I don’t use AI for motivation.
I use it to move faster.
Last week we refreshed ads for our brick-and-mortar youth sports training facility.
Instead of sitting there “brainstorming angles,” I did this:
Fed AI deeper context about our offer
Uploaded screenshots of our funnel
Uploaded screenshots of the actual program
Gave it constraints (who it’s for, what outcome, survey feed back from parents)
Asked for multiple emotional hook variations
Then I filmed and launched.
No debating.
No perfecting.
No overthinking.
Just structured testing.
From Feb 13 – Feb 16 (3 days):
37 website leads
$2.58 CPL (cost per lead) on one ad set
$3.81 CPL on another
~$114 spent per ad set
Same targeting.
Same budget.
The only thing that changed:
The hook and emotional framing.
What Actually Made It Work
It wasn’t production quality.
It wasn’t editing.
It wasn’t targeting tweaks.
It was emotional positioning.
The ads that won:
Spoke directly to parent identity
Highlighted fear (falling behind, lack of confidence, clumsy movement)
Painted the outcome emotionally (“they move differently,” “they run with confidence”)
We didn’t sell drills.
We didn’t sell workouts.
We sold transformation in identity.
Features didn’t move the needle.
Emotion did.
If you’re new to this:
A hook is simply the first emotional idea that makes someone feel seen.
That’s what changed performance.
Why AI Helped (And Why Most People Use It Wrong)
AI didn’t magically write a winner.
Context did.
The better I fed it:
Funnel screenshots
Offer structure
Customer journey steps
Real objections we hear
Clear constraints
The better the output.
AI without context = generic garbage.
AI with context = leverage.
If you’re getting mid results, it’s usually because your inputs are shallow.
The Part Most People Miss
Finding a winning ad is not the win.
Extracting it is.
Now that we know this emotional hook works, we are:
Turning it into organic reels
Breaking it into carousels
Recording new paid variations
Rewriting retargeting ads with the same emotional angle
Testing longer VSL openings with that identity trigger
Same hook.
Different executions.
When something gets easy, that’s when you press harder.
Low CPL isn’t the win.
Recognizing signal and exploiting it fully is the win.
I attached screenshots so you can see what the ad manager looks like and the two videos that were the winners so you can learn from them.