The author shares a week-long experience creating website copy using 95% AI, claiming the output was significantly better than what they could produce alone. They emphasize that a simple prompt often results in poor quality, which they term "AI-slop." Instead, a structured workflow was developed to guide the AI effectively. The process includes several steps, starting with establishing a core message and preparing a base of inputs that capture customer language and context. The first step involves preparing a positioning memo and brand voice guide. This foundation helps the AI understand how to operate within specific boundaries. Next, the author gathered extensive customer call transcripts, extracting insights into customer perspectives, pain points, and success stories. This data was transformed into Persona Briefs, which are critical for clarity in messaging. The third step involved creating a Website Styling Guide, which focuses on eliminating jargon and ensuring the copy resonates with real customer experiences. The last steps involved structuring the website using AI to determine the order of content and benefits. Finally, all documents and guidelines were combined to generate nearly complete copy. The author emphasizes that effective AI output relies heavily on the quality of input data. They argue that many users fail to invest adequate effort upfront, leading to mediocre results. By dedicating time to crafting detailed prompts, the author was able to create multiple iterations of high-quality copy efficiently.