Asking for a perfect result in one message gets you a first draft dressed as a final one. Split it: one message to draft, one to critique against your criteria, one to rewrite. You see every step, and you can stop at any of them.
Write the perfect 30-second script for our new juice launch.
Message 1: draft three script options. Message 2: score each against our brief. Message 3: rewrite the strongest one using the critique.
The prompt
Copy-ready prompt
MESSAGE 1 — draft Write [NUMBER] options for [DELIVERABLE] based on this brief: [BRIEF]. MESSAGE 2 — critique (send after reading the drafts) Score each option from 1 to 5 against these criteria: [CRITERION 1], [CRITERION 2], [CRITERION 3]. Quote the line that earns or loses each point. Do not rewrite anything yet. MESSAGE 3 — refine Rewrite option [NUMBER] using only the critique above. Keep what scored 5. Fix what scored 3 or lower.
Why it works
Anthropic’s guide calls this the most common chaining pattern: “generate a draft → have Claude review it against criteria → have Claude refine based on the review,” with each step separate “so you can log, evaluate, or branch at any point.” Google’s Gemini guide recommends the same: break complex tasks down and chain prompts so each output becomes the next input.
Use it for
Scripts, taglines, campaign concepts — anything where you would normally ask for “the best one.” Write the criteria yourself; they are where your brief lives.
Sources
- Anthropic — Claude prompting best practices — platform.claude.com (accessed 29 Sep 2026)
- Google — Prompt design strategies for the Gemini API (updated 17 Sep 2026) — ai.google.dev (accessed 29 Sep 2026)
More tricks and templates in the Prompt Lab. Want this applied to your brand’s brief? Start a project.