Loading...
Loading...
Change the ask, edit what is left, then let someone with taste choose. When we tested it on four leading AI models, a better ask cut ads with a tell from 71% to 2%. Here is the exact ask, and what it cannot do.
By Ad Legends ยท
This is the exact ask from our study, word for word, with your brief where ours went. The list in the middle is the one every Legend works under.
We sent the same 50 small-business briefs to the same four models twice: once with a plain ask, once with the ask above.
| Share of ads | Plain ask | Better ask |
|---|---|---|
| Ads with at least one tell | 71% | 2% |
| Headlines that were mostly the brief | 68% | 2% |
| Share of the five most common buttons | 63% | 19% |
| Ads with a long dash | 50% | 0% |
| Ads that kept the briefโs fact | 99% | 84% |
It costs you some facts. Asked for an idea the owner had not stated, the models kept the briefโs own fact 84% of the time, down from 99%. Check that your proof made it in.
And it cannot choose. The better ask gave us four different headlines for every brief, most of them decent, some of them good. Which one runs is a judgment about your customer and your brand, and a model will pick whichever it was asked to pick.
Paste the draft into the free slop meter. It counts the tells, shows which ones it found, and never guesses who wrote it. When it reads clean, hand it to a Legend and let a blind judge tell you if the idea holds.
Change the ask first: give the model the words it may not use and ask for an idea you did not already state. In our test that cut ads with a tell from 71% to 2%. Then edit what is left and have someone with taste choose.
The one on this page, which we tested on four leading models: your brief, a request for an idea you did not state, and a list of words and habits the model may not use. Copy it and put your own brief in the brackets.
From ads, mostly yes. 52% of the AI-written ads in our study used one, which makes it one of the easiest tells to spot. A comma or a full stop almost always works.
It is not built to, and detectors are not reliable on thirty words anyway. It is built to remove what readers notice: the tells. Check what is left on the free slop meter.
It replaces the first draft. It does not replace the person who knows which idea is right for the brand. That is why every Legend round ends with a blind judge.