Why it is written this way
When unvetted marketing copy goes live without checking for exaggerated claims, the points of failure are usually predictable: words like "best," "only," and "100%" that legally require substantiation. Writers often insert them out of habit, but consumers read them as guaranteed promises. That is why a screening stage is critical before polishing a draft.
Defining the role as a copy editor rather than a legal judge is the core premise of this prompt. Without this clear boundary, AI models often hallucinate non-existent regulations or cite fake statute numbers. Fabricated legal reasoning is not only unhelpful for review, it is dangerous. What you need here is not a legal verdict, but viable rewriting alternatives.
In the output format, the fourth column providing two alternative phrasings is the key deliverable. Merely highlighting problematic words leaves the writer stuck with blank spots. Offering alternatives—such as replacing "100% naturally derived" with "Made with XX% naturally derived ingredients"—allows for immediate editing. The two summary lists below the table serve the same purpose, distinguishing between claims that can be salvaged with documentation and those better dropped entirely.
The final verification self-check ensures the alternatives do not water down the copy into meaningless fluff. Wrapping the input in """ delimiters ensures that imperative sentences within the ad (like "Buy now!") are treated as text to evaluate rather than instructions to execute.
Unfamiliar terms? See Aha AI: self-consistency, prompt-injection
Compared with a bad example
Check if this ad copy has any legal issues.
(paste ad copy)
When asked for legal advice, AI tends to fake authority. It either returns a vague "looks completely fine" or fabricates plausible-sounding legal clauses. Neither can be trusted. Crucially, it fails to provide rewritten copy, leaving the problematic draft completely intact.
Variations
Quick Alternative Generation
Review the {{industry type}} ad copy below, identify any unverified or absolute claims, and suggest two alternatives for each that keep the original marketing intent intact. Do not include detailed explanations; provide only a concise list formatted as "Original Phrase → Alternative 1 / Alternative 2."
""" {{advertising copy}} """
Use this when you already know which parts are risky and just need quick phrasing options to choose from.
Full Landing Page Audit
Below is the full text for a {{industry type}} landing page. Identify all claims that require verification and sort them by risk level from highest to lowest. Format the output as a table with the columns: "Location (Paragraph #) | Original Text | Flagged Phrase | Alternative Phrasing". Group similar types of claims together. Provide a prioritized review checklist rather than definitive legal rulings.
""" {{advertising copy}} """
For long-form copy, large tables become unwieldy. Grouping by claim type and sorting by risk helps prioritize revisions from top to bottom.
Model notes
This prompt is designed for editorial refinement, not legal advice. Use the suggestions as draft alternatives, and always verify final copy with your legal or compliance team before launching.
AI models may occasionally attempt to cite specific legal statutes or regulatory clauses. If this happens, disregard the legal citations and prompt: "Exclude all references to legal clauses and provide only the copy critique and alternative phrasings."
Related prompts
Last updated 2026-09-02 · Found a mistake? Let us know