Write a content brief from real citations
Stop briefing writers on what you think ranks in AI. Brief them on what the models actually cite.
The prompt
Use Nimt to find the most cited pages for our tracked prompts about [topic] over the last 30 days. Open the top 10, analyze their structure, angle, and coverage, and note where our brand appears. Then write a content brief for a page that would compete: recommended structure, questions to answer, and gaps the cited pages leave open.
- Claude
- ChatGPT
- Cursor
- n8n
- Lovable
Works with any MCP client
What you get
The pages winning citations for your target topic
What they cover, how they're structured, and what they miss
Where your brand stands in the current answer set
A brief your writer can execute without further research
Coverage requirements mapped to real fan-out queries, not keyword hunches
Why this works
The brief is assembled from three Nimt layers most tools don't connect: citation data (what wins), page inspection (why it wins), and query fan-out (what the models search for behind the prompt). The writer gets a target built from model behavior instead of assumptions.