The impact of perceived anthropomorphism of the conversational agent (CA) on social presence, flow, and users’ behavioral intention

1 Introduction

Artificial intelligence has revolutionized marketing by enabling personalized customer experiences through the exploitation of data to create tailored solutions (Dieshbah 2006) adapted to specific profiles (Acquatella 2018). Companies must understand their customers’ journeys to develop applications or websites that strengthen their relationship. The integration of chatbots, as intelligent interfaces, can significantly improve the quality of a company’s customer service by offering unique, positive, and personalized experiences. According to Xu et al. (2020), these high-quality interfaces are essential for optimizing user interactions, while (Aladwani and Dwivedi 2018; Allal-Ch’erif et al. 2021; Bhawiyuga et al. 2017) highlight their role in creating customer experiences that are increasingly tailored to individual needs. In the field of digital marketing and e-commerce, numerous activities are now carried out through conversational agent systems, as observed by (Kar et al. 2021; McLean et al. 2019; Moore 2018; Wang et al. 2021). This highlights the growing importance of these tools in transforming interactions between companies and customers. Providing a seamless customer experience is a key objective for messaging companies, which have widely adopted chatbots as a new means of communicating with their customers (Androutsopoulou et al. 2019; Cui et al. 2017; Nagi et al. 2021); to achieve this, understanding the psychological factors that influence online behavior is essential for marketers to personalize experiences that resonate with consumers (Meparishvili 2023).
Customer behavior is strongly influenced by their experiences, leading to either positive or negative responses. Research indicates that positive experiences often result in favorable behaviors, such as positive electronic word-of-mouth (eWOM) sharing on symmetrical platforms, while negative experiences can lead to disengagement and negative engagement behaviors, including complaints and negative eWOM (Chen et al. 2023; Do and Bowden 2024; Do 2022). Optimal online experiences, or flow, have been defined by (Hoffman and Novak 1996) as a state of intense satisfaction reached when individuals are fully engaged in an activity. They demonstrated that businesses can utilize the digital environment to create engaging user experiences, emphasizing that online flow fosters a seamless interaction between user and machine. Marketers assert that consumers in a state of flow are more likely to purchase and return to a site (Bridges and Florsheim 2008). Furthermore, research on human-computer interactions indicates that pleasure and engagement during computer interactions generate positive emotions (Sandelands and Buckner 1989; Starbuck and Webster 1991). According to the social response theory, anthropomorphism of technology occurs when the user interacts with a screen or a virtual agent, creating a psychological link that prompts the attribution of human characteristics to the computer (Moon 2000; Lemoine and Cherif 2012). Studies show that this perception of social presence is crucial in human-computer interaction (Shin and Shin 2011), where an anthropomorphic virtual agent can strengthen the user’s sense of psychological proximity.
As mentioned by Nass and Moon (2000) and Reeves and Nass (1996), the key paradigm in the design and interaction with conversational agents is grounded in their perception as social actors. The CASA (Computers As Social Actors) model proposes that humans apply social expectations to technologies with human-like characteristics, which has been confirmed by experiments showing that individuals use social categories when interacting with humanoid artifacts. Seeger et al. (2018) identify three dimensions in the anthropomorphic design of conversational agents: human identity, verbal signals, and nonverbal signals. Users often attribute a social role to these technologies, such as a member of a customer service team, and the more human-like traits an artifact possesses, the more intense social responses it elicits (Nass and Moon 2000). Conversational agents incorporate various social cues, such as natural language and emotional expression, prompting users to respond meaningfully (Feine et al. 2019). Focusing specifically on human identity portrayal and the use of verbal and nonverbal cues to communicate effectively with users in a social context as highlighted by (Diederich et al. 2020; Pfeuffer et al. 2019).
By utilizing the SOR model with flow theory and CASA paradigm. We examine how the perception of anthropomorphism influences users’ intentions; In another manner, our research study focusing on the impact of perceived flow and social presence on user behavior intentions. The central question of this study is:

„How does perceived anthropomorphism influence users’ behavioral intention through perceived flow and perceived social presence?“

In addressing this issue, the study focuses on a chatbot-type conversational agent with which users interact in writing, accompanied by a static virtual avatar; This method allows us to explore the impact of the combination of textual and visual elements on the user experience in customer service where the state of flow is crucial for the efficiency of online experience .By integrating these communication modalities, we examine how visual representation influences users’ perception of flow and social presence towards the chatbot, as well as improving behavioral intention in critical situations. we have developed a theoretical framework that forms the basis of our research, evaluated among an online experiment. This approach aims to evaluate our theoretical framework while enriching the understanding of interactions between users and conversational agents in digital environments.

2 Theoretical framework

Building upon research, on anthropomorphic conversational agents and CASA, flow theories in online environments. We have crafted a research framework consisting of four hypotheses grounded in established principles of human computer interaction, for validation through an online study.

Figure 1: Research model
Source: Authors

2.1 Stimulus

According to SOR model stimulus originate from external environment, that triggers a reaction in an organism. In this study, Stimuli are signals that encourage users to interact with the conversational agent; we examine how the anthropomorphic characteristics of the conversational agent associated with social cues can influence the interaction.

2.1.1 Social cues

Building on the work of Nass and Moon (2000), technological artifacts are more likely to trigger social reactions when they exhibit human-like characteristics. Feine et al. (2019) and Wünderlich and Paluch (2017) indicate that social cues lead people to perceive conversational agents as human beings. Additionally, Seeger et al. (2018) have suggested that a CA incorporating social cues gives rise to three dimensions: a human identity, verbal cues, and nonverbal cues.

2.1.2 Anthropomorphic conversational agent

The theory of anthropomorphism posits that the more human-like the visible traits of non-human agents, the more likely users are to perceive them as such (Epley et al. 2007). An anthropomorphic conversational agent is defined by the attribution of human characteristics to an artefact. This similarity can manifest in both the appearance and behavior of the agent (Waytz et al. 2010b). One study revealed that participants interacting with a chatbot identified characteristics that made it seem more human, such as its ability to maintain a conversation over an extended period, consequently increasing its realism. Moreover, anthropomorphism can engender social behaviors in the user (Epley et al. 2007; Waytz et al. 2010b), as a conversational agent engages in interactions similar to those of face-to-face encounters (Qiu and Benbasat 2009). This prompts users to apply the usual social rules, which increases their perception of social presence when interacting with an anthropomorphic agent. Thus, we formulate two hypotheses:
H1: Perceived anthropomorphism of the conversational agent has a positive impact on the user’s perceived social presence.
H2: Perceived anthropomorphism of conversational agents positively influences the user’s state of flow.

2.2 Organism

Given that the stimulus-organism-response framework proposes that how stimuli impact outcomes is influenced by individuals’ states. Our main focus is presence (Biocca et al. 2003; Kreijns et al. 2007). Flow (Hoffman and Novak 1996; Koufaris 2002). These internal states can be shaped by the stimuli in our research model. Have an impact intention.

2.2.1 Social presence

Our goal, in this study is to explore how the social presence of a human assistant affects people’s behavioral intentions. Conversational agents represent a particularly interesting phenomenon given that users show positive social responses to these agents (Diederich et al. 2020; Pfeuffer et al. 2019). Research has indicated that perceived social presence positively influences user perception, which changes their behavioral intention Qiu and Benbasat (2009). For this purpose, we propose the hypothesis:
H3: Perceived social presence positively influences users’ behavioral intentions.

2.2.2 Flow

In response to evolving consumer behavior with technological tools (artifacts), organizations are using IA to enhance the quality of the services they provide to their customers. This allows them to deliver positive experiences as referenced by (Aladwani and Dwivedi 2018; Allal-Ch’erif et al. 2021; Bhawiyuga et al. 2017). Zhou (2013) showing that flow influences the decision to keep using mobile payment services. Zhou et al. (2010) showed the positive effect of flow on bringing users closer to an online mobile service; Hausman and Siekpe (2009) indicated that flow predicts the online user’s intention to return to a site to relive the same experience. Based on these findings we suggest the following hypothesis:
H4. Perceived flow positively affects users’ behavioral intentions.

2.3 Response

The psychological effects mentioned earlier play a role as they influence both attitudes and behaviors significantly.

2.3.1 Behavioral intention

The more similar people are, the more they try to communicate and understand each other, facilitating the exchange of messages between them (Rogers and Bhowmik 1970); Individuals tend to evaluate favorably those who are similar to them (Goethals and Nelson 1973). This suggests that if individuals see a computer assistant as similar, to themselves; they are likely to give it an assessment. Indeed, studies have shown that a CA with a human-like morphology is more likely to be perceived as more likable than an agent with a less human-like morphology (Koda 1996; Wexelblat 1998).

3 Empirical study

This empirical study aims to analyze the influence of design characteristics of a conversational agent, or chatbot, on user intention within the online customer service of CASH Assurances. The chatbot, which interacts with users via instant messaging, has been integrated into the Facebook Messenger platform and is also accessible on a dedicated website. Participants were invited to converse with this virtual assistant to obtain information about the company and its offerings. After a preparatory briefing, participants first consulted the frequently asked questions about insurance before using the chatbot to explore the various insurance options. Inspired by previous research on conversational agents (Diederich et al. 2019c; Gnewuch et al. 2018), we selected a set of representative tasks to evaluate the chatbot’s ability to answer diverse questions. The experience lasted approximately nine minutes for each participant, with a sample of 329 individuals aged 19 to 59 (mean: 27.8 years) and comprising 56.9% women. Participants were recruited from personal networks, primarily employees and university students.

3.1 Design prototyping

The aim was to give the impression that the agent had human qualities through the incorporation of social cues. to detect user intent and provide appropriate responses for user, we developed our conversational agent using Dialog flow, a Google natural language platform (2019), which enables the creation of natural conversational experiences through machine learning (Canonico and De Russis 2018). To create a human identity for our agent, we relied on the three dimensions of anthropomorphic design proposed by (Seeger et al. 2018). We first gave our agent’s name and gender (Cowell and Stanney 2005; Nunamaker et al. 2011), as well as an avatar representing a customer service employee (Gong 2008), which allowed for more natural responses. Next, we integrated verbal cues to enrich the experience, such as self-disclosures (Schuetzler et al. 2018), self-references (Sah and Peng 2015), and personalized greetings (Cafaro et al. 2016). Finally, nonverbal cues like emoticons (Wang et al. 2008) and dynamic response delays (Gnewuch et al. 2018). By integrating these elements, our goal was to make the agent appear human by using various social signals (Feine et al. 2019). Figure 2 presents our prototype.

3.2 The measures

As previously mentioned, each participant completed a questionnaire measuring their perceptions of anthropomorphism, social presence, flow state, and behavioral intention after online experience with chatbot. To measure flow state, we adapted a five points Likert scale based on the studies of (Novak et al. 2003). Perceived anthropomorphism was measured also on a five points Likert scale using items from (Holtgraves and Han 2007). then, we used a five points Likert scale based on the work of (Keeling et al. 2010) to measure behavioral intention. In the end, social presence was measured on a five points Likert scale using items from (Qiu and Benbasat 2009).

Table1: Survey items
Source: Authors (measurement constructs derived from the literature)

3.3 Structural equation modelling (SEM)

Structural equation modelling (SEM) technique was used to analysis our multivariate data. Given the higher flexibility provided by the alternative approach of SEM, namely the partial least square (PLS-SEM) compared to more basic covariance-based CB-SEM one (Latan 2018), we deemed it necessary to estimate our SEM model following this approach. The model processed via the PLS-SEM method follows a classic two-step approach as described by PLS experts. This approach involves ensuring the quality of the measurement model and then, once confirmed, the structural model can be examined, thereby allowing for testing the research hypotheses (Sarstedt et al. 2021). SmartPLS.4 (Ringle et al. 2024), the leading software for PLS-SEM, was used to perform all required analyses.

3.3.1 Evaluation of the measurement model

When the model consists of solely reflective constructs, its evaluation focuses on the reliability and validity of the measures (Sarstedt et al. 2014). The reliability of an indicator is established if its loading is greater than 0.708. Although loadings of at least 0.4 are acceptable as long as they do not compromise the reliability and convergent validity of the construct (Hair and Alamer 2022). The internal consistency of the construct can be evaluated by various measures such as Cronbach’s alpha or Jöreskog’s rho, but PLS experts’ advice for the use Dijkstra and Henseler’s rho (A) coefficient (2015) instead, given its greater accuracy. Values of rho (A) greater than 0.7 or even 0.6 for exploratory studies indicate good internal consistency (Sarstedt et al. 2022). Convergent validity is achieved when the average variance extracted (AVE) index is greater than 0.5, which means that the construct explains more than 50% of the variance of his associated items (Ringle et al. 2020). Discriminant validity is confirmed when the shared variance between the constructs is less than their respective AVEs, known by Fornell-Larcker criterion (Hair et al. 2021). The analysis of our model using the PLS algorithm showed that this latter meets the above referenced standards of good measurement quality after the elimination of certain items.

Table 2: Evaluation of the reliability and convergent validity of the measurement model
Source: Authors

Table 3: Testing the discriminant validity using the Fornell-Larcker criterion
Source: Authors

3.3.2 Structural model evaluation and hypothesis testing

The structural model incorporates the regression relationships linking the constructs together, in other words, the research hypotheses (Hair et al. 2014). It is often referred as the inner model in PLS-SEM terminology (Hair et al. 2017). The examination of the structural or inner model using the PLS-SEM approach involves assessing the significance and relevance of the path coefficients, and then evaluating the explanatory and predictive power of the model (Cheah 2019). However, given that the inner model is formed by regression relationships, it is necessary to ensure that it does not present with a multicollinearity problem, as this can affect the quality of the estimates later (Ghasemy et al. 2020). To this, Variance inflation factor (VIF) indices for endogenous constructs of less than 3 or at the limit less than 5 already indicate the absence of multicollinearity (Alamer et al. 2022). This was confirmed in our model (see Table 3).
The estimation of path coefficients and model parameters in PLS-SEM is usually performed using a non-parametric analysis called bootstrapping, as this method does not assume normality of the data (Risher and Hair 2017). For this reason, PLS-SEM is particularly effective for testing indirect effects (Nitzl et al. 2016; Nitzl and Chin 2017). A parameter is considered to be statistically significant when its t-value is greater than |1.96|, its p-value is less than 0.05 (a commonly used threshold in marketing studies (Hair et al. 2022)), or even when zero is not included in the its respective confidence interval (Risher and Hair 2017). Regarding the relevance of structural coefficients, Cohen’s f² effect sizes measure this relevance, with values greater than 0.35 indicating a strong impact of the independent variable over the dependent variable, while values between 0.15 and 0.35 indicate medium impacts and between 0.02 and 0.15 a weak ones (Manley et al. 2021). In our study, the estimation of the model using 10,000 subsamples and a percentile method for constructing confidence intervals (Hair et al. 2022), revealed that all hypotheses were confirmed with strong effect sizes in most case. Moreover, partial mediations of flow and social pressure between the anthropomorphic character and behavioral intention were identified as their respective VAF (Variance Accounted For) indices ranging between 20% and 80% (Nitzl et al. 2016).

Table 4: Testing of direct effects
Source: Authors

Table 5: Testing of indirect effects
Source: Authors

Inner model validation involves also evaluating its explanatory and predictive power. Explanatory capacity usually is measured by the coefficient of determination R² in PLS-SEM models. This coefficient indicates the percentage of explained variance of the endogenous variable, with values close to 1 or greater than 0.75 being desirable (Ghasemy et al. 2020). However, the interpretation of R² must take into account the complexity of the model, particularly the number of exogenous variables. The more predictors there are for a construct, the higher the expected R² (Hair et al. 2019). In marketing studies, R² values of 0.75, 0.5, and 0.25 correspond to high, moderate, and low explanatory powers, respectively (Hair et al. 2011). In our model, an R² of 0.718 was obtained for the behavioral intention variable, the model’s final variable, indicating therefore a high explanatory power.
To assess the predictive capacity of the inner model, the researcher can use either the predictive relevance index Q² or the more advanced PLS predict analysis. Although the calculation of Q² is based on the omission of some part of the original data, it cannot be considered as a full measure of predictive power rather than an indicator of both explanatory and predictive power together. Usually, a positive Q² indicates the presence of some predictive relevance, while values greater than 0.5 suggest a high predictive relevance (Ringle et al. 2020). On the other hand, PLS predict evaluates the predictive capacity of a PLS model more accurately by using out-of-sample data or holdout samples (Shmueli et al. 2019). PLS predict provides statistics which quantify the amount of prediction errors relative to the PLS model and two benchmarks formed by average indicators (AI benchmark) and linear regressions (benchmark LM). The more the PLS model shows less prediction errors than its benchmark, the higher will be its predictive power.
PLS predict prediction error summaries are respectively the Q² predict, then the MAE (mean absolute error) and/or the RMSE (root mean squared error) indices. Q² predict must be positive for all items of the targeted variable to demonstrate that the PLS model has at least more predictive ability than its most naive benchmark which is the AI one. If confirmed, the MAE and/or RMSE of the PLS model (indicators of the targeted variable) should be compared with those of the less naïve LM benchmark. High predictive ability of PLS model is established if his MAE and/or RMSE are lower than the LM’s ones (Shmueli et al. 2019). The evaluation of the predictive power of our model has shown that the latter has a high predictive power since it has a Q²=0.408 for the model’s final construct (behavioral intention) and it outperforms its two benchmarks in terms of out-of-sample predictive performance (see table 5).

Table 6: PLS predict analysis and examination of the predictive power of the PLS-SEM model (k=10, r=10)
Source: Authors

4 Results and discussion

This study contributes to the academic literature in two primary ways. First, this research enriches previous work on social presence and its effect on behavioral intention, while also improving anthropomorphism through social cues (Hossain et al. 2023; Munnukka et al. 2022; Nisha. et al. 2024). Thus, social cues have a positive impact on perceived anthropomorphism and that the state of flow can influence behavioral intention towards the chatbot.
The effects of anthropomorphism raise concerns among marketing researchers, as they challenge the notion that human-like resemblance positively influences users’ behavioral intentions, as highlighted by (Følstad and Skjuve 2019) and (Araujo 2018). So, this article confirms that integrating an avatar into an insurance company’s online customer service chatbot enhances the state of flow.
We also found that the S-O-R model provided a better fit, confirming that the flow and social presence variables mediate the relationship both directly and indirectly. Furthermore, this study highlights the potential of the flow state generated during interactions. Our results show that perceived anthropomorphism, reinforced by social cues, contributes to this positive psychological state. To our knowledge, no previous research has explored the effects of anthropomorphism on flow in the context of online customer service while using the S-O-R model. The innovative results of this study enrich the literature on anthropomorphism by providing additional information about the factors that can induce a positive flow state, thus favorably influencing users’ behavioral intentions.
This study has significant practical implications for chatbot designers and online marketing specialists looking to integrate a chatbot as a virtual assistant or improve the user experience in human-machine interactions. Based on the experimental results, it is recommended that designers create avatars by incorporating social cues to simulate human conversation. The integration of avatars could mitigate the risks of misperception of social presence, as highlighted by (Kear 2010) and (Payne et al. 2012). Moreover, anthropomorphism linked to appearance has an impact on behavioral intention, as a negative experience can harm this intention. It would therefore be wise to design chatbots capable of inspiring flow in users by using natural visual cues and reinforcing their expertise.

5 Limitation

Future research should focus on exploring the underlying mechanisms by introducing varied experimental conditions to better understand the mediating effect of flow and social presence. for insurance related information gathering a service that often demands thorough research before committing to a decision. Additionally, as highlighted by Kunreuther et al. (2015) and Kunreuther and Pauly (2018), certain individual factors were not considered in this study. The experimental scenario involved interacting with chatbots to obtain information about insurance, a high-involvement service that typically requires extensive research before making a decision.
In our study, it is possible that unmeasured factors, such as how participants felt about insurance before taking part. That could have affected our findings. To ensure the validity of our conclusions, it would be important for future studies to check and manage how participants feel about intelligence to make sure our findings are accurate. Additionally, although our study used a chatbot-type conversational agent, it would be necessary to generalize the results to other forms of conversational agents, such as embodied agents, that have added features, like voice and nonverbal cues.

6 Conclusion

Exploring how anthropomorphism influences human computer interaction is an aspect of using chatbots that resemble humans in ways such, as having expressive avatars and providing personalized responses to enhance user engagement and immersion in the interaction experience based on our studies. The seamless flow of conversation enhances the feeling of being engaged – it’s, like interacting with a sentient and attentive being.
The blend of human characteristics, in chatbots with a tone and engaging presence positively influences how users behave online. For instance, an avatar that resembles a person promoting an item may spark curiosity and desire for purchase among users compared to a robotic chat interface focused solely on functionality. These conclusions are in line with the SOR framework that examines how stimuli from the environment impact individual’s organismic processes and subsequent responses. By infusing chatbots with traits they serve as stimuli that evoke various cognitive and emotional reactions, in users ultimately shaping their favorable behavioral intentions.
Conclusively speaking like chatbots signify a progression, within the realm of computer interactions, with humans. By nurturing an interaction and social connection they have the capacity to promote acceptance of novel technologies and enhance user satisfaction across various domains, including customer support, online commerce and healthcare services.

Figure 2: Chatbot agent
Source: Authors

Literatúra/List of References

1. Acquatella, F., Fernandez, V. and Houy T., 2018. Quelles dynamiques stratégiques pour les plateformes digitales? Proposition d’un cadre d’analyse à partir du cas Coursera. i3 Working Papers Series, 18-SES-02. 2018. [online]. [cit. 2024-10-30]. Available at: <https://i3.cnrs.fr/wp-content/uploads/2018/01/WPi3-18-SES-02-Acquatella-Fernandez-Houy.pdf>
2. Aladwani, A. M. and Dwivedi, Y. K., 2018. Towards a theory of socio-citizenry: Quality anticipation, trust configuration, and approved adaptation of governmental social media. In: International Journal of Information Management. 2018, 43, 261-272. ISSN 1873-4707. Available at: <https://doi.org/10.1016/j.ijinfomgt>
3. Allal-Chérif, O., Aránega, A. Y. and Sánchez, R. C., 2021. Intelligent recruitment: How to identify, select, and retain talents from around the world using artificial intelligence. In: Technological Forecasting and Social Change. 2021, 169. ISSN 1873-5509. Available at: <https://doi.org/10.1016/j.techfore.2021.120822>
4. Androutsopoulou, A., Karacapilidis, N., Loukis, E. and Charalabidis, Y., 2019. Transforming the communication between citizens and government through AI-guided chatbots. In: Government Information Quarterly. 2019, 36(2), 358-367. ISSN 1872-9517. Available at: <https://doi.org/10.1016/j.giq.2018.10.001>
5. Araujo, T., 2018. Living up to the chatbot hype: The influence of anthropomorphic design cues and communicative agency framing on conversational agent and company perceptions. In: Computers in human behavior. 2018, 85, 183-189. ISBN 1873-7692. Available at: <https://doi.org/10.1016/j.chb.2018.03.051>
6. Bhawiyuga, A., Fauzi, M. A., Pramukantoro, E. S. and Yahya, W., 2017. Design of e-commerce chat robot for automatically answering customer question. November. In: International Conference on Sustainable Information Engineering and Technology (SIET), Batu: IEEE, 2017, 159-162. ISBN 978-1-5386-2182-0.
7. Biocca, F., Harms, C., and Burgoon, J. K., 2003. Toward a more robust theory and measure of social presence: Review and suggested criteria. In: Presence: Teleoperators & Virtual Environments. 2003, 12(5), 456-480. ISSN 1531-3263. Available at: <https://doi.org/10.1162/105474603322761270>
8. Cafaro, A., Vilhjalmsson, H. H. and Bickmore, T., 2016. First impressions in human-agent virtual encounters. In: ACM Transactions on Computer-Human Interaction. 2016, 24(4) , 1-40. ISSN 1073-0516.
9. Canonico, M. and De Russis, L., 2018. A comparison and critique of natural language understanding tools. Article presented at CLOUD COMPUTING 2018: The Ninth International Conference on Cloud Computing, GRIDs, and Virtualization, Barcelona, 2018, 110-115. ISBN 978-1-61208-607-1.
10. Cheah, J. H., Magno, F. and Cassia, F., 2024. Reviewing the SmartPLS 4 software: the latest features and enhancements. In: Journal of Marketing Analytics. 2024, 12, 97-107. ISSN 2050-3326. Available at: <https://doi.org/10.1057/s41270-023-00266-y>
11. Cowell, A. J. and K. M. Stanney, 2005. Manipulation of non-verbal interaction style and demographic embodiment to increase anthropomorphic computer character credibility. In: International Journal of Human Computer Studies. 2005, 62(2), 281-306. ISSN 1095-9300.
12. Cui, L., Huang, S., Wei, F., Tan, C., Duan, C. and Zhou, M., 2017. Superagent: A customer service chatbot for e-commerce websites. In: Proceedings of ACL. Vancouver: Association for Computational Linguistics, 2017, 97-102.
13. Diederich, S., Janßen-Müller, M., Brendel, A. B. and Morana, S., 2019c. Emulating empathetic behavior in online service encounters with sentiment-adaptive responses: Insights from an experiment with a conversational agent. In: Proceedings of the International Conference on Information Systems (ICIS). Munich, 2019, 1-17.
14. Diederich, S., Brendel, A. B. and Kolbe, L. M., 2020. Designing anthropomorphic enterprise conversational agents. In: Business & Information Systems Engineering. 2020, 62(3), 193-209. ISSN 1867-0202.
15. Dieshbah, G. and Chandonal, 2006. Impact of the presence of an embodied virtual agent on the stickiness power of a website: a behavioral analysis using log files. 2006. [online]. [cit. 2024-10-30]. Available at: <http://leg2.ubourgogne.fr/CERMAB/zoutils/documents/actesJRMB/JRM B112006/Diesbach%20Galan%20et%20Chandon.pdf>
16. Dijkstra, T. K. and Henseler, J., 2015. Consistent partial least squares path modeling. In: MIS Quarterly. 2015, 39(2), 297-316. ISSN 0276-7783. Available at: <https://www.jstor.org/stable/26628355>
17. Epley, N., Waytz, A. and Cacioppo, J. T., 2007. On seeing human: A three-factor theory of anthropomorphism. In: Psychological Review. 2007, 114(4), 864-886. ISSN 0033-295X.
18. Feine, J., Gnewuch, U., Morana, S. and Maedche, A., 2019. A taxonomy of social cues for conversational agents. In: International Journal of Human-Computer Studies. 2019, 132, 138-161. ISSN 1071-5819.
19. Følstad, A. and Skjuve, M., 2019. Chatbots for customer service: user experience and motivation. In: Proceedings of the 1st international conference on conversational user interfaces. Dublin: Association for Computing Machinery, 2019, 1-9. ISBN 978-1-4503-7187-2. Available at: <https://doi.org/10.1145/3342775.3342784>
20. Ghasemy, M., Teeroovengadum, V., Becker, J. M. and Ringle, C. M., 2020. This fast car can move faster: A review of PLS-SEM application in higher education research. In: Higher education. 2020, 80(6), 1121-1152. ISSN 1573-174X.
21. Gnewuch, U., Morana, S., Adam, M. T. P. and Maedche, A., 2018. Faster is not always better: Understanding the effect of dynamic response delays in human-chatbot interaction. In: Proceedings of the European Conference on Information Systems (ECIS). Portsmouth, 2018, 1-17
22. Goethals, G. R. and Nelson, R. E., 1973. Similarity in the influence process: The belief value distinction. In: Journal of Personality and Social Psychology. 1973, 25(1), 117-122. ISSN 1939-1315. Available at: <https://doi.org/10.1037/h0034266>
23. Gong, L., 2008. How social is social responses to computers? The function of the degree of anthropomorphism in computer representations. In: Computers in Human Behavior. 2008, 24(4), 1494-1509. ISSN 1873-7692.
24. Hair, J. F., Hult, G. T. M., Ringle, C. M. and Sarstedt, M., 2022. A primer on partial least squares structural equation modeling (PLS-SEM). Thousand Oaks, CA: Sage, 2022. ISBN 978-1-5443-9640-8.
25. Hair, J. and Alamer, A., 2022. Partial least squares structural equation modeling (PLS-SEM) in second language and education research: Guidelines using an applied example. In: Research Methods in Applied Linguistics. 2022, 1(3), 100027. ISSN 2772-7661.
26. Hair Jr, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M., Danks, N. P. and Ray, S., 2021. Partial least squares structural equation modeling (PLS-SEM) using R: A workbook. Springer Nature, 2021. ISBN 978-3-030-80519-7.
27. Hair, J. F., Risher, J. J., Sarstedt, M. and Ringle, C. M., 2019. When to use and how to report the results of PLS-SEM. In: European business review. 2019, 31(1), 2-24. ISSN 1758-7107.
28. Hair Jr, J. F., Sarstedt, M., Hopkins, L. and Kuppelwieser, V. G., 2014. Partial least squares structural equation modeling (PLS-SEM) An emerging tool in business research. In: European business review. 2014, 26(2), 106-121. ISSN 1758-7107.
29. Hair, J. F., Ringle, C. M. and Sarstedt, M., 2011. PLS-SEM: Indeed a silver bullet. In: Journal of Marketing theory and Practice. 2011, 19(2), 139-152. ISSN 1944-7175.
30. Hausman, A. V. and Siekpe, J. S., 2009. The effect of web interface features on consumer online purchase intentions. In: Journal of business research. 2009, 62(1), 5-13. ISSN 1873-7978.
31. Henseler, J., Ringle, C. M. and Sarstedt, M., 2015. A new criterion for assessing discriminant validity in variance-based structural equation modeling. In: Journal of the academy of marketing science. 2015, 43, 115-135. ISSN 0092-0703.
32. Hoffman, D. L. and Novak, T. P., 1996. A new marketing paradigm for electronic commerce. In: Information Society. 1996, 13(1), 43-54. ISSN 1087-6537.
33. Holtgraves, T. and Han, T. L., 2007. A procedure for studying online conversational processing using a chatbot. In: Behavior Research Methods. 2007, 39(1), 156-163. ISSN 1554-351X.
34. Hossain, M. A., Yesmin, N., Jahan, N. and Reza, S. M. A., 2023. Effect of social presence on behavioral intention to social commerce through online social capital. In: International Journal of e-Collaboration. 2023, 19(1),1-23. ISSN 15483681.
35. Kar, A. K., Kushwaha, A. K., 2023. Facilitators and barriers of artificial intelligence adoption in business – insights from opinions using big data analytics. In: Information Systems Frontiers. 2023, 25(4), 1351-1374. ISSN 1572-9419. Available at: <https://doi.org/10.1007/s10796-021-10219-4>
36. Keeling, K., McGoldrick, P. and Beatty, S., 2010. Avatars as salespeople: Communication style, trust, and intentions. In: Journal of Business Research. 2010, 63(8), 793-800. ISSN 1873-7978.
37. Kear, K., 2010. Social presence in online learning communities. In: Proceedings of the 7th international conference on networked learning. Aalborg, 2010, 541-548. ISBN 978-1-86220-225-2.
38. Keeling, K., McGoldrick, P. and Beatty, S., 2010. Avatars as salespeople: Communication style, trust, and intentions. In: Journal of Business Research. 2010, 63(8), 793-800. ISSN 1873-7978. Available at: <https://doi.org/10.1016/j.jbusres.2008.12.015>
39. Koda, T., 1996. Agents with faces: A study on the effects of personification of software agents (Unpublished master thesis). Boston, MA, USA: Massachusetts Institute of Technology, 1996.
40. Koufaris, M., 2002. Applying the technology acceptance model and flow theory to online consumer behavior. In: Information Systems Research. 2002, 13(2), 205-223. ISSN 1047-7047.
41. Kreijns, K., Kirschner, P. A., Jochems, W. and van Buuren, H., 2007. Measuring perceived sociability of computer-supported collaborative learning environments. In: Computers & Education. 2007, 49(2),176-19. ISSN 1873-782X.
42. Kunreuther, H. and Pauly, M., 2018. Dynamic insurance decision-making for rare events: The role of emotions. In: Geneva Papers on Risk and Insurance-issues and Practice. 2018, 43, 335-355. ISSN 1468-0440. Available at: <https://doi.org/10.1057/S41288-017-0068-X>
43. Kunreuther, H. and Pauly, M., 2015. Insurance decision-making for rare events: The role of emotions. In: Social Science Research Network. 2015, 1-15. ISSN 1556-5068. Available at: Available at: <https://doi.org/10.3386/W20886>
44. Latan, H., 2018. PLS path modeling in hospitality and tourism research: The golden age and days of future past. In: Applying partial least squares in tourism and hospitality research. 2018, 53-83. Bingley: Emerald Publishing Limited. ISBN 978-1-78756-700-9.
45. Legate, A. E., Hair Jr, J. F., Chretien, J. L. and Risher, J. J., 2023. PLS‐SEM: Prediction‐oriented solutions for HRD researchers. In: Human Resource Development Quarterly. 2023, 34(1), 91-109. ISSN 1532-1096.
46. Liengaard, B. D., Sharma, P. N., Hult, G. T. M., Jensen, M. B., Sarstedt, M., Hair, J. F. and Ringle, C. M., 2021. Prediction: coveted, yet forsaken? Introducing a cross‐validated predictive ability test in partial least squares path modeling. In: Decision Sciences. 2021, 52(2), 362-392. ISSN 1540-5915.
47. Magno, F., Cassia, F. and Ringle, C. M., 2022. A brief review of partial least squares structural equation modeling (PLS-SEM) use in quality management studies. In: The TQM Journal. 2022, 36(5), 1242-1251. ISSN 1754-274X.
48. Manley, S. C., Hair, J. F., Williams, R. I. and McDowell, W. C., 2021. Essential new PLS-SEM analysis methods for your entrepreneurship analytical toolbox. In: International Entrepreneurship and Management Journal. 2021, 17, 1805-1825. ISSN 1555-1938.
49. McLean, G. and Osei-Frimpong, K., 2019. Chat now… Examining the variables influencing the use of online live chat. In: Technological Forecasting and Social Change. 2019, 146, 55-67. ISSN 1873-5509. Available at: <https://doi.org/10.1016/j.techfore.2019.05.017>
50. Meparishvili, K., 2016. Psychology of online customer behavior. In: Caucasus Journal of Social Sciences. 2016, 9(1), 73-86. ISSN 2960-9380. Available at: <https://doi.org/10.62343/cjss.2016.158>
51. Moore, S., 2018. Gartner says 25 percent of customer service operations will use virtual customer assistants by 2020. 2018. [online]. [cit. 2024-10-09]. Available at: <https://www.gartner.com/newsroom/id/3858564>
52. Munnukka, J. Talvitie-Lamberg, K. and Devdeep, M., 2022. Anthropomorphism and social presence in human-virtual service assistant interactions: The role of dialog length and attitudes. In: Computers in Human Behavior. 2022, 135, 107343. ISSN 1873-7692. Available at: <https://doi.org/10.1016/j.chb.2022.107343>
53. Nass, C. and Moon, Y., 2000. Machines and mindlessness: Social responses to computers. In: Journal of social issues. 2000, 56(1), 81-103. ISSN 1540-4560.
54. Ngai, E. W., Lee, M. C., Luo, M., Chan, P. S. and Liang, T., 2021. An intelligent knowledge-based chatbot for customer service. In: Electronic Commerce Research and Applications. 2021, 50. ISSN 1567-4223. Available at: <https://doi.org/10.1016/j.elerap.2021.101098>
55. Nitzl, C. and Chin, W. W., 2017. The case of partial least squares (PLS) path modeling in managerial accounting research. In: Journal of Management Control. 2017, 28, 137-156. ISSN 2191-477X.
56. Nitzl, C., Roldan, J. L. and Cepeda, G., 2016. Mediation analysis in partial least squares path modeling: Helping researchers discuss more sophisticated models. In: Industrial management & data systems. 2016, 116(9), 1849-1864. ISSN 1758-5783.
57. Nisha, P. Asokk, D. Prasanna, S. and Ansari, A., 2024. Investigating chatbot users’ e-satisfaction and patronage intention through social presence and flow: Indian online travel agencies (OTAs). In: Journal of Systems and Information Technology. 2024, 26(1), 89-114. ISSN 1758-8847. Available at: <https://doi.org/10.1108/JSIT-04-2023-0062>
58. Novak, T. P., Hoffman, D. L. and Duhachek, A., 2003. The influence of goal-directed and experiential activities in online flow experiences. In: Journal of Consumer Psychology. 2003, 13(1), 3-16. ISSN 1532-7663.
59. Nunamaker, J. F., Derrick, D. C., Elkins, A. C., Burgoon, J. K. and Patton, M. W., 2011. Embodied conversational agent-based kiosk for automated interviewing. In: Journal of Management Information Systems. 2011, 28(1), 17-48. ISSN 1557-928X. Available at: <https://doi.org/10.2753/MIS0742-1222280102>
60. Payne, J., Szymkowiak, A., Johnson, G. I., Robertson, P. and Henderson, R., 2012. Beyond embodiment and social presence: preferences for virtual assistant gender and clothing style. In: International Society for Presence Research Annual Conference. Philadelphia: Temple University, 2012. ISBN 978-0-9792217-5-0.
61. Pfeuffer, N., Benlian, A., Gimpel, H. and Hinz, O., 2019. Anthropomorphic information systems. In: Business & Information Systems Engineering. 2019, 61, 523-533. ISSN 1867-0202.
62. Qiu, L. and Benbasat, I., 2009. Evaluating anthropomorphic product recommendation agents: A social relationship perspective to designing information systems. In: Journal of management information systems. 2009, 25(4), 145-182. ISSN 0742-1222.
63. Reeves, B. and Nass, C., 1996. The media equation: How people treat computers, television and new media like real people and places. Stanford, CA: Center for the Study of Language and Information, 1996. ISBN 1-5758-6053-8.
64. Rogers, E. M. and Bhowmik, D. K., 1970. Homophily-heterophily: Relational concepts for communication research. In: Public Opinion Quarterly. 1970, 34(4), 523-538. ISSN 1537-5331. Available at: <https://doi.org/10.1086/267838>
65. Ringle, C. M., Wende, S. and Becker, J. M., 2024. Smart-PLS 4. Boenningstedt: Smart-PLS GmbH. 2024. [online]. [cit. 2024-10-09]. Available at: <http://www.smartpls.com>
66. Ringle, C. M., Sarstedt, M., Mitchell, R. and Gudergan, S. P., 2020. Partial least squares structural equation modeling in HRM research. In: The international journal of human resource management. 2020, 31(12), 1617-1643. ISSN 0958-5192.
67. Risher, J. and Hair Jr, J. F., 2017. The robustness of PLS across disciplines. In: Academy of Business Journal. 2017, 1, 47-55. ISSN 2381-313X.
68. Sah, Y. J. and Peng, W., 2015. Effects of visual and linguistic anthropomorphic cues on social perception, self-awareness, and information disclosure in a health website. In: Computers in Human Behavior. 2015, 45, 392-401. ISSN 1873-7692. Available at: <https://doi.org/10.1016/j.chb.2014.12.055>
69. Sarstedt, M., Hair Jr, J. F. and Ringle, C. M., 2023. PLS-SEM: indeed a silver bullet –retrospective observations and recent advances. In: Journal of Marketing Theory and Practice. 2023, 31(3), 261-275. ISSN 1069-6679.
70. Sarstedt, M., Hair, J. F., Pick, M., Liengaard, B. D., Radomir, L. and Ringle, C. M., 2022. Progress in partial least squares structural equation modeling use in marketing research in the last decade. In: Psychology & Marketing. 2022, 39(5), 1035-1064. ISSN 1520-6793.
71. Sarstedt, M. and Cheah, J. H., 2019. Partial least squares structural equation modeling using SmartPLS: a software review. In: Journal of Marketing Analytics. 2019, 7, 196-202. ISSN 2050-3318.
72. Sarstedt, M., Ringle, C. M., Smith, D., Reams, R. and Hair Jr, J. F., 2014. Partial least squares structural equation modeling (PLS-SEM): A useful tool for family business researchers. In: Journal of family business strategy. 2014, 5(1), 105-115. ISSN 1877-8585.
73. Schuetzler, R. M., Grimes, G. M. and Giboney, J. S., 2018. An Investigation of conversational agent relevance, presence, and engagement. In: Proceedings of the americas conference on information systems, New Orleans, 2018, 1-10. ISBN 978-0-615-55907-0.
74. Seeger, A. M., Pfeiffer, J. and Heinzl, A., 2018. Designing anthropomorphic conversational agents: Development and empirical evaluation of a design framework. In: Proceedings of the international conference on information systems, 2018. ISBN 978-0-9966831-7-3.
75. Sharma, P. N., Liengaard, B. D., Hair, J. F., Sarstedt, M. and Ringle, C. M., 2022. Predictive model assessment in PLS-SEM: extensions and guidelines: an abstract. In: Academy of Marketing Science annual conference. Cham: Springer Nature Switzerland, 2022, 83-84. ISBN 978-3-031-24687-6.
76. Shmueli, G., Sarstedt, M., Hair, J. F., Cheah, J. H., Ting, H., Vaithilingam, S. and Ringle, C. M., 2019. Predictive model assessment in PLS-SEM: guidelines for using PLSpredict. In: European journal of marketing. 2019, 53(11), 2322-2347. ISSN 1758-7123.
77. Wang, S., Cheah, J. H., Wong, C. Y. and Ramayah, T., 2023. Progress in partial least squares structural equation modeling use in logistics and supply chain management in the last decade: a structured literature review. In: International Journal of Physical Distribution & Logistics Management. 2023, 54(1). ISSN 0960-0035.
78. Wang, C., Teo, T. S. and Janssen, M., 2021. Public and private value creation using artificial intelligence: an empirical study of AI voice robot users in Chinese public sector. In: International Journal of Information Management. 2021, 61. ISSN 1873-4707. Available at: <https://doi.org/10.1016/j.ijinfomgt.2021.102401>
79. Wang, N., Johnson, W. L., Mayer, R. E., Rizzo, P., Shaw, E. and Collins, H., 2008. The politeness effect: Pedagogical agents and learning outcomes. In: International Journal of Human Computer Studies. 2008, 66(2), 98-112. ISSN 1095-9300.
80. Waytz, A., Gray, K., Epley, N. and Wegner, D. M., 2010b. Causes and consequences of mind perception. In: Trends in Cognitive Sciences. 2010, 14(8), 383-388. ISSN 1364-6613.
81. Wexelblat, A., 1998. Don’t make that face: A report on anthropomorphizing an interface. In: Intelligent Environments. 1998, 173(179), 98-02. ISSN 2199-4676.
82. Wünderlich, N. V. and Paluch, S., 2017. A nice and friendly chat with a bot: User perceptions of ai-based service agents. In: Proceedings of the International Conference on Information Systems, 2017. ISBN 978-0-9966831-5-9.
83. Xu, Y., Shieh, C.H., van Esch, P. and Ling, I. L., 2020. AI customer service: task complexity, problem-solving ability, and usage intention. In: Australasian Marketing Journal. 2020, 28(4), 189-199. ISSN 1839-3349. Available at: <https://doi.org/10.1016/j.ausmj.2020.03.005>
84. Zhou, T., Li, H. and Liu, Y., 2010. The effect of flow experience on mobile SNS users’ loyalty. In: Industrial Management & Data Systems. 2010, 110(6), 930-946. ISSN 1758-5783.
85. Zhou, T., Li, H. and Liu, Y., 2010. The effect of flow experience on mobile SNS users’ loyalty. In: Industrial Management & Data Systems. 2010, 110(6), 930-946. ISSN 1758-5783. Available at: <https://doi.org/10.1108/02635571011055126>
86. Zhou, T., 2013. An empirical examination of continuance intention of mobile payment services. In: Decision Support Systems. 2013, 54(2),1085-1091. ISSN 1873-5797.

Kľúčové slová/Key words

anthropomorphic conversational agent, flow, social presence, social cues, behavioral intention
antropomorfný konverzačný agent, tok, sociálna prítomnosť, sociálne náznaky, behaviorálny zámer

JEL klasifikácia/JEL Classification

M31

Résumé

Vplyv vnímaného antropomorfizmu konverzačného agenta (CA) na sociálnu prítomnosť, tok a zámer správania používateľov
Výskum sa uskutočnil s cieľom preskúmať vplyv sociálnej prítomnosti konverzačného agenta na zámer správania používateľov na základe vnímaného antropomorfizmu. Pomocou S-O-R modelu sme využili vzorku 329 účastníkov a výsledky získané metódou PLS-SEM potvrdzujú, že antropomorfizmus posilnený sociálnymi podnetmi významne zvyšuje úmysel správania. Okrem toho vzťah medzi antropomorfizmom a behaviorálnym zámerom je sprostredkovaný tokom a sociálnou prítomnosťou prostredníctvom čiastočnej mediácie. Tieto zistenia majú významné teoretické dôsledky pre pochopenie toho, ako antropomorfizmus podporuje stav online zážitkov.

Recenzované/Reviewed

7. November 2024 / 11. November 2024