Create advertising concepts and copy for the campaign below. Work as a performance marketer and brand copywriter: match the message to the audience’s actual problem, the channel’s constraints, and the evidence we can substantiate. Product or service: [PRODUCT OR SERVICE] Target audience and buying context: [TARGET AUDIENCE] Offer and conversion goal: [OFFER AND GOAL] Channel, placements, and required format: [CHANNEL AND FORMAT] Approved proof, claims, and brand voice: [APPROVED PROOF AND VOICE] Restrictions, competitors, or claims to avoid: [RESTRICTIONS] Start with a brief audience-message map: likely problem, desired outcome, strongest approved proof, primary objection, and message angle. Then produce 3 clearly different ad concepts, not minor rewrites. For each concept provide: - Angle and audience insight in one sentence. - Primary text, headline, description, and call to action in the exact length limits for the channel. If limits are not supplied, state the assumed limits. - The offer framing and the specific proof used. - A suggested visual or creative direction that does not require text embedded in the image. - One likely compliance, credibility, or audience-risk concern. Make the copy concrete and readable. Do not invent testimonials, performance results, scarcity, certifications, pricing, customer counts, comparisons, or legal claims. Avoid unsupported superlatives and personal-attribute targeting. If the brief cannot support a meaningful proof point, use a transparent product-focused message rather than fabricated social proof. Before answering, self-check every factual or comparative claim against the approved proof and confirm each CTA matches the conversion goal. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [PRODUCT OR SERVICE] | Describe what you sell, how it works, and the main customer benefit. | A scheduling tool for independent physical therapy clinics that sends appointment reminders and fills cancelled slots from a waitlist. |
| [TARGET AUDIENCE] | Describe the audience, their situation, pain point, and buying stage. | Clinic owners with 3–15 staff who lose revenue to late cancellations and still manage schedules by phone. |
| [OFFER AND GOAL] | State the offer and the action the ad should drive. | Book a 20-minute product demo; no discount or free trial. |
| [CHANNEL AND FORMAT] | Name the ad channel, placement, character limits, and asset requirements. | LinkedIn single-image ads; primary text under 150 characters, headline under 70 characters, one CTA button. |
| [APPROVED PROOF AND VOICE] | Provide approved evidence, differentiators, required language, and brand tone. | Approved proof: automated SMS reminders, configurable waitlists, and calendar integration. Voice: practical, calm, and clinical. |
| [RESTRICTIONS] | List prohibited claims, legal or platform restrictions, competitors, and topics to avoid. | Do not claim revenue increases, guaranteed no-shows reduction, HIPAA compliance, or use patient stories. |
How to use
- Paste only claims your legal, product, or marketing team can support in the approved-proof field.
- Specify the placement and length limits, because feed, search, video, and display ads require different copy structures.
- Check every promised outcome against your evidence and remove any claim that would need a qualification you have not supplied.
- Follow up with: “Adapt concept two into five headline tests that preserve the same approved proof and CTA.”
Variations
Search ads
Use when you need intent-led copy for paid search.
Write paid-search ads for [PRODUCT]. Target search themes: [KEYWORDS OR INTENT]. Audience and stage: [AUDIENCE]. Offer and landing-page action: [OFFER]. Approved claims: [PROOF]. Platform limits: [LIMITS]. Produce 12 headlines and 4 descriptions organized into three message groups: problem-led, outcome-led, and proof-led. Include suggested sitelinks only if their destination is provided: [DESTINATIONS]. Avoid keyword stuffing and unsupported comparisons. Ask up to 3 questions only if required. Self-check that every ad matches a stated search intent and uses only approved claims.
Video script
Use when you need a short paid social video concept.
Create a [DURATION]-second paid social video script for [PRODUCT]. Audience: [AUDIENCE]. Offer and CTA: [OFFER]. Approved proof: [PROOF]. Platform: [PLATFORM]. Return a scene-by-scene table with time, visual action, on-screen copy, voiceover, and purpose. Put the product problem or outcome in the first three seconds, show the product only where accurately described, and end with one clear CTA. Do not invent results, customers, or feature behavior. Ask up to 3 questions only for missing inputs. Self-check timing and claim support.
Landing page ads
Use when several ads must align tightly with a campaign landing page.
Write three ad variants that accurately pre-qualify visitors for this landing page. Landing-page copy: [LANDING PAGE]. Audience: [AUDIENCE]. Channel and limits: [CHANNEL]. Conversion action: [CTA]. Brand voice: [VOICE]. For each variant, provide primary copy, headline, description, CTA, and the exact landing-page statement that supports its central promise. Use distinct angles without changing the offer. Do not add claims absent from the landing page. Ask up to 3 questions only if required. Self-check message match between the ad and page.
Tips
- Build each ad around one approved proof point; stacking several features usually weakens the message and makes the claim harder to substantiate.
- Match the CTA to buying stage: a demo request fits evaluation, while a guide download fits early research.
- Keep the ad promise aligned with the first visible section of the landing page, or clicks may rise while conversions fall.
- Treat each concept as a distinct hypothesis, such as operational time saved versus fewer scheduling gaps, so test results teach you something.
FAQ
Can AI write ads without making risky claims?
Yes, if you provide an approved claims list and restrictions. The prompt explicitly asks the model to avoid invented results and testimonials.
How many variants should I test?
Start with a few meaningfully different angles, then test controlled headline or opening-line changes within the strongest one. Testing many near-duplicates rarely reveals why an ad worked.
Should I include competitor names?
Only include them when comparative messaging is authorized and you can substantiate it. Otherwise describe the alternative process or category without naming competitors.