Hi, it's Peggy.
Here's a benefit line, the kind Claude will hand you fifty of when you ask:
"Cut your reporting time in half."
Read it the way your buyer does, in about half a second, already skeptical. One question forms, and it's the one that decides everything: says who, and how?
The line has no answer. It makes a promise and gives the reader nothing to believe it with.
That gap is the most common weakness I find when I review AI copy, and it has a name and a fix that both predate the technology by a hundred years.
Let me walk you through it.
Every Claim Needs a Because
Claude Hopkins wrote it down in Scientific Advertising in 1923:
Don't just make the claim. Give the reason why. A price is low for a reason. A product works for a reason.
Name the reason and the claim becomes believable. Leave it out and the claim is an adjective with a period after it.
Modern direct response calls the same thing the mechanism. It's the specific because underneath the promise.
Why does this tool cut reporting time? Because every dashboard reads from one source instead of four, so there is nothing left to reconcile by hand. Now the promise has something holding it up.
It works the same in plainer copy. "Our onboarding is painless" is a claim the reader has heard from every vendor who then handed them a forty-field form.
"Onboarding is one screen, because we pull your settings straight from the tool you already use" answers the doubt in the same breath it raises it.
The bare version leaves the doubt sitting there with nothing to push against.
The mechanism earns its place twice. It gives the reader a reason to believe, and it gives you a line a competitor can't lift by writing the same adjective.
"Faster" is available to everyone. "Faster because your data stops living in four places" belongs to the one product that can actually say it.
I've written before about one way AI copy comes out hollow, when it opens on the product instead of a desire the reader already has. This is a different hollow. The copy names a real benefit the reader wants, then hands them no reason to trust that they'll get it.
Why Claude Hands You the Claim and Stops
Ask Claude for benefits, or to "make this more persuasive," and a bare claim is the cheapest thing that satisfies the request.
"Cut reporting time in half" is complete, grammatical, and on brief. The model has no reason to reach past it for the mechanism, because the mechanism is not what you asked for.
You asked for a benefit, and a claim reads as a benefit.
The reason behind the claim is also the part the model is least able to supply on its own. Why the promise is true lives in the product, the engineering, the real before-and-after. That information is yours.
When it is missing from the prompt, the model leaves the claim bare or, under pressure to be specific, invents a because. Nova watched that happen in her last issue, when she pushed a draft to sound more specific and it manufactured a survey statistic out of nothing.
So hand the model the real reason. Don't let it write you a fake one.
The Because Pass
Run this on any draft before it goes out. It walks every claim and asks for the reason behind it.
Go through this copy and list every claim, benefit, or promise it makes to the reader. For each one, give me the specific reason the reader should believe it, the mechanism that makes it true. If the copy already provides that reason, quote it. If it does not, mark the claim as unsupported and tell me what you would need from me to support it. Do not invent reasons, numbers, or mechanisms. A claim you cannot support from what I have given you gets flagged, not filled.
[paste your copy, plus any product notes or research you have]The instruction to flag instead of fill is the whole job.
A model told to support every claim will support every claim, and it will invent the support for the ones your product cannot back.
A flagged claim is more useful than a filled one, because the flag is a decision waiting for you: go get the reason, or cut the claim.
Two limits are worth calling out:
A reason-why only works when it is real, so when the flag comes back and nothing about the product actually explains the promise, that flag is really about the offer. Mark already wrote the issue on why you cannot write your way out of one. And a because does not rescue a claim the reader never cared about.
The mechanism has to sit under a promise that matters to them. Prove an irrelevant claim and all you have proved is something nobody asked about.
The Burnett Matrix
A claim asks the reader to take your word.
The reason why is what lets them stop taking it on faith.
Before the next AI draft ships, run one pass: for every promise on the page, point to the because.
Where you can point to it, quote it in the copy.
Where you cannot, you are holding a claim to prove or a claim to cut, and both of those are forward motion.
More clicks, cash, and clients,
Peggy Burnett
