The development of non-oil exports has become a strategic priority for Iran due to vulnerability associated with oil dependence and the need for sustainable economic growth. Within this framework, the food manufacturing sector, as a key component of the agricultural value chain, possesses considerable potential for value creation, employment generation, and integration into regional and global markets. However, despite significant production capacities and geographical advantages, Iran’s food industry has not fully achieved its export potential, highlighting the need for effective industrial transformation pathways. This study develops an artificial intelligence-based framework to assess transformation pathways in Iran’s food industry by integrating historical evidence with expert knowledge. Machine learning models, including Random Forest and XGBoost, were applied using annual time-series data from 2000 to 2022 to identify major determinants of food industry export performance. The predictive results were interpreted through Explainable Artificial Intelligence (SHAP) to reveal the contribution and direction of influential factors. Furthermore, an expert-based Self-Organizing Map (SOM) clustering approach was employed using perceptions of 150 industrial experts to identify feasible transformation priorities. The findings indicate that exchange rate volatility, policy uncertainty, and insufficient international market capabilities constitute major constraints to export competitiveness. The integrated AI framework identifies four strategic transformation pathways: productivity-oriented industrial upgrading, export market and branding development, institutional and policy transformation, and capability-based technological advancement. The study provides an evidence-based roadmap for policymakers to enhance the competitiveness, resilience, and global integration of Iran’s food manufacturing sector.