There are two kinds of proof in this product, and confusing them is how guides like this one go bad.

The first kind is ours, and we’ll meet it.

Every sample script, hook and generated asset in these materials was actually generated, and the prompt that produced it is printed alongside it. When a take flopped, the flopped take is in the appendix with the others. That’s a standard we can hold ourselves to, because it’s a claim about our own work.

The second kind isn’t ours to give. It would read like this:

Readers typically earn per brand deal within of finishing Act Two, and % go on to .¹

We couldn’t print that paragraph, so we didn’t print it. The words are gone rather than softened, because a hedged version of an invented number is still an invented number.

What the guide teaches instead is a habit, not an outcome: show the real generated output, show the prompt that made it, show what the client actually saw. Audiences start more skeptical of AI content than of anything else you will make. Honest provenance is the only thing that survives that, and it costs nothing.

One stated exception, so you don’t have to work it out yourself.

The week sketched at the top of this page did not happen. It is labeled imagine this, it is typeset as a treatment (the document a production writes to propose how something will feel, before anything is shot), and it contains no rate, result or figure presented as real. That is a different category from a sample asset: an exhibit claims this is what we made, a treatment claims this is what we are proposing. We would rather name the distinction here than let a reader wonder whether we quietly exempted ourselves from our own rule.

1. 1.Redacted because the evidence doesn’t exist yet and won’t belong to us when it does. Proof of your results sits with your clients, after you’ve done the work. Any figure printed here would have been invented, so nothing is printed here.