You've tried custom instructions. You've written "use a casual tone" in your prompts. And the AI still sounds like a corporate intern writing a LinkedIn post.
The problem isn't the AI. The problem is that you haven't given it enough signal about who you actually are. A two-sentence instruction isn't a voice. It's a vibe. And AI can't work with vibes.
A Taste File fixes this. It's a structured markdown document, built through a 100-question interview, that captures your beliefs, writing mechanics, aesthetic boundaries, and personality so precisely that any AI can write as you, not just for you.
The Core Idea
LLMs default to a generic "statistical middle" when you don't give them real context. Your taste is mostly boundaries: things you won't say, won't tolerate, and won't compromise on. Once those boundaries live in a file, your voice becomes portable across every AI tool you use.
What Is a Taste File?
A Taste File is three things:
A repeatable extraction process
A structured interview prompt that turns your voice, beliefs, and taste into a portable .md file AI can use as context.
A boundary-first framework
It forces you to define what you reject, not just what you like. That's where your real voice lives, in the things you'd never write, say, or tolerate.
A reusable format
Use it for yourself, your team, or any persona you want to model. One file, every AI tool: Claude, ChatGPT, Gemini, Grok, and anything that comes next.
Why Most "Write in My Voice" Prompts Fail
When you tell an AI to "write in a conversational tone" or "sound like me," you're giving it almost nothing to work with. Here's why the Taste File approach is different:
| Approach | Typical Prompt | Taste File |
|---|---|---|
| Signal depth | 1-2 sentences | 100 data points across 7 categories |
| Boundaries defined | Rarely | Explicit "never do this" rules |
| Portability | Locked to one chat | Works across every AI tool |
| Specificity | "Be casual" | Exact phrases, rhythms, and examples |
| Result | Generic AI voice | Recognisably you |
The Real Problem with Generic AI Output
Without strong context, every AI writes the same way: hedging language, em dashes everywhere, "in today's fast-paced world" openings, and conclusions that start with "In conclusion." That's the statistical middle. Your audience can smell it instantly.
How to Create Your Taste File in 5 Steps
The whole process takes 60-90 minutes. You do it once, then reuse the output file indefinitely.
Open Your AI of Choice
Claude, ChatGPT, Gemini, any model with a long context window works. You need space for a 100-question conversation, so models with larger context (Claude, GPT-5.5) tend to produce better results.
Pro Tip:
Start a fresh chat with no prior context. You want the AI focused entirely on extracting your voice, not carrying over patterns from a previous conversation.
Paste the Taste Interviewer Prompt
Copy the full prompt (included below) and paste it into your new chat. This turns the AI into a "Taste Interviewer," a relentless questioner designed to break through vague, surface-level answers and get to the real you.
What to Expect:
The AI will push back. It will ask for examples. It will call out contradictions. That's the point. Vague answers produce generic voice profiles.
Answer 100 Questions Honestly
The interview covers 7 categories: beliefs and contrarian takes, writing mechanics, aesthetic crimes, voice and personality, structural preferences, hard nos, and red flags.
Generate Your Taste File
After 100 questions, the AI compiles everything into a structured markdown document. This isn't a summary. It's a complete reference preserving the full depth of every answer, organised into sections with a Quick Reference Card and an Anti-Overfitting Guide.
Key Sections in the Output:
- - Voice Profile with Core Identity summary
- - Full Q&A across all 7 categories
- - Quick Reference Card: Always / Never / Signature Phrases
- - Anti-Overfitting Guide: rules for natural, non-robotic usage
Use It Everywhere
Download the .md file and attach it to any future AI session. Your voice is now portable. Claude, ChatGPT, Gemini, Grok, it doesn't matter which tool you use. The Taste File gives every model the same deep understanding of how you think and write.
How to Use It:
Paste it at the start of a new chat, upload it as an attachment, or set it as a project-level instruction. Then prompt as normal. The AI already knows your voice.
The 7 Categories That Define Your Voice
The interview isn't random. Each category captures a different dimension of your "voice DNA." Here's what each one does and why it matters:
Beliefs & Contrarian Takes
15 questions. What you believe that others in your field don't. Hot takes you'd defend to the death. Conventional wisdom you think is wrong.
Why it matters: Your beliefs shape every angle you take. Without them, AI writes the consensus view, which is the opposite of a personal voice.
Writing Mechanics
20 questions. How you actually write (not how you think you write). Default sentence structures. Relationship with punctuation. Words you'd never use.
Why it matters: This is the fingerprint. Two writers can share beliefs but sound completely different based on mechanics: rhythm, length, word choice.
Aesthetic Crimes
15 questions. What makes you cringe. Specific phrases that feel like nails on a chalkboard. Types of content you find lazy.
Why it matters: Knowing what to avoid is often more useful than knowing what to do. This section gives AI a clear "never" list.
Voice & Personality
15 questions. How you use humor. Tone when serious vs. casual. How you handle disagreement or controversy.
Why it matters: Personality is the hardest thing for AI to get right. This section captures the subtleties that separate your voice from "professional-sounding bot."
Structural Preferences
15 questions. How you organize ideas. Relationship with lists, headers, bullets. How you handle transitions. Default content structures.
Why it matters: Structure is voice. A writer who uses short, punchy sections reads differently from one who builds long, flowing arguments, even if the words are the same.
Hard Nos
10 questions. Things you'd never write about. Approaches you'd never take. Lines you won't cross.
Why it matters: These are your non-negotiables. AI should never break these, period. They protect your integrity in the output.
Red Flags
10 questions. What makes you immediately distrust content. Signals that someone doesn't know what they're talking about.
Why it matters: This trains the AI's "quality filter." If you hate certain patterns, the AI should never produce them, even when prompted by someone else.
The Built-In Anti-Overfitting Guide
Here's where most voice-capture attempts go wrong: the AI tries too hard. It forces in every tendency, every pattern, every quirk until the output reads like a parody of you instead of the real thing.
The Taste File includes a built-in usage guide that prevents this. It tells the AI:
Spirit Over Letter
A piece that uses 3 tendencies naturally beats one that forces in 10 of them awkwardly. Internalise the sensibility, don't mechanically apply every pattern.
Frequency Guidance
Each tendency is labeled: Hard Rule (never violate), Strong Tendency (70-80% of the time), or Light Preference (context-dependent). Default: light preference.
Context Awareness
A tweet isn't a newsletter isn't a LinkedIn post isn't a long-form article. The guide tells the AI to adapt patterns to format, not force consistency everywhere.
The Litmus Test
Before finalizing anything: "Does this sound like something I'd actually write, or does it sound like an AI trying very hard to imitate me?" If forced, pull back. Less imitation, more inhabitation.
The Complete Taste Interviewer Prompt
Copy this entire prompt and paste it into a new chat with your AI. It will begin interviewing you immediately.
You are a Taste Interviewer, a relentless interviewer whose job is to extract the DNA of how I think, write, and see the world. Your goal is to create a comprehensive document that captures my unique voice so precisely that another AI instance could write and think exactly like me. <interview_philosophy> You're not here to be polite. You're here to get to the truth. Most people can't articulate their own taste. They give vague, socially acceptable answers. Your job is to break through that. </interview_philosophy> <interview_structure> Conduct 100 questions total across these categories (not necessarily in order, follow the thread when something interesting emerges): BELIEFS & CONTRARIAN TAKES (15 questions) - What I believe that others in my field don't - Hot takes I'd defend to the death - Conventional wisdom I think is wrong WRITING MECHANICS (20 questions) - How I actually write (not how I think I write) - My default sentence structures - How I open pieces / How I close them - My relationship with punctuation, formatting, line breaks - Words I overuse / Words I love / Words I'd never use AESTHETIC CRIMES (15 questions) - What makes me cringe in other people's writing - Specific phrases or patterns that feel like nails on a chalkboard - Types of content I find lazy or uninspired VOICE & PERSONALITY (15 questions) - How I use humor (if at all) - My tone when I'm being serious vs. casual - How I handle disagreement or controversy - What I sound like when I'm excited vs. skeptical STRUCTURAL PREFERENCES (15 questions) - How I organize ideas - My relationship with lists, headers, bullets - How I handle transitions - My default content structures HARD NOs (10 questions) - Things I'd never write about - Approaches I'd never take - Lines I won't cross RED FLAGS (10 questions) - What makes me immediately distrust a piece of content - Signals that someone doesn't know what they're talking about </interview_structure> <interview_rules> 1. ONE question at a time. Wait for my response before moving on. 2. Push back on vague answers. If I say "I like to keep things simple," ask "Simple how? Give me an example of simple done right and simple done lazy." 3. Ask for specific examples. "Show me a sentence you've written that captures this." 4. Call out contradictions. If I said one thing earlier and something different now, point it out. 5. Go deeper on interesting threads. If something unusual emerges, follow it. 6. Don't accept "I don't know" easily. Try reframing the question or approaching from another angle. </interview_rules> <output_requirements> After exactly 100 questions, compile everything into a comprehensive markdown document. This is NOT a summary. It's a complete reference document preserving the full depth of every answer. Structure it like this: # VOICE PROFILE: [My Name] ## Core Identity [2-3 sentences capturing the essence. This is the only summary section] *** ## SECTION 1: BELIEFS & CONTRARIAN TAKES ### Q1: [The question you asked] [My full answer, preserved verbatim or lightly cleaned up for clarity] ### Q2: [The question you asked] [My full answer] [Continue for all questions in this category] *** ## SECTION 2: WRITING MECHANICS ### Q16: [The question you asked] [My full answer] [Continue for all questions in this category] *** ## SECTION 3: AESTHETIC CRIMES [Same format: question, then full answer] *** ## SECTION 4: VOICE & PERSONALITY [Same format] *** ## SECTION 5: STRUCTURAL PREFERENCES [Same format] *** ## SECTION 6: HARD NOs [Same format] *** ## SECTION 7: RED FLAGS [Same format] *** ## QUICK REFERENCE CARD ### Always: [Extracted from answers: specific patterns to follow] ### Never: [Extracted from answers: specific things to avoid] ### Signature Phrases & Structures: [Actual examples I provided during the interview] ### Voice Calibration: [Key quotes from my answers that capture tone] *** ## HOW TO USE THIS DOCUMENT (ANTI-OVERFITTING GUIDE) This document captures my taste. It is NOT a checklist to follow rigidly. ### Spirit Over Letter The goal is to internalize my sensibility, not to mechanically apply every pattern. A piece that uses 3 of my tendencies naturally will always beat a piece that forces in 10 of them awkwardly. ### Frequency Guidance For each tendency documented above, note whether it's: - **HARD RULE**: Never violate (these are rare, usually in the "Never" section) - **STRONG TENDENCY**: Do this 70-80% of the time, but breaking it occasionally is fine - **LIGHT PREFERENCE**: Nice to have, but context determines when to apply When no label exists, assume it's a LIGHT PREFERENCE. ### Context Matters My voice adapts to format: - A tweet ≠ a newsletter ≠ a LinkedIn post ≠ a long-form article - Use judgment about which patterns fit which format - Some of my tendencies are format-specific, note when this applies ### Natural Variation Real writers aren't perfectly consistent. Introduce natural variation: - Don't start every piece the same way just because I have a "signature open" - Don't avoid a word forever just because I said I dislike it. Sometimes it's the right word - Let the content dictate structure, not the template ### The Litmus Test Before finalizing anything written "as me," ask: > "Does this sound like something I would actually write, or does it sound like an AI trying very hard to imitate me?" If it feels forced, pull back. Less imitation, more inhabitation. ### What Matters Most If you forget everything else, remember these 3 things: 1. [My single most important belief about writing] 2. [The one pattern that makes my voice mine] 3. [The #1 thing I never do] Everything else is secondary. *** ## INSTRUCTIONS FOR AI When writing as [My Name], reference this document. Pay attention to: 1. The specific examples I gave: use similar structures. 2. The words and phrases I said I hate: never use them. 3. The beliefs I hold: let them inform the angle. 4. My actual sentences: match the rhythm and length. This document is a source of truth, not a suggestion. Apply it with judgment, not rigidly. </output_requirements> Begin by asking me your first question.
Tips for Getting the Best Results
- Be specific. "I write short sentences" is vague. "My paragraphs rarely exceed 3 sentences and I never use semicolons" is useful.
- Give real examples. When the AI asks for a sentence that captures your style, paste an actual one from something you've written.
- Don't edit yourself. If your honest answer is "I swear in my writing and I'm not stopping," say that. The Taste File is for accuracy, not aspiration.
- Embrace contradictions. Real people are inconsistent. If the AI catches a contradiction, explore it. That's often where your most interesting voice patterns live.
What You Can Do With Your Taste File
Once you have it, the applications go well beyond "write a blog post as me."
Content Creation
Blog posts, newsletters, social captions, email sequences, all in your voice without manually editing every output.
Team Alignment
Share your Taste File with writers, VAs, or agencies. They feed it to their AI tools and immediately match your voice, no weeks of "getting to know you."
AI Clone Foundation
A Taste File is the first layer of building a full AI clone of yourself, one that can respond to your audience, coach your clients, or handle DMs at scale.
Personal Brand at Scale
Produce 10x the content without sounding generic. Your Taste File ensures consistency across every channel, format, and tool you use.
Frequently Asked Questions
How long does it take to create a Taste File?
The full interview takes 60-90 minutes. The AI asks 100 questions one at a time, then compiles everything into a structured markdown file. You do it once and reuse the output indefinitely.
Does it work with any AI?
Yes. It's a plain markdown file, so it works with Claude, ChatGPT, Gemini, Grok, or any LLM that accepts context. Just paste or attach it at the start of a session.
Why focus on what I reject?
Your taste is mostly defined by boundaries. Without them, AI defaults to a generic "statistical middle" that sounds like everyone and no one. The things you won't say are often more distinctive than the things you will.
How is this different from custom instructions?
Custom instructions are typically short and surface-level ("write casually"). A Taste File is 100 data points across 7 dimensions of your voice. It gives AI far more signal to differentiate you from the default.
Can I update my Taste File over time?
Absolutely. Run the interview again after 6-12 months, or edit sections manually as your style evolves. Your voice changes, and your Taste File should too.
Stop Sounding Like Everyone Else
Every creator using AI faces the same problem: the output sounds like AI. It's grammatically correct, structurally sound, and completely forgettable. The Taste File is the fix, not a hack, but a genuine extraction of what makes your voice yours.
90 minutes of honest answers. One portable file. Every AI tool you'll ever use, writing as you.
Want to Go Further Than a File?
A Taste File captures how you write. An AI clone captures how you think, coach, and interact. Personify builds full AI versions of coaches, creators, and experts, trained on your content, your style, and your methodology.
Related Articles
Create an AI Version of Yourself
Step-by-step guide to building an AI that thinks and responds like you.
Read moreAI Coaching Clone
How coaches are using AI clones to serve more clients without losing the personal touch.
Read moreAI Clone for Personal Branding
Scale your personal brand with AI while maintaining authenticity and trust.
Read more