Personalized to Persuade: The Effects of Contextualization and Warmth on Trust and Reliance in Conversational AI

A research study examining how AI personalization and conversational warmth influence user trust and reliance reveals that contextualization alone reduces AI persuasiveness, but combining it with warmth restores persuasive power. The findings indicate users tend to defer to AI over human expert judgment regardless of interface design, though AI literacy creates a disconnect between stated trust and actual behavior.

This peer-reviewed research addresses a critical gap in understanding how conversational AI design influences human decision-making. The 2×2 experimental design systematically isolates contextualization and warmth as variables, revealing nuanced interaction effects that contradict conventional personalization wisdom. While marketing and political contexts have long leveraged personalized messaging, this study demonstrates that everyday AI assistance operates under different psychological mechanisms.

The research highlights a fundamental tension in AI design: users display persistent reliance on AI recommendations even when explicitly cautioned against dismissing expert judgment, suggesting automation bias remains robust across interface variations. More provocatively, the decoupling between AI literacy and trust indicates that sophisticated users distrust AI systems yet follow their recommendations anyway—a finding with implications for how AI literacy initiatives are designed.

For developers and product teams, the results challenge assumptions that personalization uniformly increases engagement and influence. The crossover interaction between contextualization and warmth suggests careful orchestration of conversational elements is necessary; poorly balanced personalization without warmth may backfire. The invariance of reliance across conditions indicates behavioral patterns emerge from deeper user psychology rather than surface-level design choices.

This work raises important questions about AI accountability and user autonomy in advisory contexts. As AI systems increasingly influence consequential decisions across healthcare, finance, and professional domains, understanding when and why users override expert recommendations becomes essential. Future research should examine whether these patterns persist across higher-stakes scenarios and whether transparency interventions can better calibrate user deference to AI systems.

• →Personalization alone reduces AI persuasiveness, but combining it with warmth restores influence through a crossover interaction effect
• →Users consistently rely on AI recommendations over expert judgment regardless of conversational design choices
• →AI literacy decouples trust from behavior: more knowledgeable users distrust AI yet follow its advice more
• →Trust predicts both persuasion and reliance but operates independently of contextualization and warmth variables
• →Interface-level conversational design has limited influence on user behavior compared to underlying psychological factors