Brand personality framework: build this, not a persona
Somewhere right now, a team is in a workshop deciding which celebrity their brand would be. Someone has suggested Zendaya. Someone else is building a moodboard. By Friday it’s a slide, and by next quarter nobody can find the file.
Brand personas are a thirty-year-old habit, and the instinct behind them is sound: give everyone the same picture of who’s talking and the writing falls into line. The catch is what the picture holds. Age, job, Spotify Wrapped, favorite sitcom. That tells you what the brand watches on a Sunday. It doesn’t tell you what the brand would actually say when a competitor picks a fight.
A workshop moodboard also can’t feed an AI tool, which is the gap a brand voice generator is meant to close, as long as it outputs rules instead of vibes.
A brand personality framework fixes that. It’s a short, structured document that names what your brand believes, what it stands against, the words it uses, and the moves it repeats, specific enough that any writer or AI tool can sound like you.
A persona describes a character. A framework gives that character a spine. Here’s how to build one.
What personas get right
The instinct is correct. Humanizing a brand does make its writing more consistent, because if everyone picturing who’s talking pictures the same person, the outputs converge. That part works, and it’s why the exercise has survived three decades of marketing fashion.
The execution is where it falls apart. Most personas stop at description. They answer what kind of person is this and call it done. A 35-year-old who reads literary fiction and drinks natural wine is a mood board. It produces a vibe. It can’t tell you what to write when a competitor makes a claim you disagree with, how to word a price increase, or which words to keep off the page. It’s a portrait when what you needed was a set of instructions.
Why personas can't brief a writer (or an AI)
A human writer can squint at a persona and fill the gaps with judgment. Tell them “she’s witty but never mean” and a decent writer will mostly land it. A model can’t. Hand ChatGPT “our brand is a witty 30-something who loves indie music” and it returns the average of every witty 30-something in its training data, which is the same thing your competitor gets from the same brief. Demographics aren’t instructions.
The persona was built for a reader who could interpret it, and that reader is exactly the one being replaced at the keyboard. Personality is what the brand is like. Voice is what it does on the page. A persona over-invests in the first and skips the second, which is fine until the writing has to scale past the people who sat in the workshop.
What goes in a brand personality framework
The upgrade from persona to framework is one question: what does this brand believe that someone could disagree with? Everything else hangs off the answer. There are four parts, and the order matters.
The belief. Your POV. Not “we value quality.” A claim a competitor would publicly argue with. Basecamp built a company, several books, and a famously contrarian blog on one: hustle culture is a lie, and it doesn’t have to be crazy at work. Every line they publish traces back to it. This is the load-bearing wall, so pick it first.
The enemy. The category convention you exist to push against. BrewDog didn’t define itself as craft beer with personality. It defined itself against bland, mass-produced corporate lager and called its flagship a beer for punks. Name what you’re against and half your writing decisions make themselves.
The vocabulary. The words that follow from the belief, and the ones that contradict it. Who Gives A Crap sells toilet paper to fund sanitation, so its vocabulary is plain, cheerful, and a little crude on purpose. “We donate 50% of profits” stays. “Synergistic impact solutions” would detonate the whole thing on contact.
The signature moves. The patterns that repeat across everything. Cards Against Humanity opens with the line that shouldn’t be allowed, then lets the silence do the work. Name the move and a writer or a tool can run it on purpose.
One brand, written two ways
Here’s the difference on the page. Take Maple, a fictional tax-filing app. The persona version of its brand reads like this: “Maple is a friendly, approachable 34-year-old who makes complex things simple and always has your back.” Hand that to a writer and you get: “Filing your taxes has never been easier. We’re with you every step of the way.” Technically on-brand. Indistinguishable from forty other apps.
Now the framework version. Belief: filing taxes shouldn’t feel like punishment for being bad at paperwork. Enemy: the fear-based tax-prep industry that sells you anxiety and upsells the cure. Vocabulary: plain money words in, “maximize your refund” out. Signature move: name the scary thing, then defuse it on the next line. Same writer, same length of brief, and now you get: “Yes, you can do this yourself. No, the IRS is not sitting there waiting to ruin your life. Here’s the one number that matters, and we’ll find the rest.”
Same app. One sounds like every tax product in the store. The other could only be Maple. Nothing changed except that the brief carried a belief and a banned list instead of a birthday and a Netflix queue.
Turn the persona you already have into a framework
You don’t have to throw the persona away. Mine it. Take “approachable, a little irreverent, values community” and force each trait into a part. What belief makes you irreverent, and who does it annoy? What does “community” make you refuse to say? Irreverent toward what, exactly?
The traits that can’t survive that interrogation were decoration. The ones that can become the framework. The test is the one every voice has to pass: cover the logo. If your framework only produces copy a competitor could have published word for word, you’ve written a nicer persona, not a personality. Push the belief until it has an edge someone could catch a hand on.
How to get AI tools to use your framework
Here’s where most personality work dies. It gets built in a workshop, written into a slide, admired once, and never reaches the place the writing actually happens: the prompt box. So the brand still sounds like itself on the homepage a copywriter sweated over, and like nothing in particular across the fifty AI-assisted emails and product descriptions nobody briefed.
You can hear that gap on the page: these tone of voice examples show what a brand running on a real belief reads like next to one running on defaults.
A framework only earns its keep if it leaves the slide. The four parts have to become machine-readable, structured so every tool writing on your behalf starts from your belief instead of from zero. That’s what Boomvoice builds: your belief, enemy, vocabulary, and signature moves turned into a voice profile that plugs into your AI tools through MCP (Model Context Protocol). Boomscore then tells you when a draft has drifted off it.
Until that profile is wired in, the fastest stopgap is a ChatGPT system prompt for brand voice that carries the same four parts into every session by hand.
Define what your brand actually believes and make it readable by the tools doing the writing, and you have a personality. Leave it as a paragraph about a fictional 34-year-old and you have a persona. The persona was never going to write anything.
Turn your framework into a profile your AI tools can use.
Your belief, enemy, vocabulary, and signature moves, structured into a machine-readable voice profile every AI tool can use. That’s what Boomvoice builds, through a conversation, not a form.
Build your voice profile → boombrand.ai
