UGC Ad Script Formulas Platforms Keep Recommending

Most underperforming UGC ads fail in the first two seconds. The hook is vague, the camera is centered like a corporate headshot, and the viewer has already swiped. The script was the problem long before anyone hit record.

The structures below are not secret. Platform creative teams hand them out in their own ad libraries and best-practice docs, top-performing creators reuse them constantly, and they survive contact with a cold audience. What follows is six of them, written out as templates you can fill in, plus the parts people usually get wrong.

Why structure beats inspiration for paid social

An organic creator can post forty times and let one hit carry the month. A paid ad does not get that runway. It gets a fixed budget, a cold audience, and a few hundred impressions to prove it deserves more. The script has to do the work on the first try.

That is why structure matters more than a clever line. A repeatable skeleton lets you produce ten variants that each test one thing: a different hook, a different proof point, a different close. You are not hunting for one perfect ad. You are building a deck of competent ones and letting the auction tell you which to scale.

Each formula below has the same shape: a hook that earns the next three seconds, a body that creates tension or proof, and a close that names the next action. The differences are in how they open and what kind of pressure they apply.

The six formulas, with fill-in templates

Read each one as a skeleton, not a script. The brackets are yours to replace. Keep total runtime between 15 and 30 seconds for paid placements; longer only earns its keep if retention holds past the halfway mark.

1. Problem -> Agitate -> Solve (PAS)

The workhorse. Opens on a frustration the viewer already feels, makes it slightly worse, then resolves it. Works because it mirrors the actual emotional sequence of wanting something.

• Hook: "I wasted [time/money] on [common wrong approach] before I figured this out."
• Agitate: "The annoying part is [specific consequence] — and it kept happening every [frequency]."
• Solve: "Then I started using [product] to [specific job]. Now [concrete new state]."
• Close: "[Action], it's [the link / the bio / on sale]."

Failure mode: agitating a problem the product does not actually solve. The middle has to point straight at the feature in the close.

2. The 3-Reasons list

A list ad feels less like an ad because it promises structure up front. The viewer knows it ends and stays to see all three. Strong for products with a few distinct, easy-to-grasp benefits.

• Hook: "Three reasons I won't go back to [old way / competitor category]."
• Reason 1: "[Benefit] — [one line of proof or a quick demo]."
• Reason 2: "[Benefit] — [proof]."
• Reason 3: "And the one that sold me: [strongest benefit]."
• Close: "Linked below if you want to try it."

Keep each reason under five seconds. Save the strongest for last so retention climbs instead of leaking.

3. The honest review / "I was skeptical"

Addresses the doubt before the viewer raises it. The skepticism in the hook is the credibility. It reads as a real person reporting back, not a brand asserting.

• Hook: "I did not think [product / category] would actually work. I was wrong."
• Doubt: "I figured it would be [the cynical expectation everyone has]."
• Turn: "What changed my mind was [the specific moment it proved itself]."
• Close: "If you've been on the fence, this is the one."

This one lives or dies on the specificity of the turn. "It just works great" is not a turn. "It cut my edit time from an evening to ten minutes" is.

4. Before / After (the transformation)

Visual and blunt. Best when the result is something you can show on screen rather than describe. The contrast does most of the persuasion; the voiceover just labels it.

• Hook: "This is what [the task / the result] looked like before." [show the bad state]
• Bridge: "Here's the only thing I changed." [name the product, show it in use]
• After: "And this is now." [show the result, hold on it]
• Close: "[How to get it.]"

Do not narrate over the after-shot too quickly. Let it sit for a beat. The silence reads as confidence.

5. The "POV / a day in the life" framing

Native to TikTok and Reels because it borrows the format people already scroll past for fun. The product appears inside a moment instead of being announced. Lower hard-sell pressure, higher watch-through.

• Hook: "POV: you finally [desirable situation the product enables]."
• Scene: "[Walk through the moment, product used naturally, no pitch.]"
• Payoff: "[The small satisfying result.]"
• Close: "[Soft mention] — [product] is how."

Weakest direct-response intent, so pair it with a retargeting audience or a strong on-screen caption that carries the call to action.

6. The "stop doing X" pattern interrupt

A command in the hook stops the thumb because it sounds like advice, not a sale. Works when there is a common mistake your audience is actively making.

• Hook: "Stop [the common mistake] if you want [the outcome]."
• Why: "Here's why it backfires: [the mechanism]."
• Fix: "Do this instead: [the product as the better way]."
• Close: "[Action.]"

Use sparingly across an account. If every ad opens with a command, the pattern interrupt stops interrupting.

An annotated example, start to finish

Here is PAS filled in for a hypothetical project-management app, with notes on each line so you can see the decisions.

• Hook: "I ran three projects on sticky notes for a year. It cost me a client." (Specific, slightly painful, names a real stake — not "staying organized is hard.")
• Agitate: "Every Monday I'd rebuild the same list from memory and forget something by Wednesday." (Concrete and rhythmic; the viewer recognizes the loop.)
• Solve: "Now everything lives in one board and nothing falls through. Setup took ten minutes." (Names the new state plus removes the friction objection.)
• Close: "Free to start — link's in the bio." (One action, no menu of options.)

Notice what is missing: no feature list, no music swell, no "and that's not all." Each line does one job. That restraint is what lets the same skeleton be re-shot ten ways without sounding like ten ads.

Turning a script into a finished ad without a shoot

A script is half the job. The other half is producing enough variants to actually learn something, which is where most small teams stall. You write three good hooks and then spend a week trying to film them.

Use this checklist before anything goes live:

1. Hook in the first 1.5 seconds. If the first line is setup, cut it. Start on the tension.
2. Burned-in captions. Most paid social plays muted by default; the script has to read silently.
3. One claim per ad. If you are testing a hook, keep the body and close fixed across variants.
4. Correct aspect ratio per placement. 9:16 for TikTok, Reels, and Shorts; 1:1 or 4:5 for feed; 16:9 only where the placement demands it.
5. A spoken or on-screen call to action. "Link in bio" is not optional just because the button exists.
6. At least three hook variants per concept. The hook is the single highest-leverage line; test it hardest.

The throughput problem is real. A creator-led shoot for six scripts in three aspect ratios is days of coordination. That is why AI-generated UGC has become a practical first pass — not to replace your best human creators, but to find which scripts deserve a real shoot before you spend on one.

FAQ

How long should a UGC ad script be?

For paid placements, aim for 15 to 30 seconds of spoken content, which is roughly 40 to 80 words. Longer scripts can work for retargeting or warmer audiences, but for cold prospecting, every extra second is a chance to lose the viewer. Write tight, then cut the first line.

Do AI-generated UGC ads actually perform against real creators?

They perform well enough to be your testing layer. AI ads are reliably good at volume and speed, which is exactly what the variant-testing phase needs. Human creators still win on authenticity for top-of-funnel scale once you know which message converts. The honest play is AI for breadth, humans for the proven winners.

Which formula should I start with?

Problem -> Agitate -> Solve, almost always. It is the most forgiving structure, it fits nearly any product, and it forces you to articulate the actual pain you solve. Once you have a PAS baseline that converts, branch into the list and the honest-review formats to fight ad fatigue.

Writing the script is the part you should own. Producing ten captioned variants in every aspect ratio is the part worth automating — and that is the job Aitachyon was built for: paste a URL, get a finished, captioned video ad in about two minutes, in 9:16, 16:9, or 1:1. Use the formulas above to write the hooks, and let the variants compete.