You've seen the post. Polished, formatted, three crisp insights, a clean close. Something feels off. Not wrong exactly. Just frictionless. Like it could have been written about anyone, or for no one.
The feeling has a name. It's provenance. You're sensing that there's no human behind it: no experience that generated the insight, no opinion formed from anything real. The words are there. The person isn't.
Most AI-generated posts have this problem. They start from a prompt, not from a person. The AI is asked to produce something on a topic, and it does: fluently, correctly, without any connection to what anyone actually knows or has lived through. The words are smooth; the origin is missing.
Ghostwriting solves for this. When a ghostwriter works with someone, the output carries the subject's actual ideas: their positions, their experiences, their way of seeing things, translated into cleaner language. The words are new. The thinking is theirs. That has always been the legitimate test: not whether you wrote every word, but whether every word honestly represents you.
What's different now is that we can show the receipts.
Every post published through Proofd displays two numbers:
Words you contributed: the raw word count of the voice note transcripts that fed the post. These are your words, spoken by you, before any shaping by AI.
Time you put in: the cumulative duration of all voice notes that contributed to the post. Three minutes and forty-seven seconds of your actual voice went into this.
These two numbers are the provenance score. Not a grade. Not a rating out of ten. A documented record of where the post actually came from.
The score doesn't say the post is good. It says the post is yours. That's a different claim, and a more honest one.
When a reader sees that 4 minutes 12 seconds of voice and 340 words of transcript went into a post, they know something real: that a person sat with this, spoke about it, and brought actual experience to the topic before any AI touched it. The AI shaped the material. The material was real.
This matters because the alternative is eroding the baseline trust of professional content: posts that claim to represent someone's professional thinking but carry no evidence of it. The feed is full of signals with no origin. The provenance score is a different kind of signal: one that points back to a person, with receipts.
There's also something simpler going on. The score is honest. It says: I used AI to help write this, and here is how much of me was in it before the AI started. That transparency doesn't diminish what you're saying. It's the move that professional publishing eventually has to make: acknowledging the tool while making clear that the thinking is yours.
We built the provenance score because that's the right way to use AI in professional publishing. Not to generate content from nothing. To express real thinking more clearly than time and blank pages allow.
The number isn't a trophy. It's a handshake.
I was here. This came from me. Proofd just helped me say it right.
