Mind the Style: Impact of Communication Style on Human-Chatbot Interaction

Conversational agents increasingly mediate everyday digital interactions, yet the effects of their communication style on user experience and task success remain unclear. Addressing this gap, we describe the results of a between-subject user study where participants interact with one of two versions of a chatbot called NAVI which assists users in an interactive map-based 2D navigation task. The two chatbot versions differ only in communication style: one is friendly and supportive, while the other is direct and task-focused. Our results show that the friendly style increases subjective satisfaction and significantly improves task completion rates among female participants only, while no baseline differences between female and male participants were observed in a control condition without the chatbot. Furthermore, we find little evidence of users mimicking the chatbot’s style, suggesting limited linguistic accommodation. These findings highlight the importance of user- and task-sensitive conversational agents and support that communication style personalization can meaningfully enhance interaction quality and performance.

💡 Research Summary

The paper investigates how a chatbot’s communication style—friendly/supportive versus direct/task‑focused—affects user experience and performance in a controlled, goal‑oriented setting. The authors built a map‑based 2‑D navigation task where participants must locate an embassy from a train station. Three experimental conditions were used: (1) a friendly chatbot (NAVI‑f), (2) a direct chatbot (NAVI‑d), and (3) a control condition with no chatbot assistance. The two chatbots delivered identical navigation instructions; the only difference lay in tone, affective wording, and politeness markers.

A between‑subjects design recruited 180 university participants (balanced by gender). Participants were randomly assigned to one of the three conditions and completed the navigation task within a five‑minute window. Afterward, they answered a post‑task questionnaire measuring satisfaction, trust, perceived warmth, and perceived competence on 7‑point Likert scales. Objective performance metrics (completion rate and time) were logged, and the dialogue transcripts were analyzed for linguistic accommodation—specifically the frequency of friendly lexical items (e.g., “great”, “thanks”) versus direct lexical items (e.g., “move”, “turn”). Prior chatbot experience was also recorded to test moderation effects.

Statistical analysis (mixed‑ANOVA and chi‑square tests) revealed several key findings. First, the friendly chatbot significantly increased subjective satisfaction compared with the direct chatbot (mean difference ≈ 0.42, p < .01). This effect was driven almost entirely by female participants; men showed no meaningful difference between styles. Second, task completion rates were higher for the friendly style (78 %) than for the direct style (62 %) and the control (65 %). Again, the gender interaction was significant: women benefited from the friendly tone, whereas men’s performance was largely unaffected. Third, linguistic accommodation was minimal—participants mirrored the chatbot’s style in less than 5 % of their utterances, with no gender or experience differences. This suggests that in short, highly goal‑oriented exchanges, users prioritize task efficiency over social alignment, challenging the generality of Communication Accommodation Theory in such contexts.

The authors interpret these results through the lenses of Social Response Theory and expectancy‑based models. They argue that a warm, supportive tone reduces affective uncertainty for women, thereby enhancing both affective trust and instrumental success. For men, who tend to value efficiency and competence cues, a direct style is sufficient and a friendly style offers no additional benefit. Prior chatbot experience moderated the style effect: experienced users were more performance‑oriented and less swayed by affective phrasing.

Practical implications are clear. Designers of conversational agents should consider user demographics and task nature when selecting a communication style. For services targeting female users (e.g., health, education, customer support), incorporating friendly, supportive language can boost satisfaction and even improve task outcomes. In contrast, efficiency‑critical tools aimed at male or performance‑focused users may favor concise, direct phrasing. Moreover, because short interactions do not elicit strong linguistic accommodation, style personalization should focus on immediate affective impact rather than expecting users to adapt their language.

Limitations include the artificial, text‑only interface, the short interaction duration, and the focus on binary gender without exploring cultural or age differences. Future work should examine multimodal (voice, visual) agents, longer longitudinal usage, and adaptive style‑switching mechanisms that respond to real‑time user feedback.

In sum, the study provides robust empirical evidence that communication style is not merely a superficial design choice; it can materially affect both subjective experience and objective performance, especially for certain user groups. Tailoring chatbot tone to user characteristics and task demands emerges as a promising avenue for creating more inclusive and effective conversational AI.