Rethinking Interaction with Conversational Agents:How to Create a Positive User Experience Utilizing Dialog Patterns

Conversational agents (CAs) are increasingly used as an additional convenient and innovative customer service channel to relieve service employees, as in the studied organization. In the process of analyzing and maintaining the present AI-based agent, however, user satisfaction is low as the CA lacks understanding and offers unsatisfactory solutions to users. Nonetheless, solving the requests and providing a positive user experience is crucial to relieve the service employees’ workload permanently. For CAs’ improvement, this study followed action design research (ADR) and used design thinking. We identified the central interaction problems (findability, welcome message, dialog control and fallback issues) with a monitoring process and analysis. Afterward, we interviewed users about their expectations and requirements and addressed these problems by creating user-centric mock-ups. Through a quantitative survey, the most popular solutions were implemented in a prototype. Finally, the resulting CA prototype was evaluated, showing a significantly improved user experience afterward, and design guidelines were discovered.