Saying "please" to AI agents wastes compute power, according to Scott Lane, Founder and CEO of Speeki, a sustainability firm. Lane notes that his firm has more AI agents than humans, which forced an evaluation of its environmental impact. While companies cannot control how major tech giants run infrastructure, they can control hosting choices and prompt writing. Because large language models require more processing power the harder they work, phrasing directly determines a query's carbon impact. Lane says cutting out unnecessary pleasantries like "please" and "thank you" is a simple, immediate way to lower AI energy consumption. A 2025 Google technical paper revealed that a median text prompt consumes 0.24 watt-hours of energy, emits 0.03g of carbon dioxide and uses five drops of water. A study titled Green Prompt Engineering for Sustainable Generative AI by researchers at Boston University, the University of California and the University of Nottingham found that optimised prompting reduces LLM inference energy and carbon dioxide emissions by up to 48%. The paper demonstrates model-agnostic techniques that cut power consumption across systems like OpenAI's GPT-4o, Cohere and Mistral AI's Mistral 7B. Lane believes rapid advancements in equipment performance, such as advanced cooling technologies, will soon bridge the gap, and that AI will likely help produce new technologies to make using the tech more efficient and sustainable.