Paper Type
Short
Paper Number
PACIS2025-1422
Description
The rise of AI-powered chatbots has transformed information-seeking by facilitating interactive and dialogical exchanges. Prior research largely focused on message delivery, while overlooking how chatbots guide users in identifying and articulating their information needs. This study examined the effect of anchored instructions and verbal embodiment in chatbot-assisted information seeking. Drawing on the mindset theory of action phases, we argue that a verbally embodied chatbot activates a deliberative mindset based on users’ perceptions of chatbot intentionality, and specified instructions congruent with the implemental mindset can improve individuals’ information-seeking outcomes. By contrast, a non-embodied chatbot activates an implemental mindset, and generalized instructions that are congruous with the mindset significantly ameliorate information-seeking outcomes. In addition, we posit the depth and breadth of information-seeking as mediators. Overall, this study contributes to the human-chatbot interaction literature by demonstrating how chatbot design facilitates identifying and articulating information needs.
Recommended Citation
Hong, Zeyuan (Stephen); Choi, Ben; and Sun, Yuan, "Framing Consumer Questions: Chatbots as the Search Intermediary in Digital Sales" (2025). PACIS 2025 Proceedings. 4.
https://aisel.aisnet.org/pacis2025/hci/hci/4
Framing Consumer Questions: Chatbots as the Search Intermediary in Digital Sales
The rise of AI-powered chatbots has transformed information-seeking by facilitating interactive and dialogical exchanges. Prior research largely focused on message delivery, while overlooking how chatbots guide users in identifying and articulating their information needs. This study examined the effect of anchored instructions and verbal embodiment in chatbot-assisted information seeking. Drawing on the mindset theory of action phases, we argue that a verbally embodied chatbot activates a deliberative mindset based on users’ perceptions of chatbot intentionality, and specified instructions congruent with the implemental mindset can improve individuals’ information-seeking outcomes. By contrast, a non-embodied chatbot activates an implemental mindset, and generalized instructions that are congruous with the mindset significantly ameliorate information-seeking outcomes. In addition, we posit the depth and breadth of information-seeking as mediators. Overall, this study contributes to the human-chatbot interaction literature by demonstrating how chatbot design facilitates identifying and articulating information needs.
Comments
HCI