Examining the Impacts of Task Interval on Subjective Workload and Operational Behavior in a Multi-Conversational Agent Control System

Abstract

This paper explores the effect of task interval within a multi-conversational agent control system on the perceived workload and operational behaviors of administrators. An experimental study was conducted employing three distinctive task intervals, namely 10s, 30s, and 60s, with a sample of 39 participants. The results indicated a direct correlation between shorter task intervals and increased subjective workload of administrators, which correspondingly resulted in expedited response times. On the contrary, longer intervals were associated with enhanced response quality and more meticulous behavior. Eye-tracking data analysis further provided insights into user engagement; shorter intervals heightened focus on chat content, while longer intervals led to an increase in distractions. The findings underscore the significance of task intervals as a pivotal mechanism for optimization of multi-conversational agent control systems and enhancement of conversation quality. For future work, we plan to investigate strategies aimed at mitigating administrators