+ Clearly differentiated all angles and checked overlap.
- The extra validation notes make the response too long.
Produces genuinely different approaches and flags when two variants are the same idea reworded.
| Category | Marketing › Ad copy |
|---|---|
| Tags | DraftingIdeationMarketer |
Write ad copy from five angles. Angles: 1. **Problem** — name the friction the customer feels 2. **Outcome** — what it is like after 3. **Comparison** — against what they do now 4. **Specifics** — lead with a number or a condition 5. **Direct address** — "if you are someone who…" Per variant: - Headline (within the Char limit given below) - One line of sub-copy - When this angle works Rules: - ***The five must genuinely differ.*** **If two say the same thing in different words, say so and replace one.** - **Do not use a number or claim not supported by the strengths I gave you.** - No superlatives.
Five variants that say one thing test nothing. This checks the angles against each other and keeps every claim inside your stated strengths.
Claude is strongest in evidence use and angle differentiation. ChatGPT is concise but repetitive, while Gemini is specific yet contains several unsupported claims.
+ Clearly differentiated all angles and checked overlap.
- The extra validation notes make the response too long.
+ Delivers every required field in a concise format.
- Problem, outcome, and comparison repeat filter upkeep.
+ Connects the target to specific search situations.
- Adds unsupported claims about delivery, hygiene, and health.
| Criterion | ChatGPT | Claude | Gemini | Leader |
|---|---|---|---|---|
| Instruction following | 6 | 10 | 9 | Claude +11% |
| Accuracy | 10 | 10 | 5 | Tie |
| Specificity | 8 | 10 | 9 | Claude +11% |
| Structure | 9 | 9 | 8 | Tie |
| Right length | 10 | 7 | 6 | ChatGPT +43% |
Scored 1–10 by gpt-5.6-sol with model names hidden (2026-09-24). This is an AI review, not a measurement.
We gave three models the same input and copied their answers unedited. Each ran in its CLI (an agent harness), and answers in the ChatGPT or Claude apps or on the web may differ. Outputs are in Korean.
Character limit: 헤드라인 20자 이내 Product: 가정용 정수기 구독 서비스 '퓨어웰' Audience: 어린 자녀를 둔 30대 신혼부부 Strengths: 월 19,900원 정액제, 6개월마다 필터 자동 교체, 설치비 무료, KC 인증 완료 Medium: 네이버 검색광고