<p>This study develops and evaluates automated natural language processing (NLP) approaches for standardizing free-text offense information into structured offense codes. Using a combination of advanced transformer-based models, the analysis demonstrates the feasibility of producing reliable, scalable, and replicable offense auto-coding. The research describes the preprocessing steps including text normalization, tokenization, deduplication, and quality control that underpin model performance. Evaluation reveals that transformer-based approaches achieve robust accuracy, precision, and recall. These findings underscore the potential of pretrained models to handle the linguistic diversity and inconsistency that often characterizes justice system data, thereby addressing a long-standing barrier to high-quality research and policy analysis. The implementation of an automated offense coding system has direct implications for modernizing criminal justice data infrastructure. First, auto-coding enables greater consistency and comparability across jurisdictions, reducing errors that arise from fragmented coding practices. Second, by automating a labor-intensive process, it expands research capacity and lowers costs, making data more accessible for agencies and policymakers. Third, improved data standardization supports real-time evidence generation, aligning criminal justice data with the broader trend toward real-world evidence frameworks that have transformed decision-making in healthcare. Finally, adopting offense auto-coding provides a foundation for building interoperable, timely, and actionable data systems that can support evidence-based interventions and equitable policy reforms.</p>