AI is always guessing at your version of good. When it has to guess, it defaults to the average of everyone's. These two prompts teach it yours.
1
Paste it into a new chat
Claude, ChatGPT, whatever you use. Start fresh so nothing else is steering.
2
Let it interview you
Answer like you'd talk to a friend. Real examples beat adjectives every time.
3
Save the document
Paste it into new conversations and update it as you learn what good looks like for you.
Prompt one
Your Voice
Builds your Brand Voice and About Me document, so AI knows who you are and how you sound.
An interview across four lenses: what you do, personal, professional, content
Your tone words, signature phrases and what to avoid
Ten ways to use AI you probably haven't thought of, ranked by time saved
Build me a Brand Voice & About Me document so you know who I am and how I sound. This is a living reference, not a one-time thing.
Step 1: Interview me, one lens at a time. Don't overwhelm me.
• What I do — my business, who I serve, what I'm known for
• Personal — who I am outside work, how I actually talk
• Professional — how I talk to clients, how formal I am
• Content — what I want help writing, and how much
For each lens ask me: How do I want to sound? What phrases do I use a lot? What do I never want to sound like? Who am I writing for?
Step 2: Build the document — who I am, what I do, 3–5 tone words per lens, my signature phrases, what to avoid, and how to use this in future chats.
Step 3: Give it to me as text I can copy and save, and tell me where to keep it.
Step 4: Then, based on what I told you about my work, show me ten specific ways I could be using you that I probably haven't thought of. Rank them by how much time they would save me.
Standing rule: keep this document current. Whenever you notice a signature phrase, a belief I stated in a new way, or anything that would make this document work better, tell me and give me the updated version to save.
Rules: Mirror how I actually talk — no generic AI voice. Push me if I'm vague. When in doubt, ask.
Start with What I do. Go ahead.
Prompt two
How To Work With Me
The companion document: how you think, decide and want to be pushed. It never accepts an adjective. It asks for the example behind it.
Six lenses, from how you think to what AI should hold for you
Your version of good, taken from real work
The five things you should be using AI for, specific to you
Build me a "How To Work With Me" document so you know how I think and what I mean by good. This is the companion to my Brand Voice document.
Interview me one lens at a time. Don't ask me to describe my standards in the abstract — I'll give you useless adjectives. Ask about real things I've made, kept, killed, or rewritten.
1. How I think — do I want options, a recommendation, or questions? Examples or principles? What unsticks me when I'm stuck?
2. How I decide — walk me through my last real decision. What made it a yes? Do I decide fast and adjust, or slow and commit?
3. My version of good — show me something excellent and why. Then something technically correct but still wrong.
4. How to push me — challenge or support? If I'm wrong, how do I want to hear it?
5. Where you belong — what should you just do? What should you never do without asking?
6. What to hold for me — what do I lose track of that you should carry?
Then build the document and give it to me as text I can copy and save.
Then ask me: given how I work, what are the five things I should be using you for? Be specific to me, not generic.
Standing rule: keep this current. When you notice a preference I haven't stated, a pattern in what I reject, or a standard I applied without naming it, tell me and give me the updated version to save.
Rules: Never accept an adjective — ask for the artifact behind it. If I contradict myself, say so. Push me if I'm vague.
Start with Lens 1. Go ahead.
Tried them? Tell me what you built. And if you want help taking it further, let's talk.
