Humanoid social robots (HSRs) are an innovative technology revitalizing various service sectors, such as the hospitality industry. However, limited research has explored how anthropomorphic features of HSRs influence user satisfaction with the services delivered by HSRs. To address this, a research model was proposed to evaluate how three distinct anthropomorphic features: appearance, voice and response, impact the perceived values (i.e. utilitarian, social and hedonic values) of HSRs, which, in turn, influence user satisfaction.
Data from an online survey of hotel customers was utilized to test the research model (N = 509).
The results indicated that appearance, voice, and response affect perceived utilitarian, hedonic and social values differently. The response feature of HSRs demonstrated the strongest impact on perceived utilitarian, social and hedonic values. In addition, voice affected all three perceived values, while appearance only affected perceived utilitarian and social values. Furthermore, perceived utilitarian, hedonic and social values showed positive impacts on user satisfaction, with hedonic value being the most influential factor.
This study contributes to the literature on HSRs and anthropomorphism by explaining how different anthropomorphic features affect users’ value perceptions and user satisfaction with HSR services by utilizing the stimulus-organism-response (SOR) framework.
1. Introduction
Physical and autonomous agents that are capable of operating based on assigned roles in socially appropriate ways during interactions within dynamic and complex social settings are typically introduced as humanoid social robots (HSRs) (Duffy, 2003). Distinguished by their advanced cognitive capabilities, HSRs function as social actors or companions across various service contexts, distinguishing them from traditional information systems (IS) (Song and Kim, 2022). As a result, HSRs have been increasingly deployed in diverse service industries, including education (Conti et al., 2017; Sisman et al., 2024), healthcare (Kwon et al., 2018), retail (Chuah and Yu, 2021; Song and Kim, 2022), and hospitality (Liao and Huang, 2024; Murphy et al., 2019; So et al., 2024). Among these, the hospitality industry has been a highlight in implementing HSRs into their daily operations, aiming to reduce operating costs by using robots for repetitive or monotonous tasks and to enhance customer experiences through the provision of novel and personalized interactions with guests (De Kervenoael et al., 2020). Notably, HSRs have been employed in various hotel services, such as butlers (e.g. Botlr), concierges (e.g. Connie), bellboys (e.g. Sacarino), and waitresses (e.g. Theresa) (Ozturk et al., 2023). In addition, the necessity for contactless services and service delivery due to the outbreak of COVID-19 boosted HSRs’ adoption in hospitality services (Song et al., 2024).
Recent research has highlighted the importance of anthropomorphism in HSRs, emphasizing the need to consider these features when investigating user reactions to HSRs (Klüber and Onnasch, 2022; Li et al., 2023). Previous studies have shown that anthropomorphic features influence user perceptions, attitudes, and behaviors concerning HSRs by either measuring a general perception of robot anthropomorphism (e.g. Moran et al., 2015) or investigating a specific anthropomorphic feature of HSRs, such as appearance (Appel et al., 2020; Prakash and Rogers, 2015), voice (Fan et al., 2016; Konya-Baumbach et al., 2023; Tamagawa et al., 2011), or response (Diederich et al., 2021; Okanda and Taniguchi, 2022). However, these studies have typically focused on a single anthropomorphic feature independently, leaving a gap in understanding how different anthropomorphic features might collectively or differentially impact user attitudes and behaviors. These effects could provide important guidelines for HSR design from the anthropomorphic design perspective. Thus, in-depth research is required to investigate the varying roles of different anthropomorphic features in triggering user perceptions, attitudes, and behaviors regarding HSRs.
Additionally, user satisfaction has been suggested to be a critical factor influencing various user behaviors related to technology-supported services, such as loyalty (Soliman et al., 2024), purchase intention (Jia et al., 2021), continuous usage intention (Lee et al., 2021a), and the long-term success of information systems (IS) implementation (Vaezi et al., 2019). Examining user satisfaction with HSR services could provide an understanding of users’ various behaviors regarding HSRs as well as the long-term success of HSRs in different contexts. However, previous studies on social robots have mainly examined how anthropomorphic features affect users’ continuous behavior intentions (Lee et al., 2021a; Moussawi and Koufaris, 2019), while research on how anthropomorphic features affect user satisfaction with HSR services remains limited.
Furthermore, previous studies have highlighted the utilitarian, social, and hedonic values of IS as antecedents that affect user satisfaction with IS (Evelina et al., 2020; Konya-Baumbach et al., 2023). HSRs could be regarded as multi-purpose IS that provide utilitarian, hedonic, and social values for users. While interacting with HSRs, users could perceive different types of IS values. For instance, users may perceive HSRs’ utilitarian value when the robots help them complete tasks and achieve certain goals. Users may also perceive social value when the robots interact with them while listening to and responding to inquiries. The humorous faces and charming voices of HSRs during communication can make users feel joyful, eliciting hedonic value. Further, users could perceive hedonic value when they feel enjoyment and happiness during their interactions with HSRs. However, limited research has attempted to uncover how HSRs’ various anthropomorphic features affect the perceived IS values of these robots. Thus, it is unclear whether the appearance, voice, and response of HSRs have similar or different influences on user perceptions of the utilitarian, social, and hedonic values of HSRs, which could further affect user satisfaction with the robot-delivered services.
To address these research gaps, this study applies the stimulus-organism-response theory (SOR) (Mehrabian and Russell, 1974) to develop a research model that explores how the three anthropomorphic features of HSRs—appearance, voice, and response—influence user perceptions of the utilitarian, social, and hedonic values of HSRs, which in turn affect user satisfaction. This study employs a quantitative approach, utilizing survey data from 509 hotel customers who had prior interactions with HSRs during their stays in hotels. This targeted sample allows for a focused examination of the hospitality industry’s use of HSRs.
By adopting the SOR theory, the current study addresses a critical gap in HSR and anthropomorphism literature through the exploration of the distinct effects of anthropomorphic features of HSRs, such as appearance, voice, and response on user satisfaction via perceived utilitarian, social, and hedonic values, which has been often overlooked by the existing research on social robots (Diederich et al., 2021; Klüber and Onnasch, 2022; Seo, 2022). In doing so, this study broadens the scope of investigation in the field of social robotics and anthropomorphism, offering a comprehensive understanding of how different anthropomorphic features of HSRs influence user perceptions, with the response feature of HSRs emerging as the most impactful. In addition, this study deepens the understanding of value perceptions within the SOR framework by considering the perceived IS values as the organism, particularly in the context of HSRs, emphasizing the critical role of perceived hedonic value in driving user satisfaction. Further, in contrast to prior research that considered traditional technological features of IS as stimuli (Cheng et al., 2022; Hlee et al., 2023), the findings of the current study emphasize the importance of recognizing anthropomorphic features of HSRs as essential stimuli influencing users’ attitudinal responses. This extends the application of the SOR framework to the context of HSRs, further reinforcing the theory’s relevance in explaining human-robot interactions.
By uncovering the distinct roles of HSRs’ appearance, voice, and response in shaping user perceptions of IS values and, subsequently, user satisfaction, this research offers valuable guidelines for the HSR designs targeting service settings, specifically focusing on the hospitality sector. Further, the study serves as a guide to HSR developers by emphasizing the importance of optimizing the response feature to enhance user satisfaction while also noting that voice articulation should be refined for better human-robot interactions (HRIs). For hotel managers, the findings highlight the need to align HSR features with their intended purpose and to ensure high-quality responses to meet customer expectations, thereby boosting satisfaction and service quality. Thus, the current study contributes to the broader discourse on humanlike technologies, providing empirical evidence that supports a deeper understanding of how anthropomorphism in technologies such as HSRs can drive user value perceptions and satisfaction.
2. Theoretical background
2.1 HSRs’ anthropomorphic features
The anthropomorphism of robots is portrayed as the tendency of the “attribution of humanlike characteristics to nonhuman agents or objects,” such as robots (Seeger et al., 2021, p. 932). Researchers have investigated the anthropomorphism of various social robots, including chatbots (virtual and physical), voice assistants (Seeger et al., 2021), personal intelligent agents, and HSRs (Moussawi et al., 2022). Anthropomorphic or humanlike agents exhibit traits such as emotionality, desire, and warmth, which are conveyed through anthropomorphic design features (Fan et al., 2016). Previous research has explored various anthropomorphic features of social robots, including their appearance (Appel et al., 2020; Prakash and Rogers, 2015), voice (Fan et al., 2016; Konya-Baumbach et al., 2023; Tamagawa et al., 2011), and response (Diederich et al., 2021; Schuetzler et al., 2020).
Social robots’ diverse anthropomorphic features could lead to varying perceptions of their anthropomorphism and elicit conscious and unconscious responses from users (Kim and Sundar, 2012). Previous studies have categorized virtual robots’ anthropomorphic features as verbal or non-verbal, following the communication literature, and have included human identity as an additional feature of social robots (Seeger et al., 2021). In addition, Feine et al. (2019) claimed that social robots’ anthropomorphic features provide invisible, auditory, verbal, and visual social cues for virtual robots. Robots’ verbal and non-verbal features could induce humanlike communication behavior between robots and humans (Seeger et al., 2021). Their verbal features include spoken language, verbal styles, and social dialogues (Rossi et al., 2018). Most HSRs can use speech recognition to understand and respond to users during HRIs, including language, voice, and responses. Non-verbal features include facial expressions, gaze, and body language (van Pinxteren et al., 2019). These non-verbal cues, such as facial expressions and gaze, could also be related to HSRs’ appearance (Li et al., 2023).
Scholars have attempted to examine how robots’ anthropomorphic features affect user perceptions, attitudes, and behaviors. Table 1 summarizes existing literature focusing on anthropomorphic features. Accordingly, in prior literature, significant attention has been given to the appearance, voice, and response features, and these features appear to substantially affect user perceptions and behaviors toward social robots. Prior studies have mainly focused on anthropomorphic features either generally (Kamide and Arai, 2017; Seeger et al., 2021) or specifically, i.e. focusing on appearance (Liu et al., 2022), voice (Fan et al., 2016), or response (Schuetzler et al., 2020). To date, only a few studies have endeavored to explore how various HSRs’ anthropomorphic features influence user perceptions (Fan et al., 2016; Liu et al., 2022), and these studies have focused primarily on appearance and voice features, ignoring the other important features, such as responses, and the combined effects of these features on user perceptions and behaviors. In addition, previous studies on anthropomorphism have examined the direct influence of anthropomorphic features on attitudes or behaviors (Liu et al., 2022; Moussawi et al., 2022). However, the existing studies fail to describe the complex cognitive processes that link anthropomorphic features with user attitudes or behaviors in the context of HSRs. Therefore, the current study focuses on HSRs’ appearance, voice, and response as anthropomorphic features to examine their diverse roles in affecting user satisfaction.
2.2 The stimulus-organism-response (SOR) framework
SOR theory assumes one’s behaviors to be learning responses to external stimuli (Jacoby, 2002; Mehrabian and Russell, 1974). This theory implies that an individual’s intrinsic condition is triggered when exposed to an external stimulus (Hsiao and Tang, 2021). A stimulus (S) is the external component that changes an individual’s inner state, while an organism (O) is an individual’s experience of affective cognition, which involves both cognitive and affective states. Specifically, the cognitive state is the thought process based on information processing, whereas the affective state is concerned with feelings or emotions (Benlian, 2015). A response (R) refers to an individual’s ultimate natural reactions to a particular stimulus, including attitudinal and behavioral reactions (Hsiao and Tang, 2021; Xie et al., 2023).
In IS domains, the SOR framework has been utilized to discuss individuals’ attitudinal and behavioral responses to different IS in diverse contexts, such as social commerce (Fu et al., 2018; Zhang et al., 2014) and e-commerce (Cheng et al., 2022). Table 2 summarizes existing IS studies that use the SOR framework. The table illustrates that the features and characteristics of IS have been widely used to capture stimuli factors, while cognitive reactions, such as various user perceptions, have commonly been used to represent organism factors. Prior literature shows that the SOR framework uncovers the intrinsic process of user responses and is more suitable for including various types of stimuli based on the studied context (Neves et al., 2024). In addition, prior IS studies have considered the technological attributes or characteristics of IS as the stimuli that trigger individuals’ perceptions or behaviors regarding IS (Cheng et al., 2022; Hlee et al., 2023; Ochmann et al., 2024; Thapa et al., 2024; Wang et al., 2024). For example, Zhang et al. (2014) used the SOR framework to study individuals’ intentions to engage in social commerce. They found that the technological elements of social commerce platforms, such as users’ perception of their interactivity, personalization, and sociability, serve as stimuli to trigger their perceptions of virtual customer experiences in social commerce (an organism) concerning flow, social support, and social presence, which further drive their intentions to participate in social commerce (a response). Visual appeal, interpersonal influence, and portability of mobile commerce platforms were reported to serve as stimuli to trigger user perception of hedonic and utilitarian browsing (organism), which leads to their impulsive purchases (organism) (Zheng et al., 2019). Furthermore, Cheng et al. (2022) used the SOR framework to study text-based chatbot usage among individuals and discovered that the perception of warmth, communication delay, and competence of a text-based chatbot served as stimuli to trigger users’ trust in chatbots (an organism).
The widespread application of the SOR framework to explore users’ attitudes and behavioral responses to different IS indicates that this framework could be an appropriate theoretical framework to study user satisfaction with robot services, particularly in relation to robots’ anthropomorphic features. In addition, the uncanny valley theory posits that robots’ anthropomorphic features can elicit spontaneous responses from users (Mori, 1970). This further supports the appropriateness of applying the SOR framework to study user satisfaction with robot services through the lens of anthropomorphic features. Unlike some studies that explore the direct relationships between anthropomorphism and user attitudes or behaviors (Seo, 2022), the SOR framework enables a deeper analysis of how anthropomorphic features influence user satisfaction by capturing the underlying cognitive mechanism through its organism component. This makes the SOR framework a well-suited theoretical stance to understand user satisfaction with HSR services.
Previous studies show that the SOR framework can effectively uncover the intrinsic process of user responses, allowing the inclusion of various types of stimuli tailored to the specific context being studied (Neves et al., 2024). In IS research, the technological attributes or characteristics of IS have often been considered stimuli that trigger individuals’ perceptions regarding IS (Cheng et al., 2022; Hlee et al., 2023; Ochmann et al., 2024; Thapa et al., 2024; Wang et al., 2024). Recent research on robots has also applied the SOR framework and set anthropomorphic features of chatbots as stimuli to understand user trust, switching intentions, and user attitudes (Cheng et al., 2022; Hlee et al., 2023). These studies provide strong support for the current research, validating the consideration of anthropomorphic features of robots as stimuli, given that these features are technological attributes of humanlike robots. Additionally, cognitive processes have often been chosen and characterized as organisms in the SOR framework (Wang et al., 2024; Zhu et al., 2023). In the context of HSRs, user perception of IS values can represent the cognitive processes of users, making them suitable as the organism in accordance with the SOR framework. User satisfaction, which refers to users’ attitudinal response to robot service, has also been categorized as a response in the SOR framework in previous research (Ochmann et al., 2024), further supporting its inclusion in our study. Therefore, in alignment with the SOR framework, HSRs’ anthropomorphic features as stimuli can trigger users’ value perceptions (organism), which further affect user satisfaction (attitudinal response) toward HSRs.
2.3 User satisfaction
User satisfaction with an IS has been a dominant research topic in IS research since satisfaction has been identified as one of the most critical and widely used indicators to measure the success of an IS (Vaezi et al., 2016, 2019). Some IS scholars have suggested that satisfaction constitutes the attitudes of users or the collection of feelings developed to assess the effectiveness or success of an IS (Petter et al., 2013), highlighting that user satisfaction with an IS pertains to user attitudes and feelings stemming from perceived benefits that an individual anticipates during interactions with an IS (Kuo and Hsu, 2022).
Previous studies have examined the factors determining user satisfaction with IS from different perspectives. For example, Cao et al. (2018) considered trust to be an antecedent in satisfaction, and continuance intention to be an outcome of satisfaction from a trust transfer perspective. Bhattacherjee (2001) posited that the perceived usefulness of an IS and users’ extent of confirmation are associated with user satisfaction with IS use, and their satisfaction could lead to users’ continuance intention to use such IS. Later, this model was widely applied to examine user satisfaction and continuance intention in IS research. Some scholars have examined user satisfaction with IS from the IS attribution view. For example, DeLone and McLean (2003) stated that attributes such as information quality, system quality, and service quality of an IS are associated with user satisfaction and have also been extensively utilized in the IS domain to investigate user satisfaction and IS success. Meanwhile, previous research has suggested other attributes such as timeliness, perceived ease of use, user attitude, convenience, perceived usefulness, and accuracy to serve as antecedents of user satisfaction based on their research contexts (Vaezi et al., 2016).
The critical role of user satisfaction in measuring the success of an IS has also drawn scholars’ attention to user satisfaction with HSRs. For example, Song and Kim (2022) categorized user satisfaction as an emotion and experimented on user satisfaction together with irritation, anger, and happiness in HSR use. Similarly, Kwon et al. (2018) found that a lower difference in empathy between a user and HSR yields higher user satisfaction with service robots in the healthcare context. Additionally, Xie et al. (2023) described how anthropomorphic cues are related to interaction satisfaction by arguing that satisfaction is a reaction attitude toward smart home assistants. Though anthropomorphic features of HSRs have been argued to be important in understanding user perception, attitude, behavior regarding HSRs, and the success of HSR implementation, only a few studies have attempted to elaborate on how the anthropomorphic features of HSRs influence value perceptions associated with using HSRs and how these perceived values further affect user satisfaction with HSR services.
3. Research model and hypotheses
Drawing on the SOR framework, we developed a research model (see Figure 1) to examine the varied effects of different anthropomorphic features on perceived utilitarian, hedonic, and social values, which, in turn, determine user satisfaction with HSR services. In the context of social robots, prior studies have highlighted appearance (Appel et al., 2020; Prakash and Rogers, 2015), voice (Fan et al., 2016; Konya-Baumbach et al., 2023; Tamagawa et al., 2011), and response (Diederich et al., 2021; Okanda and Taniguchi, 2022) features as prominent anthropomorphic features. HSRs’ appearance refers to how HSRs look and mimic humanlike physical appearance, including the face, body, legs, and dynamic characteristics such as facial expressions (Appel et al., 2020; Klüber and Onnasch, 2022; Liu et al., 2022; Prakash and Rogers, 2015), and personality features such as gender (Seo, 2022). These elements elicit non-verbal cues during HRIs. Further, robots’ appearance has been reported to impact user perceptions regarding robots (Castro-González et al., 2016). HSR’s voice is also a key design element in communication, providing verbal cues during HRIs. Additionally, the capability to communicate through voice is a key determinant when assessing user preferences for robots, and a robot’s voice has been found to be a critical factor in user engagement (Klüber and Onnasch, 2022). Moreover, responses of HSRs can be either verbal or non-verbal in nature based on their interaction style, environment, or robot type (Złotowski et al., 2015). HSRs could possess affective and cognitive capabilities and combine emotional and rational thinking, thereby understanding users’ requirements (Čaić et al., 2019). When HSRs’ responses to users resemble the responses that occur during human–human interactions, users may not discriminate between HSRs and humans (Okanda and Taniguchi, 2022), and these interactions produce positive perceptions regarding HSRs. However, HSRs’ failure in responses could lower users’ expectations concerning HSRs’ human likeness, leading to uncertainties concerning HSRs use and user dissatisfaction with robot services (Diederich et al., 2021; Fan et al., 2016). Accordingly, appearance, voice, and response features of HSRs are selected in this study.
Previous studies have considered perceived utilitarian, social, and hedonic values of IS as dominant motivators for IS use (Evelina et al., 2020; Stock et al., 2015). Therefore, the current study conceptualizes these three types of perceived IS values for HSRs as the organism in the SOR framework, which are triggered by HSRs’ three anthropomorphic features and subsequently influence user satisfaction. Specifically, HSRs’ appearance, voice, and response are characterized as external stimuli (S) that trigger users’ perceived utilitarian, social, and hedonic values (O). These perceived values, in turn, affect users’ attitudinal response (R), such as user satisfaction with HSR services (Ochmann et al., 2024). Prior research has suggested that users’ age, gender, and previous robot use experience affect their perceptions of using IS, such as social robots (Kang et al., 2023). Thus, we set users’ age, gender, and previous experience with robot services as control variables.
Humanlike designs have been found to transfer human characteristics to HSRs and influence user perceptions (Duffy, 2003). When users see the humanlike appearance of HSRs, such as physical appeal and facial expressions, they might perceive HSRs as similar to humans and assume HSRs to be capable of performing tasks like human employees. Therefore, users are likely to infer humanlike service quality from HSRs when HSRs have humanlike congruence, thus influencing their perceptions of the utilitarian value of HSRs. Similarly, when HSRs look like humans, users tend to interact with HSRs in a manner similar to how they interact with humans and carry out social interactions like human-human interactions. HSRs with humanlike faces can be perceived to be sociable (Broadbent et al., 2013). Facial expressions and maintaining eye contact are elements of human-human communication. The facial expressions of HSRs and eye contact with humans during HRIs can increase user engagement with HSRs, similar to human-human interactions. Therefore, the humanlike appearance of HSRs could strengthen user perceptions of the social value of HSRs. Furthermore, the humanlike appearance of HSRs has been found to increase users’ positive emotions, such as making users feel enjoyment and pleasure in the frontline service context (Belanche et al., 2020). HSRs with a humanlike appearance could evoke users’ curiosity and enjoyment while performing their dedicated tasks (Chi et al., 2023; De Graaf and Ben Allouch, 2013; Liu et al., 2022). Users’ interest and curiosity during interactions with HSRs can further make users feel a sense of joy. Thus, the humanlike appearance of HSRs could trigger user perceptions of the hedonic value of HSRs. Overall, HSRs’ humanlike appearance can be reasonably assumed to trigger users’ cognitive responses, including their perception of the IS values related to HSRs, and it is plausible to assume the following hypotheses:
HSRs’ appearance is positively linked to user perceptions of robots’ utilitarian value.
HSRs’ appearance is positively linked to user perceptions of robots’ social value.
HSRs’ appearance is positively linked to user perceptions of robots’ hedonic value.
HSRs can communicate with users through voice-based speech. As an expressive communication medium, voice can reflect speakers’ features, such as age, gender, emotion, and personality (Li et al., 2023). Compared to a mechanical or technical voice, an HSR’s humanlike voice can convey richer implicit information (e.g. emotion and personality), facilitating social expression and interactions between users and robots (Moran et al., 2015). For instance, the humanlike voices of HSRs, which are soft, faintly accented, and lively, could make human-robot communication more similar to human-human communication than machinelike voices, thereby making users perceive such HSRs as having stronger humanlike cognitive capability to communicate with humans and perform tasks following their requests. Thus, the humanlike voice of an HSR could enhance user perceptions of the utilitarian value. In addition, the humanlike voice of an HSR can make users feel that they are talking with a human, perceiving the social interactions with HSRs as similar to human-human interactions and fulfilling social experiences during HRIs (Conti et al., 2017). When an HSR possesses a lively voice without monotonous articulation, users are likely to feel connected and develop trust and rapport during HRIs (Li et al., 2023). The openness and sense of comfort in communication with a robot with a soft and lively voice could even make users perceive an HSR as a social companion, unlike a machine. Therefore, the humanlike voice of HSRs may improve user perception of the social value. Additionally, a humanlike voice is deemed more pleasurable than a synthesized voice (Kühne et al., 2020; Schreibelmayr and Mara, 2022). Also, Li et al. (2023) argued that the soft and lively voices of robots during HRIs can make users feel more engaged and enjoyable. Based on the above grounds, the HSRs’ voice can be assumed to improve the perceived utilitarian, hedonic, and social values. The more humanlike the voice of an HSR, the stronger the perception of the IS values related to HSRs. Therefore, the following hypotheses are proposed:
HSRs’ voice is positively linked to user perceptions of robots’ utilitarian value.
HSRs’ voice is positively linked to user perceptions of robots’ social value.
HSRs’ voice is positively linked to user perceptions of robots’ hedonic value.
The response feature is another key anthropomorphic feature of an HSR. HSRs usually interact with users via verbal or non-verbal responses (Złotowski et al., 2015). Effective and timely responses are required when performing specific tasks assigned to HSRs, and users might not perceive such HSRs to be useful or to have good utilitarian value if the robots fail to provide meaningful, prompt responses to users or engage with users in interactive dialogues (Fan et al., 2016). Prior research has indicated that robot responses can affect user perceptions regarding robots (Belanche et al., 2020; Birnbaum et al., 2016). Specifically, a more responsive robot is perceived as more competent and sociable than a less responsive robot (Birnbaum et al., 2016). Chatbots with higher-level response skills, such as offering tailored and varied responses, have been reported to be more anthropomorphic and socially present than chatbots with less-skilled response features (Schuetzler et al., 2020). In the context of HSRs, the appropriate responses delivered promptly by HSRs can enable smooth engagement between the robots and users and mimic human-human interactions, which further develop the users’ belief that HSRs are competent in completing tasks and sociable enough to interact with, enhancing user perceptions of the utilitarian and social values (Song and Kim, 2022). Similarly, meaningful and real-time interactive dialogues are likely to make users feel that HSRs are vivid and realistic, making their experience more pleasant (Okanda and Taniguchi, 2022). HSRs’ responses to users, such as gestures and physical movements carried out according to users’ requests, could trigger users’ perceived enjoyment and happiness, similar to the enjoyment of playing games. Therefore, it is reasonable to assume that HSRs’ responses to users during HRIs would affect users’ perceived utilitarian, social, and hedonic values; thus, the more prompt, appropriate, and real-time responses HSRs provide, the stronger the perception of the HSRs’ IS values. Therefore, it is reasonable to suggest the following hypotheses:
HSRs’ response is positively linked to user perceptions of robots’ utilitarian value.
HSRs’ response is positively linked to user perceptions of robots’ social value.
HSRs’ response is positively linked to user perceptions of robots’ hedonic value.
HSRs’ utilitarian value, such as their adaptability and usefulness, influences user satisfaction with robots (Bhattacherjee, 2001; Moussawi et al., 2022). Users tend to be satisfied if their expectations are fulfilled during technology use (De Graaf et al., 2015). When users perceive HSRs to have high utilitarian value, i.e. the robots are effective in fulfilling their intended functions and tasks, users could perceive the robots to be useful in service delivery, leading to satisfaction. Additionally, better task performance of HSRs can meet users’ expectations for using HSRs, affecting user satisfaction with the service delivered by HSRs. HSRs with higher utilitarian capabilities can perform tasks with high levels of competency and reliability, such as efficiently handling queries, which are more likely to fulfill users’ expectations and make them more satisfied with the service delivered by HSRs (Chen and Girish, 2023). Therefore, we argue that the higher the perception of the utilitarian value of HSRs, the higher the satisfaction among users, and propose the following hypothesis:
User perception of the utilitarian value of HSRs positively affects their satisfaction with HSR services.
HSRs are identified diversely, for instance, as social actors, social companions, or social partners, depending on the environments in which they operate and the nature of social interactions they perform during HRIs. The social value of HSRs entices users to perceive robots as sufficiently humanlike (Pelau et al., 2021). When users perceive that HSRs can engage in communication that closely mimics human-to-human interaction—through language, understanding, and responsive behavior—they are more likely to feel a sense of connection and relational bonding with the robots, similar to their connections with other humans (Fan et al., 2016). In addition, when users perceive an enhanced social value, they could feel that the HSRs can fulfill their expectations for social interaction, replicating human-human interactions and making them highly satisfied with robot services. Further, prior literature has shown that when users perceive a higher social value of a technology, they are more likely to be satisfied with their usage (Evelina et al., 2020; Huang et al., 2021). Accordingly, it can be inferred that increased user perceptions of the social value of HSRs could improve user satisfaction with the service delivered by the robots; thus, the following hypothesis is suggested:
User perception of the social value of HSRs positively affects user satisfaction with HSR services.
The unique anthropomorphic features of HSRs could make users experience enjoyable interactions with HSRs (Kang et al., 2023; Kim and Sundar, 2012). Prior research has found that users are more satisfied with robots when they perceive fun and joy during their use and engage in interactions with robots (Huang et al., 2021). For instance, in healthcare settings, users perceive the high social value of social robots since social robots offer entertainment and emotional support to patients, thereby enhancing their overall experience and making them satisfied with robot usage. Studies in the hospitality industry have also shown that users’ emotional experiences, including fun and surprise, generated when using robots in hotels can increase their satisfaction with robotic services (Luo et al., 2021). Therefore, it is reasonable to assume that when users perceive a higher hedonic value of HSRs, they are likely to have higher satisfaction with the service delivered by HSRs. Thus, the following hypothesis is proposed:
User perception of the hedonic value of HSRs positively affects user satisfaction with HSR services.
4. Research methodology
4.1 Construct measurements
Items that had been validated by prior studies were adapted to assess all constructs included in this study (the measurement items are presented in the Appendix). Slight modifications were applied to the measurement items of each construct to fit the context of HSR usage. The items for appearance, response, perceived utilitarian value, and perceived social value were evaluated using a seven-point Likert scale, which ranged from 1 = strongly disagree to 7 = strongly agree, while the items for voice, perceived hedonic value, and user satisfaction were assessed using a similar seven-point semantic differential scale. Appearance was measured with items adapted from the work of Liu et al. (2022), voice with items adapted from the work of Stern et al. (2002), and response with items adapted from the work of Lankton et al. (2015). Measurement items from the work of Im and Workman (2004) were adapted to assess perceived utilitarian value. A combination of items gathered and modified from Pelau et al. (2021) and Zhang et al. (2021) was used to measure perceived social value, while items adapted from van der Heijden (2004) were utilized to measure perceived hedonic value. User satisfaction was evaluated in accordance with the items proposed by Bhattacherjee (2001).
4.2 Data collection
An online survey, developed in English, was conducted to gather data from hotel customers since the hotel industry has pioneered the deployment of HSRs to serve customers (Chi et al., 2023). Prior to the main study, a pilot study was carried out with 12 respondents to collect feedback on the research instruments. The measurements were modified to improve their clarity and accuracy based on the pilot study’s feedback. The survey was designed using Qualtrics and launched for data collection on Amazon Mechanical Turk (AMT) from 6–15th September 2022. AMT, a crowdsourcing platform, has been extensively used in research areas such as business management and psychology to enlist participants for research. It is recognized as an appropriate source of research data since the AMT user population is reflective of the general public and responds to experimental conditions in a manner comparable to traditional participant groups (Schuetzler et al., 2020; Steelman et al., 2014).
To ensure the quality of the responses, a filtering question: “Have you ever used humanoid social robots in hotels?” was included at the beginning of the survey, aiming to exclude participants who had not experienced HSRs. Then, the respondents were inquired about the types of HSRs they had experienced in hotels to confirm their previous use. Finally, questions regarding respondents’ demographic background, usage experience, and perceptions regarding HSRs were presented. Overall, the survey included three attention checks. Each research participant who completed the survey was provided with monetary compensation ($1.50) upon completion, which took about 10 min.
In total, we received 567 responses. Afterward, 58 responses with similar answers to all questions were excluded. In accordance with Armstrong and Overton (1977), an analysis was performed to compare the initial 25% and last 25% of survey participants in terms of their age and gender to evaluate nonresponse bias. The results revealed no significant differences, indicating that nonresponse bias was not a concern in this study. Finally, 509 responses were used as the study’s valid data sample. Gender, age, travel frequency, and frequency of robot use among the survey respondents are presented in Table 3. Accordingly, the majority of respondents were between 30 and 39 years old (38.5%), and respondents were almost equally distributed by gender. Additionally, most respondents had traveled 1–2 times per year (52.7%) and experienced robots 2–3 times (59.7%) during their stays.
4.3 Validity and reliability
We tested for common method bias (CMB) by employing Harman’s single-factor test, as proposed by Podsakoff et al. (2003). The highest variance was 37.974%, which accounted for less than 50% of the total variance, indicating that CMB was not a critical issue in the current study. In addition, the marker variable (MV) approach was used to test for the presence of CMB (Podsakoff et al., 2003). Symbolic patriotism (Parker, 2010) was selected as the MV because the variable is unrelated to the variables under investigation and the context of this study. The bootstrapping process did not reveal any alterations in the path coefficients in the presence of the MV, and the R2 values showed minor variations: ΔR2 (perceived utilitarian value) = +0.064, ΔR2 (perceived social value) = +0.076, ΔR2 (perceived hedonic value) = +0.050, and ΔR2 (user satisfaction) = +0.002, verifying that the constructs were unaffected by CMB (Tehseen et al., 2017). The variance inflation factor (VIF) values were also tested to examine multicollinearity. The VIF values ranged from 1.311 to 1.642, falling below the threshold of 3.300 (Kock and Lynn, 2012), indicating the absence of multicollinearity issues.
Smart PLS 4.0 was used to assess the measurement model. The convergent validity of the model is established when the standard threshold values of factor loadings are greater than 0.700, the composite reliability (CR) of constructs exceeds 0.700, and the average variance extracted (AVE) surpasses 0.500 (Hulland, 1999). Some items were removed from the constructs due to issues with construct validity and factor loadings. Subsequently, the factor loadings of the measurement items surpassed the threshold of 0.700, while the CR and AVE values were also above 0.700 and 0.500, respectively (see Table 4), confirming the convergent validity of the current study. Further, Cronbach’s alpha values above 0.7 (Henseler et al., 2016) indicated that the constructs exhibit a satisfactory level of internal consistency reliability.
Both the Fornell–Larcker criterion (Fornell and Larcker, 1981) and the heterotrait–monotrait ratio of correlation (HTMT) (Henseler et al., 2015) were applied to test discriminant validity. As presented in Table 5, the AVE square root of each construct exceeded the correlation values of any other constructs. The cross-loadings (see Table 6) indicate that each construct item had a stronger loading on its parent construct than on other constructs (marked in italic). All HTMT values for the constructs were below 0.900 (see Table 7), which is the threshold for the HTMT ratio suggested by Henseler et al. (2015). Accordingly, the discriminant validity of the research data was established.
4.4 Results
The structural model was tested using the PLS bootstrapping process. As postulated, appearance had a significant and positive relationship with perceived utilitarian value (β = 0.251, p < 0.001) and perceived social value (β = 0.265, p < 0.001). However, the relationship between appearance and perceived hedonic value was insignificant. The voice also showed significant positive correlations with perceived utilitarian value (β = 0.171, p < 0.001), perceived social value (β = 0.190, p < 0.001), and perceived hedonic value (β = 0.308, p < 0.001). Similarly, response was significantly and positively related to perceived utilitarian value (β = 0.452, p < 0.001), perceived social value (β = 0.368, p < 0.001), and perceived hedonic value (β = 0.372, p < 0.001). In addition, perceived utilitarian value (β = 0.217, p < 0.001), perceived social value (β = 0.121, p < 0.05), and perceived hedonic value (β = 0.514, p < 0.001) displayed significant positive correlations with user satisfaction. Therefore, as depicted in Figure 2, all other proposed hypotheses were supported, except H1c. The model explained 50.1% of the variance in perceived utilitarian value, 43.2% of the variance in perceived social value, 40.1% of the variance in perceived hedonic value, and 52.9% of the variance in user satisfaction. Age, gender, and previous experience with robot services, included as control variables, showed no significant effects on user satisfaction.
Stone–Geisser’s Q2 in SmartPLS 4.0 was used to assess the predictive relevance of the model (Geisser, 1974; Stone, 1974). The Q2 values for perceived utilitarian value, perceived social value, perceived hedonic value, and user satisfaction were 0.483, 0.413, 0.386, and 0.334, respectively, indicating strong predictive relevance. Moreover, the model fit was calculated by utilizing the standardized root mean square residual (SRMR) (Henseler et al., 2016). A better model fit was indicated through the SRMR value, which was 0.067, falling below the recommended threshold of 0.080 (Hu and Bentler, 1999).
5. Discussion
The current study investigated the varied effects of HSRs’ different anthropomorphic features on perceived utilitarian, social, and hedonic values, which subsequently influenced user satisfaction with robot services. The results revealed that HSRs’ three anthropomorphic features: appearance, voice, and response, as stimuli in the SOR framework, play distinct roles in triggering users’ perceptions of the utilitarian, social, and hedonic values of HSRs. Specifically, the responses of HSRs showed the strongest influence on all perceived values. Apart from an insignificant influence of appearance on perceived hedonic value, all other relationships between anthropomorphic features and perceived values were significant. Additionally, all perceived IS values as organism in the SOR framework, contributed to user satisfaction (attitudinal response), with perceived hedonic value exerting the strongest effect.
The findings of this study raised several intriguing insights. First, the findings revealed that the three anthropomorphic features as stimuli affect the perception of the utilitarian value of HSRs differently. The responses influenced user perceptions of utilitarian value most strongly (β = 0.452, p < 0.001), followed by appearance (β = 0.251, p < 0.001), with voice having the weakest influence (β = 0.171, p < 0.001). This finding aligns with Birnbaum et al. (2016), who claimed that robots’ responses primarily trigger user perceptions concerning the robots’ competency in performing different tasks. HSRs’ physical appearance and voice may induce user perceptions of the robots’ humanlikeness and the degree to which user–HSR interactions could resemble traditional human–human service encounters at hotels. However, the responses from HSRs could elevate personalized interactions more than appearance and voice, as these features are mostly static in nature. The appearance and voice of an HSR may initially attract the attention of users due to curiosity. Nevertheless, the responses from HSRs could result in positive and lasting user experiences that provide utilitarian value, making them the critical factor affecting the perceived utilitarian value more than HSRs’ appearance and voice.
Second, the study found that appearance, voice, and responses, as stimuli, influence the perceived social value of HSRs differently. The response feature demonstrated the strongest link with user perceptions of social value (β = 0.368, p < 0.001), followed by appearance (β = 0.265, p < 0.001), with voice exerting the least influence (β = 0.190, p < 0.001). As Birnbaum et al. (2016) state, responsive robots are perceived as more sociable than unresponsive robots. Meanwhile, robots’ responses, when understood meaningfully by users, foster friendships and lead to psychological bonds between users and robots during HRIs (MacDorman et al., 2009). In hotels, HSRs, such as receptionist robots, are expected to interact with users by mimicking human social behaviors (Belanche et al., 2020). Users expect personalized, context-specific, and timely responses when they engage with hotel service robots. Robots that provide relevant and personalized responses are perceived as more socially engaging, elevating their social value (Vishwakarma et al., 2024). This aligns with the idea that the responsive behavior of HSRs mimics human social interactions, enhancing the perceived social value of HSRs. Appearance, which is important for the initial impression of users, plays a secondary role as it does not engage with users during ongoing interactions. However, appearance can still influence the perception of social value by conveying a level of human likeness and making users more open to interactions (Broadbent et al., 2013). The weaker impact of voice on perceived social value may be attributed to its function primarily as a medium for communication rather than a driver of dynamic social interactions. This observation aligns with the findings of Li et al. (2023), which suggest that while a robot’s humanlike voice may capture users’ initial attention, it does not, however, independently contribute to deeper or sustained social engagement. Thus, HSRs’ responses more strongly trigger users to perceive robots’ social value than their voice and appearance.
Third, from the three anthropomorphic features, only response and voice, as stimuli, were found to trigger user perception of HSRs’ hedonic value. The response had a stronger influence on user perceptions of the hedonic value (β = 0.372, p < 0.001) than voice (β = 0.308, p < 0.001), while appearance exhibited no effect. The study by Birnbaum et al. (2016) on social robots showed that effective responses from social robots positively affect their appealing traits and the tendency of users to approach them while realizing their utility or function. Similarly, according to the findings of this study, HSRs’ responses are the main stimuli for the perception of hedonic value. While HSRs’ responses could arouse joy and fun among hotel customers, they also rely on robots’ voices to perceive the hedonic value during their use. Previous studies have found that realistic-sounding voices could evoke a high level of enjoyment among users when interacting with voice assistants (Schreibelmayr and Mara, 2022). However, the role of HSRs’ appearance as visual stimuli is mostly static during the HRIs in hotel services, and therefore, it may not contribute to the perception of hedonic value for users at the level of HSRs’ response or voice.
Finally, as hypothesized, user perception of the utilitarian, hedonic, and social values of HSRs positively impacted user satisfaction (attitudinal response) with robotic services. The perceived hedonic value (β = 0.514, p < 0.001) had the strongest impact on user satisfaction, followed by the utilitarian value (β = 0.217, p < 0.001) and the social value (β = 0.121, p < 0.05). These results align with prior findings of De Graaf and Ben Allouch (2013), who noted that a stronger perception of the IS values involving adaptability, usefulness, enjoyment, companionship, and sociability leads to greater satisfaction levels with robotic services. Accordingly, HSRs can substitute for human employees by evoking joy and positive emotions during interactions, which eventually enhances overall user satisfaction. In addition, the findings of this study are consistent with the findings of Odekerken-Schröder et al. (2022), which revealed that both utilitarian and hedonic values of HSRs influence customer repatronage intentions, mirroring user satisfaction. Although the perceptions of utilitarian, hedonic, and social values affect user satisfaction, the perceived hedonic value exerts the strongest influence due to its emotional aspect, fostering positive memories and emotional connections within overall user experiences. Further, the perception of enjoyment or pleasure influences users’ intrinsic motivation to use robots (Said et al., 2024). As a result, in contexts such as hotels where users often seek enjoyment, the hedonic perception of HSRs could surpass the perception of utilitarian value or social value.
6. Conclusion
6.1 Theoretical contributions
This paper makes several theoretical contributions to the existing literature. First, this study is one of the first studies to investigate the effects of different anthropomorphic features such as appearance, voice, and response of HSRs on user satisfaction through the organism of users’ value perceptions (utilitarian, social, and hedonic values). Unlike prior research, which has predominantly concentrated on a single anthropomorphic feature when disentangling the role of anthropomorphic features of HSRs on user perceptions and behaviors (Diederich et al., 2021; Klüber and Onnasch, 2022; Seo, 2022), the current study broadens the scope by examining three distinct features: appearance, voice, and response. This approach not only extends the existing framework of anthropomorphic features of HSRs but also offers a comprehensive understanding of how various anthropomorphic features of HSRs shape user perceptions and behaviors differently. Specifically, by unveiling the distinct effects of the three different anthropomorphic features on user satisfaction through the mechanism of user perceptions of the utilitarian, social, and hedonic values of HSRs, our findings reveal that response, as a stimulus, strongly influences all three perceived IS values compared to voice and appearance, with appearance particularly lacking influence on the perception of hedonic value. These insights highlight the complexity of the cognitive process acting as the organism, providing a detailed explanation of how various anthropomorphic features of HSRs shape users’ attitudinal responses.
Second, the current study deepens the understanding of the role of IS value perceptions as an organism within the SOR framework, particularly in investigating the impact of different anthropomorphic features on user satisfaction with HSR services. While HSRs are generally designed to offer utilitarian, social, and hedonic values, especially within the hotel context where they are primarily intended for utilitarian purposes (Ozturk et al., 2023), our findings emphasize the critical importance of hedonic value perception in determining user satisfaction. The study reveals that perceived hedonic value exerts the most substantial impact on user satisfaction, surpassing the effects of both perceived utilitarian and perceived social values.
Finally, this study provides novel insights into the HSR literature by confirming that the anthropomorphic features of HSRs can effectively function as stimuli within the SOR framework to trigger users’ intrinsic cognition and attitudinal responses, such as user satisfaction. While previous studies have predominantly applied traditional technological features of IS as stimuli in the SOR framework (Cheng et al., 2022; Hlee et al., 2023), our findings demonstrate that the anthropomorphic features of HSRs—distinctive and innovative aspects of humanlike technologies—should be recognized as vital stimuli. This recognition advances the understanding of how users cognitively and behaviorally interact with humanlike technologies, thereby influencing their attitudinal responses. This contribution not only extends the applicability of the SOR framework but also enriches the broader discourse on HRIs and the design of socially interactive technologies.
6.2 Practical contributions
The findings of this study deliver actionable insights for robot developers and hotel management, offering practical guidance on optimizing design and deployment of HSRs to enhance user satisfaction.
For HSR developers, the study highlights the critical importance of the response feature in shaping user perceptions of utilitarian, social, and hedonic values. Since responses were found to be the most influential factor across all perceived values, and with perceived hedonic value emerging as the most impactful factor on user satisfaction, developers should prioritize improving the quality and humanlikeness of HSRs’ responses, focusing on their hedonic capabilities. Enhancements to the response of HSRs could include optimizing response times, enabling more personalized, natural, and adaptive interactions similar to human-human interactions, and ensuring that responses effectively address user needs in real-time. Such improvements can elevate the utility of HSRs, simultaneously enhancing the overall user experience through more engaging and enjoyable interactions, which may eventually result in higher user satisfaction with HSR services.
The voice feature, although not as critical as the response feature, still significantly impacts all perceived values. Developers are advised to optimize the voice quality of HSRs by refining elements such as softness, tone, loudness, and accent and by developing the voice to be more humanlike. A well-modulated voice can potentially enhance user engagement with HSRs (Li et al., 2023), contributing to a more pleasant and engaging experience, ultimately leading to improved user satisfaction with HSR services.
The findings also suggest that HSR developers could deprioritize extensive customization of appearance, given its insignificant impact on hedonic value. Instead, resources could be better utilized to enhance interactive capabilities, such as the ability to engage in responsive, meaningful, socially engaging, emotionally resonant conversations as well as practical functionalities, ensuring that HSRs are more effective in fulfilling their roles in hotel settings. For example, designs could focus on optimizing HSRs’ ability to assist with assigned tasks (e.g. check-ins and navigational assistance), thereby improving the overall guest experience through the functions of HSRs rather than focusing on features such as appearance.
For hotel managers, the study’s findings underscore the importance of enhancing the perceived utilitarian, hedonic, and social values of HSRs to boost user satisfaction. Managers should carefully consider the primary purpose of the HSRs they deploy—whether it is to serve utilitarian tasks, provide entertainment, foster social interactions, or serve a combination of tasks. Depending on the intended purpose, robots with different features should be deployed. For instance, if the goal is to improve customer service, HSRs should be equipped with highly responsive features to ensure they can meet customer needs effectively. If the goal is to entertain hotel guests, an HSR with a pleasant, humanlike voice combined with effective responses is recommended, as it can significantly enhance guests’ enjoyment and satisfaction. However, despite the intended purpose, the hotel managers needs to ensure the quality of HSRs’ responses through the evaluation of feedback from hotel customers. By promoting the functionality and unique experiences that HSRs offer, managers can divert guest attention away from appearance and toward the robot’s contributions to service quality. This approach enhances user satisfaction with HSR services while aligning with the practical realities of resource allocation in HSR design and deployment.
6.3 Limitations and avenues for future research
The current study encountered a few limitations. First, it mainly included three anthropomorphic features of HSRs: appearance, voice, and response. Other factors, such as social robots’ level of autonomy, could be considered in future research to explain how anthropomorphic features influence user behaviors involving HSRs. Second, the study’s data were gathered from hotel customers who had experienced robots during their stay. In the future, researchers could consider duplicating this study across different contexts to enhance its generalizability. Third, this study examined user satisfaction with HSR services using the SOR framework. Further research could consider investigating the impact of anthropomorphic features on perceived humanness, perceived hospitality, or emotions by applying alternative theories such as social cognitive theory, social presence theory, or trust. Finally, future research could also investigate the effects of anthropomorphic features on users’ affective states, such as their positive and negative feelings or emotions, deviating from the current study’s focus on how these features impact user perception of IS values in terms of utilitarian, social, and hedonic values.
