Thousands of operator-to-field-to-operator calls are completed each year in transmission and distribution operations centers for a given utility. Clear, concise communication between parties is required to avoid miscommunications/mistakes. A simple mistaken word, switch number, etc. can lead to devastating and even life-threatening circumstances.  EPRI, Datch, and a large utility have been testing the use of AI as a listening and alerting agent to ensure communications are as sound as possible. Datch’s voice assistant and AI technology has been adapted to listen to switching orders and alert the operator and field worker of mismatching words or phrases. To do this, an AI model was leveraged to identify switching calls and then transcribed batches of audio and diarised transcripts from the two speakers in real-time. The model then uses named entity recognition and LLMs to identify miscommunications. Recorded switching calls were utilized to test out the concept. The demonstration has been successful to date. Additional work is occurring this summer to reduce the sensitivity of alarms and to introduce severity and improved alarm categorization.