Article generation

What "article" means here

In the content engine, Article is a long-form piece — guides, deep dives, evergreen reference content. The default length target is 1,200–1,800 words. The engine writes at the length the brief justifies; it does not pad.

Starting a brief

Content → New → Article. Three required fields:

• Working title. A direct, claim-style title beats a clever one for AI citation. "How to evaluate a CRM" beats "The great CRM hunt."
• Target prompt. The buyer-intent query you want this article to win. Pick one — articles that try to win five prompts win none.
• Outline angle. One sentence. "Start with the cost framing because the audience is bootstrapped."

Optional fields refine the output: target reading level, examples to draw on, anti-examples to avoid.

How the engine drafts

For each article, the engine:

1. Loads your company positioning and tone profile from workspace settings.
2. Reads the target prompt and pulls related prompts from your visibility data.
3. Researches the prompt by examining the top-cited pages from the visibility test results — so the draft is informed by what's currently working in your category.
4. Builds an outline (H2 + H3) and asks for your approval if Outline first is on.
5. Writes the draft section by section, keeping each section under 250 words for citation-friendly chunking.
6. Adds an FAQ block with 3–5 buyer questions and concise answers.
7. Runs compliance.

Best practices for citation-worthiness

• Lead with a one-paragraph answer. AI engines preferentially cite content that resolves the prompt in the first 100 words.
• Use semantic H2s. "How to choose a CRM" cites better than "Choosing the right tool."
• Include concrete numbers. "47% of users abandon checkout" cites better than "many users abandon."
• End with an FAQ. FAQ schema is heavily favoured for direct-answer citation.

The engine bakes these in by default. You can override the FAQ if you don't want one.

Tone defaults

The content engine reads your tone profile (set during onboarding, editable in Settings → Company → Voice) and reflects it in:

• Sentence length distribution.
• Use of contractions.
• Industry jargon level.
• Use of first vs. second vs. third person.

If output sounds off, refine the tone profile rather than re-prompting individual drafts.

When to use a comparison instead

If the target prompt is "X vs Y," use a comparison page, not an article. The two structures look superficially similar but the comparison engine handles competitor mentions safely and adds the right side-by-side structures.

When to use a service page instead

If the target prompt is "buy [thing]" or "[service] in [location]," use service pages or location pages. Articles target informational intent; service / location pages target transactional intent.

Auto-revision

Reviewers leave notes inline. Auto-revise regenerates the draft with the notes as constraints. See Auto-revision.

Publishing

Approved articles publish via any connected integration. The engine maps fields automatically; you can preview the destination's render before clicking Publish. See WordPress troubleshooting for the most-asked-about destination.

Bulk article generation

Domination supports bulk briefs: upload a CSV of titles + target prompts, and the engine produces drafts in parallel up to 20 at a time. Useful for filling a content calendar in a single session.