AI technologies will allow us to achieve greater autonomous operations maturity for industrial plants for Energy, Power and Chemicals. Deep Reinforcement Learning [DRL] Agents are one enabling AI technology that transitions from L3 to beyond L4 control autonomy. AVEVA is colaborating with NVIDIA for DRL Agents for Autonomous Operations. Bringing together AVEVA and NVIDIA’s expertise agents are trained using accurate digital process twins at scale using high speed compute architecture. This presentation will focus on a specific control problem for ADNOC to be able to better control feed dsiturbances into an oil separation train. The objective is to train a reinforcement learning controller to minimize operator manual interventions and to maximise the utilization of the existing hardware.