The Commissioning Brief as Creative Brief: Encoding Brand Voice Into AI Prompts

Most teams treat the commissioning brief as a logistical document—word count, deadline, topic, done. They then hand it to an AI model and wonder why the output sounds like every other AI output: competent, hollow, interchangeable.

The real problem isn't the AI. It's that the brief contains almost no information about how the brand thinks. It's a specification sheet, not a voice specification. And when you don't encode voice into your prompt, you get the statistical average of everything the model has learned—which is precisely what you don't want.

A commissioning brief that works as a creative brief does something different. It treats the AI as a collaborator that needs to understand not just what to write about, but how your brand argues, what it refuses to say, what it privileges, and what it sounds like when it's being honest.

The thing everyone gets wrong: they assume voice emerges from tone adjustments.

"Write this in a conversational tone" or "make it authoritative" are instructions that land nowhere. They're adjectives applied to the surface. Real voice lives in the choices you make about what matters, what gets questioned, what gets assumed. It lives in the ratio of certainty to doubt. It lives in whether you use "we" or "I" or neither. It lives in whether you're willing to say "we don't know" or whether that feels like weakness.

The commissioning brief that actually works encodes these choices. Not as tone notes, but as examples of thinking. A paragraph from a previous piece that shows how your brand handles nuance. A sentence that demonstrates what you refuse to do—oversimplify, hide behind jargon, pretend consensus exists where it doesn't. A statement about what your audience needs to believe about you after reading this, not just what they need to know.

Why this matters more than people realize: consistency at scale requires a system, not intuition.

When you're publishing one piece a month, a talented writer can absorb your voice through osmosis and intuition. When you're publishing daily, or when you're coordinating across multiple writers and AI systems, intuition breaks down. You need a specification. And that specification has to be detailed enough that someone (or something) who's never worked with you before can produce work that sounds like it came from your team.

This is where most AI governance fails. Teams build brand guidelines that describe visual identity with precision—hex codes, spacing ratios, typeface weights—but describe voice with vagueness. "Authentic." "Thought-leading." "Approachable." These words mean nothing to a language model. They mean nothing to a new hire either.

The commissioning brief as creative brief reverses this. It treats voice as a specification problem. It says: here's what we sound like when we're at our best. Here's what we refuse. Here's what we're willing to be uncertain about. Here's what we always come back to. And it does this not through adjectives but through examples, through explicit constraints, through the logic of what we believe.

What actually changes when you see it clearly: your output becomes defensible.

When a piece comes back from an AI system and it doesn't sound right, you can now point to the brief and say exactly why. "This is too certain about something we said we'd be tentative about." "This uses 'we' when we decided to use 'I' in this context." "This assumes the audience already agrees with us, but the brief says we need to earn that agreement first."

More importantly, your team stops treating AI output as something that arrives fully formed and needs minor tweaking. You start treating it as a draft that's been shaped by a system you've designed. You become the editor of a process, not the editor of random outputs.

The commissioning brief that works as a creative brief is longer than most. It's more specific. It might include voice constraints that seem obvious to you but aren't obvious to anyone else. That's the point. It's the difference between hoping your AI system produces on-brand work and knowing it will.