Rapport-Driven Virtual Agent: Rapport Building Dialogue Strategy for Improving User Experience at First Meeting

TL;DR

The paper addresses how to cultivate rapport in first-meeting human–agent interactions by introducing a rapport-building dialogue strategy that embeds targeted utterances into LLM-driven small talk. It compares a free-form prompting approach against a predefined sequence and evaluates their effects on user experience using the ERICA multimodal virtual agent. Key findings show that the free-form rapport-building strategy yields higher naturalness, satisfaction, engagement, and usability than the predefined approach and a Q&A baseline, with Rapport Score correlating positively with several UX metrics ($\rho$ values reported) but not with dialogue length. The work demonstrates that adaptive, rapport-focused small talk can meaningfully enhance early-stage human–agent interactions, highlighting directions for future work in emotional prosody and non-verbal cues to further strengthen rapport.

Abstract

Rapport is known as a conversational aspect focusing on relationship building, which influences outcomes in collaborative tasks. This study aims to establish human-agent rapport through small talk by using a rapport-building strategy. We implemented this strategy for the virtual agents based on dialogue strategies by prompting a large language model (LLM). In particular, we utilized two dialogue strategies-predefined sequence and free-form-to guide the dialogue generation framework. We conducted analyses based on human evaluations, examining correlations between total turn, utterance characters, rapport score, and user experience variables: naturalness, satisfaction, interest, engagement, and usability. We investigated correlations between rapport score and naturalness, satisfaction, engagement, and conversation flow. Our experimental results also indicated that using free-form to prompt the rapport-building strategy performed the best in subjective scores.

Figures (2)

• Figure 1: Comparative analysis between Free Rapport Agent and Predefined Rapport Agent.
• Figure 2: Comparative analysis between Free Rapport Agent and Q&A Agent.