PARTLY

Partly true: AI can sort a brand's Meta ads by hook and offer for pennies, but 'Hormozi's whole strategy in one picture' rests on an assumption, and the 'new AI model' is a waitlisted tagging tool

Facebook · Nov 2026

The claimA new AI model analyzed every ad in a famous marketer's Meta ad library by format, hook, offer and copy, plotted them by how long each has run, revealed the winning pattern, and did it all for 49 cents. Comment a keyword to get the exact system.

The workflow is real and cheap. Meta's public Ad Library shows every active ad a page is running and when each started, and AI models can tag ad text and transcripts by format, hook, offer and angle for fractions of a cent each, so a bill under a dollar for one brand is plausible. The model named in the reel appears to be a new decision-style model in early access through a waitlist; independent reviews and official documentation were hard to find, and secondary write-ups say it reads text and structured data rather than watching video, so video ads need transcripts first. The weaker part is the conclusion. 'The longer an ad runs, the better it performs' is a popular rule of thumb, not proof: the Ad Library does not show spend or results, and large advertisers often keep evergreen or brand ads running regardless of return. So the chart shows what a brand keeps running, not what makes money. The reel ends by trading the system for a comment, a common lead-capture tactic.

What holds up

Meta's Ad Library publicly shows active ads and their start dates
AI tagging of ad copy by hook, format and offer is cheap and works
Sub-dollar cost for analyzing one brand's ads is plausible

What doesn’t

Long-running ads are not proven winners; the library shows no spend or results
The named model is early access with little independent documentation
Text-only models need transcripts to analyze video ads

The catch

'Running longest' is not the same as 'making the most money.' Without spend and results, the chart shows what a brand keeps live, so treat the patterns as ideas to test, not a copied strategy.

How to actually do it

Open Meta's Ad Library, search the brand, and note each active ad's start date and format
Copy the ad text and transcribe video ads, then ask an AI model to tag hook, offer and format
Group the long-running ads and turn the patterns into two or three ad tests of your own
Judge the tests on your own cost per result, not on the other brand's run time

Do the same research with any capable AI model: export or copy the ad text, add transcripts for video ads, tag them, and use the patterns as test ideas for your own ads.

Effort: About 1 to 2 hours per brand
Cost: Under $1 in AI usage per brand; the Ad Library is free
Stack: Meta Ad Library, a transcription tool and any capable AI model
Confidence: Low
Posted by: a marketing automation creator offering a DM'd system for a comment, Facebook reel (about 8,800 views); reel sent to Buddy Sun, Nov 1, 2026, 8:23 PM

See the original claim →

We test hype for free. We build the real thing for a living.

Thirty minutes, no pitch — and you'll leave with something useful either way.

Book a call with Todd or start with the free Business Checkup →

The Verdict Weekly

Three verdicts every Friday. Free forever, unsubscribe anytime, no spam — that would be ironic.