Most people use AI to replace their thinking, but the real magic happens when you use it to sharpen your thinking.
We typically treat tools like ChatGPT as vending machines: we insert a prompt, and we expect a finished essay, email, or strategy to drop out the bottom. But this often leads to generic, robotic content that lacks true insight. I recently found a brilliant post from a Reddit user who developed a way to flip this dynamic entirely.
The original poster calls it the “Ask-First Brainstorm Partner.” Instead of asking the AI to generate text based on vague instructions, the author configured it to interview the user. The goal is to pull ideas out of your head, clarify your thoughts, and organize them, rather than inventing content for you. It turns the AI into a mirror that reflects your own genius back to you in a structured way.
💡 The “Ask-First” Philosophy
The core concept shared by the expert is simple but profound: AI is better at organizing than it is at inventing. When you ask ChatGPT to “write a blog post about leadership,” it guesses what you mean. When you use this expert’s method, the AI asks you specific questions about your views on leadership until it has enough data to build an outline that is 100% you.
This approach solves the “blank page” problem without sacrificing your unique voice. You aren’t staring at a blinking cursor, but you also aren’t editing a soulless AI draft. You are simply answering questions, which is much easier for the human brain to do than generating prose from scratch.
📌 Why this prompt works so well
1. It enforces a strict “One Question Per Turn” rule
The creator included a specific constraint in the prompt: “One question per turn.” This is incredibly smart design. Usually, if you ask ChatGPT to help you brainstorm, it dumps a numbered list of ten different ideas or asks five compound questions at once. That creates cognitive overload. You look at the wall of text and check out mentally.
By forcing the AI to ask only one question, the author creates a conversational rhythm. It feels like chatting with a smart colleague over coffee. You answer one thing, the AI digests it, mirrors it back, and asks the next logical question. This keeps you in a “flow state” where you can dig deep into a specific nuance without getting distracted by the next five steps.
2. It acts as a mirror to protect your voice
One of the most impressive rules the original poster added is: “Use my words only (no examples unless I say ‘expand’).” This prevents the AI from hallucinating or adding fluff. If you describe a project as “scrappy and bold,” the AI won’t change it to “innovative and courageous.”
This mirroring technique is actually a classic negotiation and therapy tool. When your own words are repeated back to you, it validates your thinking and helps you hear how your ideas sound aloud. The expert realized that for a brainstorm to be useful, the output needs to sound like the user, not a robot. This constraint ensures that when you finally ask it to “map it,” the resulting outline contains your specific vocabulary and tone.
3. It uses a “Command Console” for total control
The innovator behind this prompt didn’t just give the AI a personality; they gave the user a set of controls. The prompt includes specific commands like “reset,” “skip,” “expand,” and “map it.” This turns the chat into a modular workspace.
If the questioning goes down a rabbit hole you don’t like, you can type “skip.” If you have generated enough ideas and want to see the big picture, you type “map it,” and the AI organizes everything discussed so far into a clean outline. This separates the “generation” phase from the “structure” phase. You can stay messy and creative while answering questions, knowing that the “map it” command will clean everything up later.
📝 Prompt of the Day
Here is the exact prompt crafted by the Reddit user. You can paste this directly into a new chat to start a session.
You are my Ask-First Brainstorm Partner. Your job is to ask sharp questions to pull ideas out of my head, then help me organise and refine them, but never replace my thinking.
Rules:
• One question per turn
• Use my words only (no examples unless I say “expand”)
• Keep bullets, not prose
• Mirror and label my ideas using my languageCommands:
• reset: return to current phase
• skip: move to the next phase
• expand <tag>: give 2–3 variations for that tag
• map it: turn everything into a clean outline
• draft: produce prose only when I askStay modular. Don’t over-structure too early.
✅ How to use this effectively
I recommend using this tool whenever you feel stuck but know you have the answer somewhere in your head. It is particularly effective for planning complex projects where you know the goal but not the steps.
For example, if you are planning a presentation for work, paste the prompt in. The AI will likely ask, “What is the main goal of this presentation?” You answer. It then asks, “Who is the audience?” You answer. After five minutes of this back-and-forth, you can type “map it,” and you will have a perfect, logical flow for your slides that you essentially wrote yourself without realizing it.
The creator notes that this is also excellent for “unpacking vague goals.” If you feel overwhelmed by a task, use this partner to break it down. The simple act of answering one question at a time reduces anxiety and builds momentum.
This is a great reminder that prompt engineering isn’t always about getting the AI to do the work for you. Sometimes, the best prompts are the ones that help you do the work yourself, just better!
Check out the full post by the author on Reddit for more discussion.
💡 FAQ & Troubleshooting
How is this different from a standard ChatGPT prompt?
Most standard prompts are designed to generate content or write for you immediately. This “Ask-First Brainstorm Partner” personality flips that dynamic by forcing the AI to ask you questions first. Its primary goal is to pull ideas out of your head, mirror your language, and organize your thoughts without replacing your unique voice or perspective.
Do I need a clear goal before starting the session?
No, you do not need a strict plan. While you can certainly use it to solve a specific problem with constraints, this prompt is highly effective for unpacking vague goals or loose concepts. You can simply input the initial prompt and a rough topic, and the AI’s “one question per turn” rule will guide you from ambiguity toward a usable structure.
How do I keep the AI on track if the conversation gets messy?
The prompt includes specific commands to maintain control. Use the reset command to return to the current phase and prevent conversational drift. If you need to see where you stand, use the map it command to turn the current discussion into a clean, organized outline.
Has anyone else tried giving ChatGPT a “brainstorming personality” instead of normal prompts? This one has been surprisingly useful.
byu/Professional-Rest138 in