Source-of-Truth Matrix for AI-Assisted Publishing
Define approved facts, claims, examples, CTA routes, and exclusions before using AI to draft or rewrite public website copy.
Information asset -> diagnostic/template/inquiry route
The Prompt To Publish Qa Lab library is designed as a working shelf, not a link directory. Start with Source-of-Truth Matrix for AI-Assisted Publishing when the problem is still broad, then use Draft Diff Checklist for AI-Assisted Website Copy or Factuality Gate for Public Website Drafts to turn the finding into a decision record.
A good pass through this library should produce one artifact: a checklist result, scorecard, route map, or repair queue that another operator can review. When the artifact shows repeated unknowns, use the services route with concrete examples; when it resolves the issue, keep the page as a reference and move to the next bottleneck.
For backlink value, each page should be useful without a sales conversation: it must define the failure mode, show the fields to inspect, and leave the reader with a reusable operating object.
Define approved facts, claims, examples, CTA routes, and exclusions before using AI to draft or rewrite public website copy.
Compare an AI-assisted draft against approved source notes with a checklist for facts, claims, CTA routes, schema, and brand boundaries.
Use a factuality gate rubric to check names, dates, claims, proof, pricing, sensitive wording, and unsupported AI-generated details.
Review public website copy for self-promotion risk, brand bleed, fake proof, route confusion, and CTA mismatches before publishing.
Prepare rollback notes, approval evidence, changed URLs, measurement events, and owner decisions before releasing public website changes.
If several pages in this library point to the same missing evidence, broken route, or measurement gap, turn the finding into a scoped diagnostic note before adding more pages.