An AI-Based Voice Commentary System for Dashboard Interpretation

Abstract

In today’s data-driven world, dashboards are widely used to visualize analytical data and support decision-making. However, understanding complex dashboards can be difficult for non-technical users because the information is often dense and multi-dimensional. To address this challenge, this paper presents an enhanced Voice Commentary Generation System (VCGS) that automatically converts dashboard data into simple, human-like audio explanations in real time. The proposed system uses a modern web-based architecture with a React and TypeScript frontend, Supabase backend services, and AI-powered Natural Language Generation (NLG) with Text-to-Speech (TTS) technology. Unlike traditional systems that depend on manual analysis or text summaries, the proposed approach generates meaningful insights automatically and delivers them through voice narration. Experimental results show that the system generates accurate insights within 5–7 seconds per dashboard. It improves accessibility and helps users easily understand complex data through AI-based voice interaction.