Most AI tools give you a tidy little answer and call it done. Short, clean, confident. The trouble is that tidy is usually shallow. The model can go deep, it just defaults to fast and concise, so it compresses its own reasoning before you ever see it. You get the surface. You wanted the thinking.
I got tired of that, so I started ending my prompts with the same three lines every time. Now I just paste them in.
Expand every point with practical details, implications, risks, tradeoffs, and supporting context. Do not summarize, compress, or skip sections. Address each point individually and preserve the original structure. Be thorough, specific, and actionable while avoiding filler or repetition.
It works because it cancels the default. Instead of collapsing six ideas into one neat paragraph, the model has to walk each one out, show the tradeoffs, and say the part it would normally bury. Run it once on something you know well and you'll watch it surface a point you'd have skimmed past.
What I've learned using it
- Reach for it at the front of a task, when you don't yet know what you're looking for. That's when the depth pays off.
- Skip it when you already know the shape of what you want. Ask for three bullets and it'll just bloat them.
- Treat what comes back as raw material. You're going to cut most of it, and that's fine.
- The goal is better thinking, not better writing. You drive; the model keeps up.
When a document is long or dense enough to be worth a heavier push, I swap in a longer version that asks for examples and tells it to protect nuance. Same idea, more horsepower.