Abstract

Swine disease surveillance is critical to the sustainability of globalagriculture, yet its effectiveness is frequently undermined by limitedveterinary resources, delayed identification of cases, and variability indiagnostic accuracy. To overcome these barriers, we introduce a novelAI-powered, multi-agent diagnostic system that leverages Retrieval-AugmentedGeneration (RAG) to deliver timely, evidence-based disease detection andclinical guidance. By automatically classifying user inputs into eitherKnowledge Retrieval Queries or Symptom-Based Diagnostic Queries, the systemensures targeted information retrieval and facilitates precise diagnosticreasoning. An adaptive questioning protocol systematically collects relevantclinical signs, while a confidence-weighted decision fusion mechanismintegrates multiple diagnostic hypotheses to generate robust diseasepredictions and treatment recommendations. Comprehensive evaluationsencompassing query classification, disease diagnosis, and knowledge retrievaldemonstrate that the system achieves high accuracy, rapid response times, andconsistent reliability. By providing a scalable, AI-driven diagnosticframework, this approach enhances veterinary decision-making, advancessustainable livestock management practices, and contributes substantively tothe realization of global food security.