Your brand voice is what keeps your product pages from sounding like everyone else's. Define it well and an AI writes like you; define it badly and it writes like everyone else. And most voices are defined badly, because describing a voice gets confused with listing adjectives. Here's the practical way to do it, so that an AI — or a new copywriter — can truly live and breathe it.
The mistake: describing the voice with adjectives
"Approachable, professional, fresh, trustworthy." Your brand says it, and so do the other fifty in your industry. Adjectives are too abstract: neither a model nor a human knows what to do with them. "Approachable" can mean an emoji, or it can mean speaking informally but with respect. The label isn't enough.
What works: examples, not labels
Show, don't tell. Give the AI two or three real examples of text written in your voice — a description you love, a couple of sentences that sound like you — and it will stop guessing and start imitating the pattern. One concrete example says more than ten adjectives.
The golden trick: contrasting examples
Nothing teaches like contrast. For the same idea, show how you would say it and how you wouldn't. The "not like this" marks the edges just as well as the "like this" marks the center.
Define the what, not just the how
- Vocabulary: words you use and words you avoid.
- Form of address: formal or informal, and how you speak to the customer.
- Length and rhythm: short sentences or longer ones.
- Honesty limits: what you never claim without data to back it up (promised results, health benefits).
One voice, applied to five thousand product pages
Here's the real payoff: you define the voice once and it's applied to the whole catalog, without blurring by the time you reach product page 300. That consistency at scale — impossible to keep up by hand — is what keeps a large catalog sounding like a single brand.