Abstract
This study examines conversational business analytics, an approach thatutilizes AI to address the technical competency gaps that hinder end users fromeffectively using traditional self-service analytics. By facilitating naturallanguage interactions, conversational business analytics aims to empower endusers to independently retrieve data and generate insights. The analysisfocuses on Text-to-SQL as a representative technology for translating naturallanguage requests into SQL statements. Developing theoretical models groundedin expected utility theory, the study identifies conditions under whichconversational business analytics, through partial or full support, canoutperform delegation to human experts. The results indicate that partialsupport, focusing solely on information generation by AI, is viable when theaccuracy of AI-generated SQL queries leads to a profit that surpasses theperformance of a human expert. In contrast, full support includes not onlyinformation generation but also validation through explanations provided by theAI, and requires sufficiently high validation effectiveness to be reliable.However, user-based validation presents challenges, such as misjudgment andrejection of valid SQL queries, which may limit the effectiveness ofconversational business analytics. These challenges underscore the need forrobust validation mechanisms, including improved user support, automatedprocesses, and methods for assessing quality independently of end users'technical competencies.