What this gives you:
A Source Ledger to record where each fact came from, and a Claim Check table to mark draft claims as Supported, Partial, or Unsupported. Use them together so a real source trail exists before the final SEO and publishing pass.
Download the templates
How to use them
- Research before drafting.
Collect useful sources first instead of asking a model to write and invent the evidence trail afterward. - Give every source an ID.
Use IDs like S001 and S002 so draft claims can point back to evidence without copying full URLs everywhere. - Extract factual claims after drafting.
Pull out prices, dates, capabilities, policies, research findings, and other checkable statements. - Match each claim to source IDs.
Mark it Supported, Partial, or Unsupported. Repair only the weak parts. - Recheck volatile facts.
Prices, plan limits, policies, and feature availability should be checked close to publication. - Get human approval.
Do not let a green automation run substitute for final judgment.
These are sample files. Replace the example rows with your own research. A source being real does not automatically mean it supports the wording of your claim.
Want the full workflow?
The complete AI Research-to-Publish Engine adds the Article Brief, full prompt sequence, operations guide, publishing checklist, n8n starter workflow, and the complete editable template set.
AI Research-to-Publish Engine
Research → sources → draft → claim check → SEO → human review. One-time USD 29 digital download.
See the full workflow kit — USD 29