The business team reports that the LLM responses generated by the customer assistant are overly verbose and repetitive. They want the output to be concise and action-oriented without changing the base model. How should the Generative AI Engineer modify the prompt to meet this requirement?