Generative AI as a tool for bank valuation analysis

This study examines the potential of generative AI, specifically ChatGPT, in bank valuation using established financial models such as the Dividend Discount Model, Free Cash Flow to Equity, Excess Return, and Relative Valuation techniques. Through a four-stage evaluation, ChatGPT's performance is compared to human-driven analysis, highlighting its ability to process complex data, apply structured frameworks, and integrate nuanced assumptions. Results show high accuracy for large banks, with minimal deviations when explicit formulas and assumptions are included, though challenges persist for smaller banks and complex models. The findings demonstrate ChatGPT's potential as a complementary tool for financial analysts, capable of automating routine valuation tasks and improving efficiency. Its adaptability to established frameworks underscores its value in financial modeling and decision-making. This study provides practical findings into integrating AI in finance while identifying opportunities for further research, including hybrid human-AI approaches, diverse datasets, and ethical considerations in AI-driven financial analysis. • Structured prompts enable ChatGPT to replicate established bank valuation models. • Providing explicit formulas and assumptions improves the accuracy of AI valuations. • Errors are larger for small banks and for complex frameworks such as Excess Return. • Human oversight improves explainability, robustness, and suitability of results. • Generative AI can automate routine valuation tasks and enhance analyst productivity.