Publicação
The impact of the inherent failures of AI Chatbots on brand satisfaction: a pre and post-purchase comparison
| Resumo: | The integration of AI-based conversational agents has become a growing practice in customer support services, allowing companies to provide consumers with immediate responses, reduce operating costs, and serve as guides throughout the purchasing journey. However, the presence of failures in these systems raises questions about the impact that such occurrences may have on brand evaluations, with a particular focus on satisfaction. With this in mind, this study aims to understand the impact of failures inherent in AI chatbots on brand satisfaction at different stages of the consumer journey- pre and post-purchase. To this end, a mixed methodological approach was employed, including in-depth interviews and an experimental study. The qualitative results identified the existence of a risk area for the occurrence of failures- postpurchase- where the feeling of frustration is more intense due to the greater investment, time and expectations in relation to the brand. However, the quantitative findings do not support a statistically significant effect of the moment of failure (pre vs. post-purchase) on brand satisfaction, nor did it support the mediation hypotheses regarding brand perception and reputation. Overall, although the failure disrupts the service experience, consumers tend to attribute responsibility to limitations in the technological system rather than to the brand, which contributes to maintaining stable levels of satisfaction. Additionally, it was concluded that the variables of brand perception and reputation, rather than a mediation system, serve as interpretive lenses that allow the failure to be recognized and contextualized, mitigating or intensifying its effect. This study explores how the overall evaluation of the brand is relatively protected from isolated failures, reinforcing the importance of considering the experiential context and the normalization of AI (and its limitations) in customer support services. |
|---|---|
| Autores principais: | Janeiro, Carolina Ribeiro |
| Assunto: | AI chatbots Service failures Brand satisfaction Brand perception Brand reputation |
| Ano: | 2026 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso embargado |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | The integration of AI-based conversational agents has become a growing practice in customer support services, allowing companies to provide consumers with immediate responses, reduce operating costs, and serve as guides throughout the purchasing journey. However, the presence of failures in these systems raises questions about the impact that such occurrences may have on brand evaluations, with a particular focus on satisfaction. With this in mind, this study aims to understand the impact of failures inherent in AI chatbots on brand satisfaction at different stages of the consumer journey- pre and post-purchase. To this end, a mixed methodological approach was employed, including in-depth interviews and an experimental study. The qualitative results identified the existence of a risk area for the occurrence of failures- postpurchase- where the feeling of frustration is more intense due to the greater investment, time and expectations in relation to the brand. However, the quantitative findings do not support a statistically significant effect of the moment of failure (pre vs. post-purchase) on brand satisfaction, nor did it support the mediation hypotheses regarding brand perception and reputation. Overall, although the failure disrupts the service experience, consumers tend to attribute responsibility to limitations in the technological system rather than to the brand, which contributes to maintaining stable levels of satisfaction. Additionally, it was concluded that the variables of brand perception and reputation, rather than a mediation system, serve as interpretive lenses that allow the failure to be recognized and contextualized, mitigating or intensifying its effect. This study explores how the overall evaluation of the brand is relatively protected from isolated failures, reinforcing the importance of considering the experiential context and the normalization of AI (and its limitations) in customer support services. |
|---|