MachineSage AI: Evidence-Based Maintenance Copilot

MachineSage AI is an evidence-based industrial maintenance copilot designed to help factory technicians investigate machine incidents more efficiently. Industrial teams often need to interpret sensor readings, error reports, and technical documentation before deciding on the next troubleshooting step. The application combines deterministic Python-based sensor checks with an AI-powered explanation workflow. Users can enter a machine incident, inspect abnormal readings, review relevant technical manual excerpts, and receive a structured report containing possible causes, evidence, suggested checks, and uncertainty warnings. The system is designed to distinguish observed readings from AI-generated interpretations. Recommendations should be supported by available evidence, and uncertain or potentially high-risk cases are flagged for human review rather than treated as confirmed diagnoses. The prototype will integrate an AMD-hosted AI model so that AMD infrastructure contributes directly to the application's core inference workflow. Its evaluation will use controlled example incidents to examine response relevance, evidence support, uncertainty handling, and processing time. MachineSage AI aims to demonstrate how AI can make industrial troubleshooting information more accessible while keeping technicians involved in important decisions.