Abstract

The highly competitive environment among Conversational Generative AI (GenAI) applications has shifted management focus toward user-perceived quality, rather than AI benchmark scores. However, existing information systems (IS) quality literature does not fully capture how users perceive the qualities of these conversational GenAI applications. To address this issue, we employ the grounded theory analysis of user reviews to inductively develop a conceptual framework of perceived conversational GenAI quality, comprising four quality dimensions, system, information, reasoning, and service, along with thirty subdimensions. Crucially, we introduce reasoning quality as a novel dimension to capture users’ unique evaluative criteria in conversational GenAI contexts. We further identify how subdimensions cluster within each quality. These contributions extend existing IS quality literature by offering a user-centered perspective on conversational GenAI quality and establishing a foundation for future measurement development.

Download

Share

COinS