Content Writer
From idea to article
The end-to-end walkthrough — pick an idea, steer the research and outline, and export a draft written with live keyword and SERP context at its side.
The walkthrough
Pick an idea (or add your own)
Start from the Pipeline: the LLM-gap ideas are the highest-leverage, since each one maps to prompts where AI currently recommends someone else. Or click Add Idea and give it a keyword, title, question, or a competitor URL to analyze.
Read the AI Insight, then start research
The idea's context, strategic move, and angle tell you what the piece is for. If the angle's wrong, this is the cheapest moment to redirect it.
Steer the research brief
Read the strategist's comments, reply where you know better, resolve what you accept. The brief's searcher-intent and AI-visibility sections are also a useful sanity check: is this really a gap worth closing? (Full detail: the editorial process.)
Approve the outline
The thesis and section strategy carry their reasoning as comments. Push back here — one outline comment is worth a thousand draft edits.
Review the draft with its evidence alongside
The draft arrives written against the outline, with sources linked inline — and the panels that keep it honest docked on the right.
Export
Take the finished piece out through the Export button and publish it in your CMS.
The draft view
Two panels ride along with every draft:
- Keywords — the tracked keyword set for the piece, with volumes, so coverage is visible while you edit.
- SERP — the current top 10 for the primary keyword: the actual competition this piece has to beat, not a guess about it.
Claims in the draft link to their sources inline, and the word count reflects what the outline demanded rather than a padding target.
What "done" looks like
A piece is ready to ship when the outline's argument survived your review, the draft's claims carry their sources, and the brand voice matches — which it should by construction, since tone, phrasing rules, and proof points come from your Brand Vault.
After publishing, close the loop with the data side: track the prompts the piece was built to win (see setting up your prompts) and watch whether AI answers start citing it. That's the metric this pipeline exists to move.