ADS & LANDING PAGES

Ad copy compliance and claim check

Use before ads go live in a regulated or claim-sensitive category.

ad-copy-compliance-and-claim-check.md

Download .md

You are a compliance reviewer checking ad copy before it goes live. Be strict and flag anything doubtful.

Ad copy to review: {{AD_COPY}}
Sector and regulator or platform policy that applies: {{SECTOR_AND_POLICY}}
Country or region: {{REGION}}
Evidence I hold for each claim: {{EVIDENCE}}
Words and claims already banned internally: {{BANNED_TERMS}}

Produce a table: Line, Claim made, Type (factual / comparative / superlative / implied / testimonial / pricing), Evidence I supplied, Verdict (supported / unsupported / needs qualifier / rewrite), Suggested fix.

Then:
1. List every unsupported claim separately, ordered by how likely it is to cause a problem, with the reasoning for that order.
2. Flag every implied claim - things the copy makes a reader believe without stating outright. These are the ones people miss, so be thorough.
3. Rewrite the copy in a compliant version that keeps as much persuasive force as possible.
4. List what evidence I would need to gather to restore each claim you had to weaken.

Rules:
- Do not tell me something is compliant if you are unsure. Say what you are unsure about.
- You are not providing legal advice; say so once at the end and recommend a qualified review for anything you marked high risk.
- Never invent a regulation, clause or policy number. If you do not know the exact rule, describe the principle instead.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{AD_COPY}}
{{SECTOR_AND_POLICY}}
{{REGION}}
{{EVIDENCE}}
{{BANNED_TERMS}}

Getting a better result

Attach the actual evidence, not a summary of it, or the check is theatre.
Implied claims cause most of the trouble - read that section first.
Treat this as a filter before human review, never as a replacement for it.

Questions about this prompt

When should I run this rather than sending copy straight to review?

Before ads go live in a category where a claim can cost you, as a filter ahead of human review rather than in place of it. Sending unfiltered copy to legal or a regulator burns goodwill on problems you could have caught. It is strict by design and will flag things an experienced reviewer would wave through.

What do I need in front of me?

The copy, the sector and the regulator or platform policy that applies, the region, your internal banned terms, and {{EVIDENCE}} attached as the actual evidence rather than a description of it. A line saying we hold data supporting this gives the check nothing to compare against and turns the exercise into theatre.

What comes back?

A claim by claim table with type and verdict, unsupported claims ordered by how likely each is to cause trouble, an implied claims section, a compliant rewrite and a list of evidence to gather. Read the implied claims first. Those are the ones reviewers miss, because the copy never states them and the reader believes them anyway.

What is the mistake?

Treating a clean output as clearance. It is not legal advice, it says so at the end, and it will never invent a regulation or clause number, which means it can be right about the principle and wrong about the specific rule where you operate. Anything marked high risk goes to a qualified reviewer with the evidence.