Start with the article job
Before checking individual sentences, confirm what the article is supposed to do. It should answer a specific search intent, support a priority topic, and help the reader make a decision or understand a concept.
If the article job is vague, fact-checking becomes scattered. The reviewer needs to know which claims matter most and which sections should carry the answer.
Check factual claims
Review definitions, statistics, product claims, timelines, legal or medical references, and any statement that could mislead a reader if wrong.
AI drafts often sound confident even when the claim is broad, outdated, or unsupported. Mark those claims before editing for tone.
Check sources
Prefer primary sources when available. When using secondary sources, check credibility, recency, and whether the source actually supports the claim.
A source should not be decorative. If it does not prove the exact statement being made, the claim needs a stronger source or a narrower rewrite.
Check recency
Some topics age quickly: AI tools, platform features, laws, pricing, market data, technical standards, and search guidance.
For time-sensitive claims, use current primary sources or avoid making the claim too specific. The date of the source matters when the fact can change.
Check usefulness
A factually accurate page can still fail if it is generic. Make sure the article answers the buyer question with enough specificity to be useful.
Look for examples, constraints, decision criteria, internal links, and practical next steps. These details make the page more helpful than a broad AI summary.
Prepare the handoff
The final handoff should show what passed, which sources were used, which claims were changed, and what still needs client approval.
For SEO content, the same handoff should also include title, meta description, internal links, and CMS notes so the page is ready to publish after approval.