The Ultimate AI Prompt Formula
A few of you may like this AI prompt framework I use across ChatGPT, Claude, Gemini etc (I use it for AI Agents and Claude Cowork instructions too.) Things like this... “Write me a post on the latest Skool building tips in 2026”, “Create a sales email for my book on cold outreach” or "Give me 12 tips for making money as a Skool owner". You'll have to enter so many corrections and updating prompts that you'll get fed up or give up. It’s not the LLM (Large Language Model). It’s the instruction. I had this myself early on. Same tool. Same prompt. Different results every time. Turns out I was basically giving it half a brief and expecting a full result. Once I fixed that, everything tightened up. The shift is simple. Stop treating your LLM like Google. Start treating it like someone you’ve hired. Give it: - a clear job - proper context - a standard to reach That alone changes the output. Here are 3 Value Tips for you: 1. What you feed it matters - Better inputs = better thinking - Tone, offer, audience and frameworks all shape the output - Mine’s trained on Alex Hormozi, Russell Brunson and Jeff Walker for a reason - Generic output usually comes from vague input 2. Keep it focused and simple - Pick a lane, content outreach offers - Frameworks do the heavy lifting - The win comes from refining, not reinventing 3. Ask better before you expect better - Get it to ask what it needs before it answers - Better questions = better outputs - Small tweaks to your prompt change everything Ways to use the prompt formula: - Drop it in as your default starting point instead of writing prompts from scratch - Attach it to any conversation and have the AI rewrite your rough prompt properly - Use it inside tools like Claude projects, ChatGPT custom GPTs or NotebookLM to train behaviour - Pair it with source material to shape thinking, not just outputs (e.g., books, notes, frameworks) - Turn rough ideas into structured prompts without overthinking - Use it as a feedback loop by asking what’s missing before it answers - Keep outputs consistent across different tools and tasks - Pressure-test your thinking by forcing clarity on role, task, specifics and context - Refine it over time and make it your own