earticle

영어

This study proposes a prompt structure based on generative AI to optimize business data analysis reporting, focusing on performance marketing. By systematically analyzing datasets containing various Key Performance Indicators (KPIs) critical in performance marketing, we develop a prompt that effectively extracts and summarizes key data. It emphasizes constructing 'Table-to-Text' prompt patterns, selecting the most suitable prompts through trial and error, and evaluating their functionality against other prompts. The study also uses ROUGE and BERT models to assess the similarity between generated comments and target comments, finding significant effectiveness in capturing nuanced semantic similarities. The proposed prompt structure aims to streamline data analysis and report generation in professional settings, reducing reliance on subjective interpretations and enhancing report objectivity and efficiency. This approach is expected to significantly improve the efficiency of data-based report writing tasks in business environments.