Authentic and Consistent: Brand Voice Across a Company, a Sub-Brand, and a Human

Two fears show up the moment you let AI write things for your business, and they pull in opposite directions.

Two Fears, Opposite Directions

Fear one: everything will sound like it was written by the beige committee that names airport lounges — the same flat “we value your business” mush, the verbal equivalent of hold music run through a laminator. Fear two: everyone will freelance, prompting their own way until your brand splinters into a dozen voices, three of which are openly feuding.

I have an unusual stake in solving this, because I run wildly different brands at once. Resistance Wine is irreverent and a little unhinged (“resist the ordinary”). CU 2.0 is a contrarian operator talking to executives. McDrake is its own thing. And then there’s just me, the founder, who sounds like none of them and all of them. If AI flattened all of that into one beige corporate voice, I’d lose the only thing that makes any of them worth following.

Voice Has Three Layers

The fix is to stop treating voice as one setting and start treating it as three layers.

The company layer is the guardrail — your values, your tone, the things you’d never say. Encode it once, feed it to the AI every time, automatically. The sub-brand or department layer flexes within that — your fun line shouldn’t sound like your serious line, and your sales voice shouldn’t sound like your support voice. And the human layer is the person actually talking, because your customer is talking to a someone, not a logo, and your best people have a way of saying things that is the relationship.

It’s a Context Problem, Not a Voice Problem

Here’s the counterintuitive part I learned the hard way: when AI writes something generic, it’s almost never a voice problem. It’s a context problem. The model wrote beige because you handed it nothing real — so it did what any of us would do with no information and a deadline, and produced confident filler. Feed it the actual material — this customer’s history, the award you just won, the reason this product exists, the thing you’re genuinely proud of — and it stops sounding like a robot, because it finally has something true to say.

Get the three layers right and you get the thing that used to be impossible: everything sounds unmistakably like your company, appropriately like the part of it talking, and genuinely like a human — at a scale you could never staff by hand. You don’t have to choose between authentic and consistent. That trade-off was always a tooling limitation, and the tooling just changed.

Final Thoughts

Why It Matters: AI lets you finally have authentic and consistent — but only if you stop treating voice as one flat setting and start treating it as company, sub-brand, and human layers.

What to Do: Write down your company voice once, let departments flex within it, and feed AI real context instead of expecting it to invent personality from nothing.

Coming Up Next: The big one — why the future belongs to founders who put their people above the AI, and why it’s coming faster than you think.

Disruption with a side of humor — Kirk Drake, signing off (until the next post).

Building a voice your AI can actually wield — across brands and people — is one of my favorite things to teach from the stage. Book me here.