Was I All Ears? A Comprehensive Evaluation of Voice Bot Service
摘要
This study conducted a comprehensive survey to develop a validated and reliable questionnaire for evaluating voice bot services. Additionally, we investigated domestic credit card companies providing voice bot services, referred to as callbots, to compare their initial approaches. In Experiment 1, the reliability and validity of design principles and strategies were confirmed, with findings indicating that design principles are more critical than design strategies. Building on these results, Experiment 2 aimed to capture the first impressions of the actual callbots used by credit card companies. While no significant differences in usefulness were observed among three companies, significant differences were found in learnability and satisfaction. Post-hoc analysis revealed that these differences were linked to realistic voice responses and appropriate guidance in scenarios of user non-response and no-match within the design principles. Regarding design strategies, substantial differences were found in the selectable options available when the callbot required clarification. In conclusion, to enhance callbot usability, it is recommended to establish a realistic persona and design features that clearly handle cases where users deviate from the normal flow.