Conversational agents and momentary user experience: an assessment using an electroencephalography device

This study investigates the momentary user experience (UX) during interactions with conversational agents, utilizing an electroencephalography (EEG) device alongside questionnaires. The aim is to examine the differences in user experience metrics between interactions through keyboard-based and voice-based inputs compared to a control group using a standard Google search. The experiment involved 36 participants divided into three groups. Group 1 used Google Search via a browser (control), Group 2 interacted with Google Assistant via keyboard, and Group 3 used voice commands with Google Assistant. All participants engaged in identical search tasks. EEG data captured metrics such as interest, excitement, engagement, relaxation, and stress. Participants also completed questionnaires to assess their subjective experiences after each task. The EEG results indicated variable engagement levels across the groups, with notable differences in stress and excitement metrics between interaction modalities. Group 3 (voice interaction) displayed greater fluctuation in engagement and excitement levels compared to Groups 1 and 2. The questionnaire data supported the EEG findings, suggesting differing levels of user satisfaction and task difficulty across the groups. Voice interaction was associated with higher engagement but also higher variability in emotional response. The modality of interaction with conversational agents influences momentary UX. These findings suggest that while voice-based interfaces may enhance engagement, they also require careful design to manage user emotions and expectations effectively. Further research is recommended to explore these interactions across a broader range of user demographics and settings.