How We Learn Your Voice, and Why We Never Stop

We wrote recently about the words we refuse to publish. This post is about something harder to pin down, which is how we learn what a client sounds like.

Before Profile Pros had a name, our founder, Eric Howie, ran the experiment on himself. After twenty years of building and operating companies, he knew exactly what he sounded like in a boardroom, so he took writing he cared about, fed it to the best AI tool of the day, and asked it for more in the same voice. What came back was polished and competent and grammatical to a fault, and it sounded like nobody he had ever met, least of all him. Nothing in it was wrong, exactly, but nothing in it was his either, and that draft became the founding memory of this company. We keep it around the way some people keep a rejection letter.

It took us longer to understand the failure than to have it, and the understanding eventually settled into four convictions about what a voice actually is.

The first is that voice is not static. A real person writes long and messy one week and short and blunt the next, so when you capture a single snapshot of someone you freeze them into a template, and readers feel a template repeating before they can name what is bothering them. A system that learns your voice once has already failed, because a voice that stops being learned starts being faked.

The second is that people cannot describe their own voice. Ask, and you get adjectives like "professional but approachable," which sound meaningful until you notice that every consultant on LinkedIn picks the same four words. What tells the truth is what your voice does on the page, in real samples, and what you cross out when someone shows you an imitation of it.

The third is that voice lives in specifics: the client story only you were in the room for, the number you still remember from 2019, the opinion your peers think is slightly wrong. A perfect imitation of your sentence rhythm wrapped around nothing is still nothing, which is exactly what that first experiment had produced.

And the fourth is that voice is learned the way taste is learned, through feedback over time. You show someone work and you watch what they keep, what they change, and what they reject without quite being able to explain why, and the rejections teach you the most.

So that became the service. We start from your real writing and your intake answers rather than adjectives, a written interview collects the moments and numbers only you have, you can email us an idea or a recent win whenever one happens, and every weekly delivery becomes another round of calibration, because what you rate highly shapes next week and what you quietly ignore shapes it too.

The draft that started this company fooled nobody, and the ones we ship now have to convince the only reader who counts, which is the one who knows you. We never entirely stopped being afraid of that first draft. The fear turned out to be useful.