FinancialLLM

This project aims to build an AI-powered reporting tool that leverages Large Language Models (LLMs) to transform raw financial data into contextualized insights, executive commentary, and concise visualizations. The tool will reduce manual reporting workload, standardize output formats, and enhance the analytical depth of finance reporting. Students will use historical financial data from Roche—along with examples of the insights and reports currently provided by the Finance department to management—to design, prototype, and implement a solution that automates: ■ Data ingestion and aggregation ■ KPI and summary metric calculation ■ Data contextualization and business logic integration ■ Prompt generation for LLMs ■ Commentary and insight generation ■ (Nice to have) Assembly of visual and textual outputs into presentation-ready reports By automating the commentary, insight, and report generation process, this solution will enabling the Finance department to shift from time-consuming manual reporting tasks to more strategic, value-adding activities.