Evaluation of AI-driven GL code mapping accuracy across Fortune 100, 250, and 500 enterprise datasets.