AI Winner Sprint: 100 AI UGC ads, a 7-day test in your ad account, money back if none beats your best.

For DTC brands on Meta

Three new winning angles for your product, tested in 7 days. Or your money back.

We write three angles from your reviews and your competitors' live ads, make 100 AI UGC ads of your real product, and test the best 30 against your best ad in your Meta account for 7 days, in two rounds. You keep the winners.

Per winner, 1% of your monthly Meta spend (at least $500), only for an ad that wins at scale. Test spend stays in your ad account. One brand per category per month.

How it works

Angles in 24 hours. Ads live in 48. Winners on day 9.

  1. Hour 0

    Pay, then share your store link, your target CPA or ROAS, your monthly Meta spend, your test budget and partner access to your ad account (or launch the ads yourself), and name your best ad. We publish your Sprint's sealed details on our public ledger as a SHA-256 hash: the budget split, the ad you named as the bar, and the test settings. Your category is fixed and shown there as taken. The rule itself is this page, so neither side can change it once the numbers come in.

  2. Hour 24

    Three angles for your approval, each with hooks and a script, built from your reviews and your top competitors' live ads.

  3. Hour 48

    100 AI UGC ads made from your real product photos: AI actors, voices, scenes and captions, in 9:16 and 4:5. We pick the 30 that go live and seal our top-3 prediction beside the rule, and each of two blind AI judges seals its own.

  4. Days 1-4, round one

    Each of the 30 runs in its own ad set at $20 a day, so every ad gets the same budget. Your ads keep running as they are; your best ad's numbers over the same days are the bar.

  5. Days 5-7, round two

    The 3 with the lowest cost per purchase in round one get fresh ad sets at $200 a day each and face your best ad again over the same days. One lucky week isn't enough.

  6. Day 9

    The seals open. The winner report: the rule, both rounds with your own Ads Manager export attached, our prediction against the result, which angle and hook won on hook rate, CTR and cost per purchase, what to scale today, and what to make next.

The AI Winner Sprint

One offer. Paid on results past the first $1,500.

What you get

  • Three new angles with hooks and scripts.
  • Every finished ad we make, licensed to you for good; only the cash refund ends it.
  • A sealed two-round test of the best 30 in your own ad account, 7 days in all.
  • The winner report, the sealed prediction against the result, and the brief for round two.
  • Your category closed to other brands for the month you buy.
  • Optional: Meta's own AI, reported alongside. An Advantage+ creative variant of your best ad runs as one more ad set on the same budget as each of ours, in both rounds (+$680 of your media), so you see whether you need us at all. It never counts as our winner.

Price

$1,500 up front, by card or bank transfer

  • Per winner: 1% of your monthly Meta spend, at least $500, for a test ad that passes $20,000 of spend at or above your target ROAS within 90 days. Nothing more if none does. The fee covers any of our ads you keep, a Sprint with no win included.
  • Test spend stays in your ad account: $4,200 over the 7 days by default. Round one is $2,400 (30 ads at $20 a day for 4 days) and round two $1,800 (3 ads at $200 a day for 3 days). You can set less, down to $2,000: both rounds shrink in proportion and the rule stays the same, but luck wins more tests (18% at $2,000 against 13% at $4,200).
Start a sprint, $1,500
Beat your best ad twice

The guarantee

Beat your best ad twice, or your money back.

One of our ads wins if its cost per purchase beats your best ad's in round one (days 1-4) and again in round two (days 5-7), each over the same days.

Your best ad

The one ad you name at hour 0: it spent $500 or more in the last 30 days and stays live through day 7, so luck can't set the bar.

If none of our ads wins, you choose

Your $1,500 back. The license to the ads ends with it.
A free second Sprint, with three new angles, on the same sealed rule.
Keep everything we made, and get $1,000 back. The winner fee still applies to any kept ad that passes $20,000 of spend at your target ROAS within 90 days.

Sealed at hour 0

The rule is on this page for everyone. Your Sprint's details are sealed as a SHA-256 hash at hour 0, so we can't change anything after the numbers come in, and neither can you.

How to claim

Ask within 14 days of the day-9 report. The test has to run: the best 30 ads live in round one and the top 3 in round two, at the test spend you set.

Who it's for

Built for brands that already sell on Meta.

A fit if

  • You're a DTC brand spending $30k or more a month on Meta.
  • Your best ad is getting tired and you need the next one.
  • You'd rather test 30 new ads than brief one shoot.

Not a fit if

  • You spend under $30k a month on Meta: a week of test media costs more than one winner is worth, so take the free ad instead.
  • You want a promised ROAS: nobody can honestly promise one.

FAQ

Questions

Are the ads AI?

Yes: AI actors, voices and scenes, built from your real product photos and checked frame by frame against your product. We keep any AI label a platform applies.

Do you need my ad account?

Partner access in Meta Business Manager lets us launch and read the test. If you'd rather launch yourself, we hand you the 30 ads and the test setup.

Why seal the rule?

So neither side can move the goalposts once the numbers are in. The rule itself is on this page for everyone. What we seal at hour 0 is your Sprint's details: the ad you named as the bar, the budget split and the test settings. They stay private until day 9; then anyone can check them against the hash.

How do I know the test ran as sealed?

It runs in your ad account, so you see every number before we do, and the report attaches your own Ads Manager export. The hash proves the rule came first; your account proves what happened.

Why offer Meta's AI in my test?

Because you'll wonder. If its variant beats ours, the report says so; it never counts as our winner and never changes your remedy.

Who owns the ads?

Every ad we made is licensed to you for good. Only the cash refund ends it.

What counts as my best ad?

The one ad you name at hour 0. It must have spent $500 or more in the last 30 days and stay live through day 7. Picking the cheapest of your ads after the test would let luck set the bar, and naming a weaker one only costs you your remedy, so name your true best.

Why two rounds?

Because one round is a lottery. With 30 ads at about two purchases each, the luckiest ad beats your best almost every time, even when none is better. Round two re-tests the top 3 against your best on fresh budget, so a lucky week isn't enough.

What if none of your ads wins?

You pick: your $1,500 back (the license to the ads ends), a free second Sprint on the same rule, or keep everything we made and get $1,000 back.

What is a winner?

An ad that passes $20,000 of spend at or above your target ROAS within 90 days of launch.

What is a category?

The product type a shopper would compare, like "collagen powder" rather than "supplements". It is fixed at hour 0 and shown on the ledger as open or taken. When yours is taken, you join the queue for next month.

Will my results be public?

Every sealed Sprint goes on our public ledger, wins and losses. Your brand's name appears only with your yes.

How fast is it?

Angles in 24 hours, ads live in 48, report on day 9.

Who is it for?

DTC brands spending $30k or more a month on Meta. Below that, a week of test media costs more than one winner is worth, so take the free ad instead.

How do I pay?

$1,500 up front at a secure checkout, by card or bank transfer; the charge shows as WHOP*PRONTOCUT. Winner fees are invoiced only when an ad qualifies.

Free

Send your store link. Get one finished ad free.

One finished AI UGC ad of your product, made from your real photos, sent to your email. Yours to keep and run.

We use your email only to send your ad and follow up about it. Reply "no" and we stop.