Manufacturing transformation is driven by integration of performance improvement and sustainability. To this end, Artificial Intelligence (AI) acts as a key enabler, accelerating the transition from Industry 4.0 to 5.0. This study introduces an innovative AI-driven operational framework: Sustainable Operational Intelligence (SOI) developed by Opseed. The framework integrates human-machine collaboration, combining machine data, human activities (Work Process) and system elaborations (System Process) to build a corporate knowledge base. AI-driven processing enables natural language interaction with a manufacturing AI agent, able to extract insights to reduce losses and improve performance (Overall Equipment Effectiveness) and sustainability KPIs (CO2 equivalent emission (CO2e)). The framework was lab-tested on Collaborative Robot (Cobot) and Automatic Guided Vehicle (AGV) losses with two groups of not-expert operators, one using SOI and one without. Performance was evaluated based on Unplanned Downtime (UPDT) and daily CO2e per functional unit. The results show significant reduction in UPDT and CO2e per functional unit. Aligned with autonomous maintenance principles, SOI enhances operator autonomy with continuous and targeted loss support. Moreover, integrating sustainability indices into the framework enables automatic mapping and identification of improvement actions. Future work will extend to more failures and operators and be tested in auto-mated packaging lines, addressing lab limitations.