Overview of analyzed articles from our systematic literature review
| Authors | Journal | Customer service context | Relational personalization constructs | Theories and concepts | Effect on | Key findings |
|---|---|---|---|---|---|---|
| Adam et al. (2021) | Electronic Markets | Customer service scenario in which participants were instructed to inquire with a chatbot about the possibility of using their debit card overseas in the United States | Small talk, empathy | Social response and commitment-consistency theory | User compliance | The results show customers are more likely to comply with a chatbot's request for service feedback when the chatbot has anthropomorphic qualities and maintains consistency. Furthermore, the study reveals that the effect of anthropomorphic design cues on user compliance is mediated by social presence |
| Crolic et al. (2022) | Journal of Marketing | Customers' interactions with a customer service chatbot in an international mobile telecommunications company | Pronouns, relational cues in current interaction | Expectancy violations | Customer satisfaction | If customers enter a chatbot-led service interaction while feeling angry, the use of chatbot anthropomorphism can have a detrimental impact on their satisfaction with the service, overall evaluation of the company, and their willingness to make future purchases. However, this negative effect is not observed in customers who are not feeling angry |
| Go and Sundar (2019) | Computers in Human Behavior | E-commerce site scenario about choosing a digital camera to purchase | Interactivity | Social presence theory | Humanness perceptions | The results indicate that a chatbot with low anthropomorphic visual cues can still provide a satisfactory level of personalization if it has a high level of message interactivity |
| Han et al. (2022) | Information Systems Research | A scenario in which customers engage with a chatbot because they are missing item from a delivery in the retail industry | Empathy | Emotional contagion and expectation–disconfirmation | Service evaluations | Positive emotion expressed by an AI agent (versus a human employee) is less effective in facilitating service evaluations because of a heightened level of expectation–disconfirmation |
| Haugeland et al. (2022) | International Journal of Human-Computer Studies | Customer service chatbot interaction in retail banking | Small talk, free flow | N/A | User perception of anthropomorphism and hedonic quality | The study discovered that topic-led conversations can enhance both anthropomorphism and hedonic quality in chatbot interactions. However, no such effect was observed in free text interactions, possibly because chatbots lack the necessary flexibility and adaptability |
| Kull et al. (2021) | Journal of Business Research | Text message on a smartphone from a travel agent chatbot | Empathy, relational history | Social response theory | Brand engagement | When chatbots initiate a conversation using a warm (vs competent) message, brand engagement increases. Brand–self distance mediates this effect, such that a warm (vs competent) initial chatbot message makes consumers feel closer to the brand |
| Li and Shin (2023) | Journal of Consumer Behaviour | Imaginary scenario of a conversation with Rolex's chatbot | Empathy | Expectancy violation theory | Brand status perception, perception of the appropriateness | Luxury brands might be better off not using emoticons in chatbot communications because it dampens the brand status perception due to perceived unexpectedness, which in turn decreases the perception of the appropriateness of the interaction with chatbots. However, this negative effect of luxury brand's use of emoticons in chatbot communication only exists for traditional luxury brands, not for masstige brands |
| Lou et al. (2022) | International Journal of Advertising | Online conversation between a customer and an online AI-powered chatbot or an online customer service employee of a fictitious clothing brand | Empathy | Schema theory | Patronage intention and consumer perception (competency and warmth) | Human employees are more competent and warmer than a chatbot. When a human employee expresses sympathy to the afflicted customer during the conversation, participants considered the employee to be more competent when he/she also exhibits contingency (vs no contingency) during the conversation, which in turn, elicited higher patronage intentions among participants |
| Naito et al. (2023) | International Journal of Social Robots | Customer trying to buy a new outfit for themselves by communicating with chatbot and human | Quantitative | N/A | Customer satisfaction | Research finds that while traditional customer service styles are best suited for human shopkeepers, bot shopkeepers using straight or data driven customer service styles increase customer satisfaction, make customers feel more informed, and feel more natural than when a human uses them |
| Nguyen et al. (2022) | Information Systems Frontiers | Chatbot data from a large electric power company and their customer service | Small talk, empathy, humor | Social presence and social cues | Information match | The study found that desirable outcomes, where customers obtain the information they requested, are associated with bi-directionality, a moderate level of social presence, and selected social cues used by both the chatbot and customers |
| Sands et al. (2021) | Journal of Service Management | Customer service encounter | Scripted | Social impact theory | Customer satisfaction and purchase intention | When human service agents use an educational script instead of chatbots, there is an increase in customer satisfaction and purchase intention. These effects are explained by emotion and rapport, as the proximity to a human service agent leads to emotional engagement and the development of rapport, which in turn influence service outcomes |
| Schanke et al. (2021) | Information Systems Research | Used clothing buy-back process | Humor | Social presence and social information processing theory | Transaction conversion | In this retail setting, anthropomorphism is beneficial for transaction outcomes, but that it also leads to significant increases in offer sensitivity |
| Shin et al. (2023) | International Journal of Consumer Studies | Interaction on the web site of a fictitious telecommunication company regarding an overcharged bill | Humor | Uncanny valley theory | Service satisfaction | The use of humor enhances service satisfaction when it is used by a chatbot but not when it is used by a human agent. The positive impact of chatbot humor is linked to increased perceptions of anthropomorphism and interestingness in the chatbot interactions. Both positive and negative chatbot humor can improve the interestingness of interactions, but socially appropriate humor (affiliative) is more likely to enhance service satisfaction compared to inappropriate humor (aggressive) |
| Song et al. (2022) | Journal of Retailing and Consumer Services | Purchase process through text-based interaction with customer service | Empathy | Social response theory, CASA paradigm and perceived communication quality | Consumers' adoption intention | Consumers have a higher communication capability perception for human beings than chatbots. Consumers' credibility perception of chatbots was lower at the time of purchase; however it has been noted that consumers' pre-purchase (browsing stage) trust in personalized recommendations from chatbots is higher. Strong/weak human interaction need will decrease/increase consumer perceptions of communication capability of chatbots |
| Tsai et al. (2021) | Journal of Research in Interactive Marketing | Online interaction with a RedBull chatbot on a website with a follow-up survey | Empathy, humor, pronouns | Social presence theory | User engagement, interaction satisfaction, brand likability | High social presence communication positively impacts consumer engagement satisfaction and brand likability. This effect can be boosted by chatbots' anthropomorphic profile design |
| Wei et al. (2022) | Information Processing and Management | Online text-based customer support log from a primary health insurance provider | Standardized | N/A | N/A | Summarize and report the characteristics of different message types and compare their usage in sessions with only human, AI, or both representatives |
| Weiler et al. (2022) | Electronic Markets | Customers' banking requests | Quantitative | Inoculation theory and the elaboration likelihood model | Users' decision to discontinue chatbot usage after a response failure | The research found that inoculation messages indicating a low performance level alleviate the negative effects of chatbots response failures on discontinuance. However, quantitative performance level information exhibits this moderating effect on users' central processing, while qualitative performance level information affected users' peripheral processing |
| Whang et al. (2022) | Journal of Business Research | Lipstick purchase interaction with chatbot | Pronouns | Interactivity Theory | Purchase intention | The results indicate that a higher level of personalized chatbot messages enhances purchase intention through a sense of ease and understanding of the product |
| Xu et al. (2022) | Frontiers in Psychology | Screenshots of customer-chatbot interactions | Empathy, small talk | Social cognition theory | Customer satisfaction | Two experimental studies reveal that using a social-oriented communication style boosts customer satisfaction. Warmth perception of the chatbot mediates this effect, while consumer attachment anxiety moderates these effects |
| Authors | Journal | Customer service context | Relational personalization constructs | Theories and concepts | Effect on | Key findings |
|---|---|---|---|---|---|---|
| Customer service scenario in which participants were instructed to inquire with a chatbot about the possibility of using their debit card overseas in the United States | Small talk, empathy | Social response and commitment-consistency theory | User compliance | The results show customers are more likely to comply with a chatbot's request for service feedback when the chatbot has anthropomorphic qualities and maintains consistency. Furthermore, the study reveals that the effect of anthropomorphic design cues on user compliance is mediated by social presence | ||
| Customers' interactions with a customer service chatbot in an international mobile telecommunications company | Pronouns, relational cues in current interaction | Expectancy violations | Customer satisfaction | If customers enter a chatbot-led service interaction while feeling angry, the use of chatbot anthropomorphism can have a detrimental impact on their satisfaction with the service, overall evaluation of the company, and their willingness to make future purchases. However, this negative effect is not observed in customers who are not feeling angry | ||
| E-commerce site scenario about choosing a digital camera to purchase | Interactivity | Social presence theory | Humanness perceptions | The results indicate that a chatbot with low anthropomorphic visual cues can still provide a satisfactory level of personalization if it has a high level of message interactivity | ||
| A scenario in which customers engage with a chatbot because they are missing item from a delivery in the retail industry | Empathy | Emotional contagion and expectation–disconfirmation | Service evaluations | Positive emotion expressed by an AI agent (versus a human employee) is less effective in facilitating service evaluations because of a heightened level of expectation–disconfirmation | ||
| Customer service chatbot interaction in retail banking | Small talk, free flow | N/A | User perception of anthropomorphism and hedonic quality | The study discovered that topic-led conversations can enhance both anthropomorphism and hedonic quality in chatbot interactions. However, no such effect was observed in free text interactions, possibly because chatbots lack the necessary flexibility and adaptability | ||
| Text message on a smartphone from a travel agent chatbot | Empathy, relational history | Social response theory | Brand engagement | When chatbots initiate a conversation using a warm (vs competent) message, brand engagement increases. Brand–self distance mediates this effect, such that a warm (vs competent) initial chatbot message makes consumers feel closer to the brand | ||
| Imaginary scenario of a conversation with Rolex's chatbot | Empathy | Expectancy violation theory | Brand status perception, perception of the appropriateness | Luxury brands might be better off not using emoticons in chatbot communications because it dampens the brand status perception due to perceived unexpectedness, which in turn decreases the perception of the appropriateness of the interaction with chatbots. However, this negative effect of luxury brand's use of emoticons in chatbot communication only exists for traditional luxury brands, not for masstige brands | ||
| Online conversation between a customer and an online AI-powered chatbot or an online customer service employee of a fictitious clothing brand | Empathy | Schema theory | Patronage intention and consumer perception (competency and warmth) | Human employees are more competent and warmer than a chatbot. When a human employee expresses sympathy to the afflicted customer during the conversation, participants considered the employee to be more competent when he/she also exhibits contingency (vs no contingency) during the conversation, which in turn, elicited higher patronage intentions among participants | ||
| Customer trying to buy a new outfit for themselves by communicating with chatbot and human | Quantitative | N/A | Customer satisfaction | Research finds that while traditional customer service styles are best suited for human shopkeepers, bot shopkeepers using straight or data driven customer service styles increase customer satisfaction, make customers feel more informed, and feel more natural than when a human uses them | ||
| Chatbot data from a large electric power company and their customer service | Small talk, empathy, humor | Social presence and social cues | Information match | The study found that desirable outcomes, where customers obtain the information they requested, are associated with bi-directionality, a moderate level of social presence, and selected social cues used by both the chatbot and customers | ||
| Customer service encounter | Scripted | Social impact theory | Customer satisfaction and purchase intention | When human service agents use an educational script instead of chatbots, there is an increase in customer satisfaction and purchase intention. These effects are explained by emotion and rapport, as the proximity to a human service agent leads to emotional engagement and the development of rapport, which in turn influence service outcomes | ||
| Used clothing buy-back process | Humor | Social presence and social information processing theory | Transaction conversion | In this retail setting, anthropomorphism is beneficial for transaction outcomes, but that it also leads to significant increases in offer sensitivity | ||
| Interaction on the web site of a fictitious telecommunication company regarding an overcharged bill | Humor | Uncanny valley theory | Service satisfaction | The use of humor enhances service satisfaction when it is used by a chatbot but not when it is used by a human agent. The positive impact of chatbot humor is linked to increased perceptions of anthropomorphism and interestingness in the chatbot interactions. Both positive and negative chatbot humor can improve the interestingness of interactions, but socially appropriate humor (affiliative) is more likely to enhance service satisfaction compared to inappropriate humor (aggressive) | ||
| Purchase process through text-based interaction with customer service | Empathy | Social response theory, CASA paradigm and perceived communication quality | Consumers' adoption intention | Consumers have a higher communication capability perception for human beings than chatbots. Consumers' credibility perception of chatbots was lower at the time of purchase; however it has been noted that consumers' pre-purchase (browsing stage) trust in personalized recommendations from chatbots is higher. Strong/weak human interaction need will decrease/increase consumer perceptions of communication capability of chatbots | ||
| Online interaction with a RedBull chatbot on a website with a follow-up survey | Empathy, humor, pronouns | Social presence theory | User engagement, interaction satisfaction, brand likability | High social presence communication positively impacts consumer engagement satisfaction and brand likability. This effect can be boosted by chatbots' anthropomorphic profile design | ||
| Online text-based customer support log from a primary health insurance provider | Standardized | N/A | N/A | Summarize and report the characteristics of different message types and compare their usage in sessions with only human, AI, or both representatives | ||
| Customers' banking requests | Quantitative | Inoculation theory and the elaboration likelihood model | Users' decision to discontinue chatbot usage after a response failure | The research found that inoculation messages indicating a low performance level alleviate the negative effects of chatbots response failures on discontinuance. However, quantitative performance level information exhibits this moderating effect on users' central processing, while qualitative performance level information affected users' peripheral processing | ||
| Lipstick purchase interaction with chatbot | Pronouns | Interactivity Theory | Purchase intention | The results indicate that a higher level of personalized chatbot messages enhances purchase intention through a sense of ease and understanding of the product | ||
| Screenshots of customer-chatbot interactions | Empathy, small talk | Social cognition theory | Customer satisfaction | Two experimental studies reveal that using a social-oriented communication style boosts customer satisfaction. Warmth perception of the chatbot mediates this effect, while consumer attachment anxiety moderates these effects |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.