QUERY AI: AN AI-POWERED DATA QUERY INTERFACE

Abstract

This paper compares two AI-powered systems for SQL query generation: WrenAI and QueryAI. Both models use advanced lan- guage models to convert natural language input into SQL queries, but they differ significantly in their design and performance. WrenAI uses the OpenAI API, which is a flexible and general-purpose solution that can handle a variety of natural language processing tasks, including SQL query generation. On the other hand, QueryAI uses the Gemini API, which is a more specialized model, optimized for working with struc- tured data, especially SQL databases. This paper discusses the strengths and weaknesses of each approach, highlighting the role of prompt engi- neering and model specialization in improving the accuracy and effi- ciency of SQL query generation. QueryAI shows better performance in generating accurate and optimized SQL queries, while WrenAI is more versatile across different use cases. The comparative assessment com- prises of query accuracy, precision, recall, and F1-score as metrics, thus pointing out the trade-off flexibility against specialization in AI- generated queries. The findings shed light on selecting appropriate AI models for SQL-based applications for the choice between general-pur- pose and domain-specific models.