.05, 95% CI = [−0.201, −0.029]). The cognitive responses of utilitarian value and perceived risk both had a significant effect on the affective response user satisfaction ($\beta$ = 0.663, t-value = 18.493, $p<$ 0.001, 95% CI = [0.604, 0.720] and $\beta$ = −0.142, t-value = 3.838, $p<$ 0.001, 95% CI = [−0.206, −0.085], respectively). Meanwhile, perceived risk showed a significant negative effect on utilitarian value ($\beta$ = −0.068, t-value = 2.004, $p<$ 0.05, 95% CI = [−0.126, −0.014]). Moreover, the two cognitive responses of user satisfaction and attitude toward AR technology had a significant effect ($\beta$ = 0.749, t-value = 30.575, $p<$ 0.001, 95% CI = [0.708, 0.789]). Consequently, user satisfaction and attitude toward AR technology had a significant effect on continuance intention ($\beta$ = 0.484, t-value = 9.661, $p<$ 0.001, 95% CI = [0.403, 0.567], and $\beta$ = 0.300, t-value = 5.154, $p<$ 0.001, 95% CI = [0.201, 0.393], respectively).

4.2.1. Serial Mediating Effect Analysis

The serial mediating effect analysis signified that, first, the utilitarian value and user satisfaction jointly mediated the indirect path from AR to continuance intention ($\beta$ = 0.210, t-value = 6.915, $p<$ 0.001, 95% CI = [0.162, 0.262]). Meanwhile, the mediation effects of utilitarian value, user satisfaction, and attitude revealed significant indirect effects on the path from AR to continuance intention ($\beta$ = 0.097, t-value = 4.238, $p<$ 0.001, 95% CI = [0.062, 0.136]). Additionally, perceived risk and user satisfaction jointly did not mediate the relationship between AR and continuance intention ($\beta$ = 0.008, t-value = 1.509, $p>$ 0.05, 95% CI = [0.001, 0.018]). Moreover, the mediation effect of perceived risk, user satisfaction, and attitude did not mediate the relationship between AR and continuance intention ($\beta$ = 0.004, t-value = 1.493, $p>$ 0.05, 95% CI = [0.001, 0.006]). Furthermore, perceived risk, utilitarian value, and user satisfaction jointly did not mediate the relationship between AR and continuance intention ($\beta$ = 0.002, t-value = 1.338, $p>$ 0.05, 95% CI = [0.000, 0.006]). Finally, the mediation effect of perceived risk, utilitarian value, user satisfaction, and attitude did not mediate the relationship between AR and continuance intention ($\beta$ = 0.001, t-value = 1.327, $p>$ 0.05, 95% CI = [0.000, 0.003]); see

Table 9. Indirect effects results.

Hypotheses	Relationship	Path Coefficient	t-Value	p-Value	Decision
H9	AR → UV → SAT → CI	0.210	6.915	0.000	Supported
H10	AR → UV → SAT → ATT → CI	0.097	4.238	0.000	Supported
H11	AR → PR → SAT → CI	0.008	1.509	0.066	Not Supported
H12	AR → PR → SAT → ATT → CI	0.004	1.493	0.068	Not Supported
H13	AR → PR → UV → SAT → CI	0.002	1.338	0.091	Not Supported
H14	AR → PR → UV → SAT → ATT → CI	0.001	1.327	0.092	Not Supported

Note: AR, augmented reality; UV, utilitarian value; PR, perceived risk; SAT, user satisfaction; ATT, attitude; CI, continuance intention.

.

4.2.2. Control Variables

With respect to the control variables reported in

Table 10. Analysis of control variables.

	Path Coefficient	t-Value	p-Value
Age → CI	−0.024	0.744	0.228
Gender → CI	−0.068	1.059	0.145
Experience → CI	0.111	3.429	0.000

, the findings indicated that shopping experience exerts a statistically significant influence on continuance intention ($p<$ 0.001), suggesting that individuals with greater prior exposure to online shopping are more likely to sustain their engagement with the platforms of online retailers. Conversely, demographic factors such as age and gender did not demonstrate a statistically significant effect on continuance intention ($p>$ 0.05), implying that consumer retention in online shopping is not inherently influenced by these variables. This suggests that online shopping behaviors are becoming increasingly independent of demographic characteristics, highlighting the growing significance of experiential and psychological factors in shaping consumer decisions in digital commerce. AR technology further reinforces this shift through enhancing the sensory and interactive aspects of online shopping, making the experience more immersive and engaging for users across all demographic groups. As a result, rather than being influenced by age or gender, continuance intention in online shopping is increasingly being driven by the perceived values and effectiveness of AR-enhanced shopping experiences, demonstrating how technology can shape consumer behaviors beyond traditional demographic boundaries.

5. Discussion

With the growth of online shopping since the COVID-19 pandemic, consumers are increasing their online purchasing behaviors and level of interaction with online retailers. The integration of AR into online shopping offers consumers precise information and the ability to visualize diverse products, with the aim of reducing uncertainties regarding product fit and ultimately boosting the online purchasing intentions of consumers and their intention to continue to use AR technology. The purpose of this study is to investigate how consumers’ cognitive and affective responses to web-based AR experiences affect their desire to continue to use AR. The results obtained through examining the proposed model revealed that shopping online using AR evokes customers’ affective and cognitive responses, consequently affecting their continuance intention. The direct effect analysis indicated that AR significantly enhances both utilitarian value and perceived risk. AR enables consumers to access essential product-related information, helping to reduce uncertainties. Consequently, the increase in utilitarian value and decrease in perceived risk positively influence user satisfaction; these findings are aligned with [53]. When consumers’ expectations are met, their satisfaction encourages favorable emotional responses, contributing to their future adoption of the product or service [52]. As a result, user satisfaction positively shapes attitudes towards AR technology, which further leads to the continuance of use of AR technology. These findings are in agreement with those of Daragmeh et al. [48], Kim et al. [53], and Nan et al. [62]. Therefore, the direct relationships posited in Hypotheses 1 to 8 are supported.

The supported serial mediation model revealed that AR stimulates continuance intention through its utilitarian value. Utilitarian value stems from the AR technology itself, while user satisfaction arises from perceiving the web-based AR as useful and facilitating the acquisition of relevant information [23]. Therefore, utilitarian value and user satisfaction partially mediate the relationship between AR and continuance intention, thus supporting H9. Furthermore, user satisfaction, as an affective response stemming from the evaluation of the shopping experience, was found to influence consumer attitudes toward the AR technology, which involves an inclination to behave either in favor of or against technology [98]. Accordingly, the serial mediation effect of utilitarian value, user satisfaction, and attitude toward AR were found to partially mediate the relationship between AR and continuance intention; therefore, H10 is supported. These findings are aligned with [19].

However, while AR can reduce concerns about the authenticity, quality, and fit of a product, the results showed that it did not reduce the perceived risks and, ultimately, did not increase customer satisfaction and did not affect their attitude toward AR. This finding is aligned with [13], in which a sole researcher investigated the impact of perceived risk on affective response in the context of AR. This study examined the impact of perceived risk on purchase intention through attitude toward virtual try-on (VTO) technology in the context of online apparel shopping in China. Therefore, the pathway from AR to continuance intention doubly mediated by perceived risk and user satisfaction was not significant; furthermore, it was also not significant through the pathway involving perceived risk, user satisfaction, and attitude toward AR. This finding suggests that user satisfaction is not affected by perceived risk, indicating that, although AR provides more information about the product, it does not reduce perceived risk as expected. A potential explanation for this outcome is the influence of individual risk tolerance, which varies across consumers. Those with higher risk tolerance may be less concerned about uncertainties associated with AR shopping and more willing to embrace the technology, despite its limitations. Conversely, consumers with lower risk tolerance may perceive AR as a complex and unfamiliar technology, amplifying their concerns about product quality and return policies. This reluctance could persist even when AR offers enhanced product visualization, thereby weakening its potential to alleviate perceived risk and enhance user satisfaction.

Additionally, variations in the design of AR interfaces may have contributed to the insignificant pathway. Poorly designed interfaces with limited interactivity, inaccurate product visualizations, or technical glitches may diminish users’ trust in the technology and its ability to provide reliable product information. Inconsistencies in AR features, such as difficulty in visualizing certain fabric textures or color accuracy issues, can further exacerbate consumer doubts, leading to the maintenance of high levels of perceived risk. These design limitations may prevent users from developing the confidence necessary to perceive AR as a trustworthy shopping tool, thereby diminishing its impacts on satisfaction and continuance intention.

Furthermore, cultural factors may have amplified these concerns. Despite the widespread integration of online shopping into daily Egyptian life, some consumers still perceive it as risky—particularly for apparel, where touching and feeling the product is important for making an accurate purchase decision. Although AR provides more information about product fit, size, and colors, concerns about product quality cannot be alleviated, and, therefore, the utilization of AR may not mitigate product risk and, accordingly, did not increase user satisfaction. Consumers still need to touch and feel items to make accurate purchase decisions. Therefore, it can be concluded that AR did not significantly affect continuance intention through the serial mediation of perceived risk, user satisfaction, and attitude toward AR. Thus, H11 and H12 are not supported.

Finally, the pathway from AR to continuance intention through perceived risk, utilitarian value, user satisfaction, and attitude was also not significant. It is essential to recognize that, although a direct relationship exists between utilitarian value and perceived risk, elevated levels of perceived risk may attenuate the positive impact of utilitarian value on user satisfaction and attitudes. This suggests that, even when AR offers substantial functional advantages, persistent concerns regarding potential risks can undermine the overall user experience and diminish users’ intentions to continue engaging with the technology. Thus, H13 and H14 are not supported.

5.1. Practical Implications

This study presents valuable insights for online retailers, offering several practical implications. First, AR is recommended as an extremely effective marketing strategy, particularly for promoting wearable products, as the utilization of AR by online retailers can foster positive psychological responses. AR can serve as a powerful communication tool in online shopping, reducing uncertainty and enabling consumers to make more confident purchasing decisions. Moreover, online platforms should incorporate AR features and provide comprehensive guidance to retailers on the effective utilization of AR, in order to increase product sales and reduce product return rates.

Second, marketers should emphasize the fun and interactive aspects of AR in their promotional strategies, whether on websites or social media campaigns. This research indicated that AR can significantly elevate consumer satisfaction through reinforcing consumer perceptions of usefulness. Encouraging consumer engagement with AR features is crucial in fulfilling their desire for entertainment, while simultaneously delivering utilitarian benefits. Moreover, online retailers must emphasize the utilitarian benefits of AR experiences, which play a vital role in mitigating product risks and strengthening consumers’ trust in their choices.

Third, retailers and technology developers are strongly encouraged to invest in the continuous improvement of AR applications. Through actively integrating user feedback, they can refine the functionality of these applications and enhance user interactions, creating a more seamless and engaging experience. This iterative approach not only boosts consumer satisfaction but also fosters more positive attitudes toward the technology. As AR applications become more intuitive and user-friendly, the likelihood of consumers making a purchase increases, ultimately driving higher conversion rates and long-term loyalty.

5.2. Theoretical Implications

This study contributes to the literature by integrating the S-O-R model and TCT within a serial mediation framework to examine how consumers’ cognitive and affective responses to web-based AR influence their continuance intention. By conceptualizing AR as the stimulus, utilitarian value and perceived risk as cognitive response, satisfaction and attitude as affective response, and continuance intention as the behavioral response, this study provides a novel perspective on AR adoption and sustained use. The integration offers a more comprehensive framework for understanding both the initial adoption of AR, as reflected in users’ immediate cognitive and affective responses, and its influence on the AR post-adoption behavior, as demonstrated by continuance usage behavior. The findings reveal that users’ continuance use of AR is influenced by both cognitive and affective responses, reinforcing the importance of both rational and emotional factors in technology adoption research. The serial mediation model further clarifies that satisfaction and attitude act as critical psychological mechanisms through which consumers process their AR experiences, ultimately influencing their decision to continue using the technology. Thus, this study emphasizes that technology continuance is not a single-step decision but a dynamic psychological process. However, interestingly, the findings indicate that perceived risk did not influence continuance intention through the serial mediation effect. This suggests that, although AR directly reduced consumers’ perceived risk, it did not subsequently enhance their satisfaction and continuance intention through the serial mediation effect. This implies that AR was not entirely effective in alleviating consumers’ uncertainty to a degree that would lead to increased satisfaction and a stronger intention to continue using the technology. This unexpected result highlights the need for further research to examine the role of perceived risk in the adoption and continued use of AR. In summary, this study enriches existing AR and technology continuance literature by providing empirical evidence of a serial mediation effect, highlighting how consumer responses evolve through interconnected stages. Additionally, the study deepens our understanding of the drivers behind the sustained use of AR technology but also suggests avenues for optimizing user engagement through tailored experiences that balance functional benefits with emotional satisfaction.

5.3. Limitations and Future Work

While this study provided valuable insights for research and practice, its limitations offer opportunities for future exploration. The current study primarily focused on AR features as the primary stimulus, while the incorporation of additional characteristics such as personalization and content customization using generative AI can further enhance the technology’s effectiveness. These enhancements not only increase the utilitarian value of AR by offering tailored experiences but also play a crucial role in mitigating perceived risks, leading to greater user satisfaction and, ultimately, foster positive attitudes toward AR.

Moreover, this research was undertaken in Egypt, where the adoption of AR remains in its early stages and is limited to a small number of companies. As a result, consumers are not yet fully familiar with this emerging technology, making it challenging to evaluate the influence of prior AR exposure on their cognitive and emotional reactions. Nevertheless, incorporating additional control variables, such as prior AR experience, could yield significant insights regarding consumer perceptions and interactions with AR in online shopping environments. Individuals who have experienced AR previously are expected to exhibit increased familiarity, diminished perceived risk, and a more favorable disposition towards AR technologies.

Likewise, future studies could explore the role of trust in greater depth, building on the findings of this study and examining its complex relationships with other moderating and mediating factors. Trust could be investigated across several key dimensions, such as trust in the AR application itself and trust in the online retailer—both of which are critical to reducing consumer uncertainty and fostering confidence in AR-driven online shopping experiences. Additionally, exploring cognitive factors such as ease of use and perceived enjoyment, alongside affective responses such as emotional engagement, could provide a deeper and more holistic perspective on user experiences and adoption of technology. By addressing these areas, future research could offer a more comprehensive understanding of the factors influencing AR adoption and its impacts on consumer behaviors.

In addition, the research specifically focused on self-augmentation through eyewear products. However, future studies could benefit from exploring other self-augmentation categories, such as apparel, beauty products, and accessories. Through expanding the focus to include these diverse categories, the study’s relevance can be significantly strengthened across multiple industries, including fashion, cosmetics, and lifestyle products. Encompassing a broader range of self-augmentation products, future research could yield more generalized insights that apply across industries, providing valuable guidance for businesses in these sectors to better understand consumer motivations, develop targeted marketing strategies, and innovate their product offerings in ways that resonate with diverse consumer needs.

Furthermore, the data were collected in Egypt and within a certain age group using convenience and snowball non-probability sampling techniques, thus reducing the study’s applicability across diverse populations. The findings may also vary across different cultures and demographic groups. Expanding the research to include diverse populations from different countries and cultural backgrounds, while applying random probability sampling techniques, will not only enhance the generalizability of the results but may also provide valuable cross-cultural insights that can make the findings more globally relevant. Moreover, the findings of this research revealed that perceived risk did not significantly mediate the relationship between AR and affective responses such as user satisfaction and attitude toward AR. Future research should address these inconsistent findings through exploring the various dimensions of perceived risk (e.g., financial risk, information risk, performance risk, and return and refund risks) individually. Examining these dimensions separately can provide deeper insights into their unique influences on the behaviors and affective responses of consumers.

As highlighted previously, AR adoption in Egypt remains in its early stages, with actual purchases facilitated by AR largely yet to be observed. As the adoption of AR progresses and the technology matures, the integration of actual behavioral data—such as purchase logs and user engagement statistics—will become increasingly feasible. Future investigations could expand upon this framework by enhancing the research model to explore the correlations between behavioral intentions and actual behaviors, assessing the degree to which self-reported purchase intentions correspond with actual purchasing trends. Additionally, future studies might employ a qualitative approach, rather than a quantitative one, allowing for a deeper exploration of customers’ challenges and issues. Moreover, further investigation may broaden the insights gained from this study to comprehend potential alterations in the attitudes of online customers towards the use of AR technology and their inclination to make online apparel purchases over an extended period.

6. Conclusions

This study underscores the significance of web-based AR experiences for eyewear products in shaping consumers’ affective and cognitive responses, ultimately influencing their behavioral intentions (i.e., continuance intention). By integrating the S-O-R model with TCT, this research provides a comprehensive framework for examining the relationships between cognitive responses (e.g., utilitarian value and perceived risk), affective responses (e.g., satisfaction and attitude), and their influences on continuance intention. This integration addresses the gaps in previous studies by emphasizing the critical role of perceived risk and highlighting the value of AR technologies in reducing uncertainties. The results of this research align with those of previous studies, such as [10,11], which highlight the significant effects of AR in the retail industry. This study builds upon these existing insights by focusing specifically on the eye wear sector, demonstrating how online AR interactions can enhance consumer decision making, satisfaction, and continuance intention. Practically, AR enables retailers to effectively bridge the gap between online and offline shopping by offering immersive and interactive product experiences that boost customer engagement. As AR technology continues to evolve, its potential to transform the retail sector and influence consumer behaviors is considerable, making it a crucial tool for businesses seeking to innovate and meet the demands of modern consumers. This research adds to the growing understanding of AR by addressing key theoretical and practical gaps, emphasizing the importance of adopting new technologies to create engaging and trustworthy shopping experiences that encourage long-term customer commitment.