The query represents a real question
We check suggested queries and meaningful variants instead of inventing long phrases from product features.
EDITORIAL SYSTEM
Question the candidate
before writing the answer.
AI provides speed; primary evidence and counterexample gates provide control. This is how ToolCriteria selects, updates and sometimes merges research.
BEFORE PUBLISHING
Six publication gates
A candidate that fails a gate is held or rejected before scoring. Avoiding the wrong article matters more than increasing article count.
The search result can be improved
We inspect whether current results leave a comparison, source or decision gap that this article can fill.
The intent is singular
Incompatible goals become separate pages; an intent already covered by an article is merged instead of duplicated.
Evidence exists before drafting
A consequential claim needs primary or public evidence. Missing evidence is shown as uncertainty, not filled by guesswork.
The question leads to a decision
The article must help a reader compare, test, configure, start, stop or update something without distorting the query.
Rejections are recorded
We preserve why a candidate was accepted, held or rejected so AI does not repeatedly propose the same weak idea.
AFTER PUBLISHING
Improve technology, search fit and evidence in parallel
CHECK 01
Publication quality
Verify canonical, hreflang, sitemap and internal links.
CHECK 02
Search recognition
Review indexing, early queries and search snippets.
CHECK 03
Reader fit
Add decision material where impressions and intent show a gap.
CHECK 04
Evidence freshness
Reassess the conclusion after price, feature or term changes.
CONTENT PORTFOLIO
Every article has a job
Guide
Procedures, settings, prices and terms: answer a concrete question and connect it to a decision.
Comparison
Choices, alternatives and thresholds that cannot be resolved from a search snippet alone.
Deep dive
Combine distributed primary sources into a process, checklist or model.
Update watch
Track price, term and program changes and show who is affected from which date.
AI & QUALITY CONTROL
What AI does—and what stops publication
AI assists with
- Query and intent variation collection
- Primary-source extraction and normalization
- Conflict, change and missing-data detection
- Conditional analysis, calculations and FAQ drafts
- Search-performance classification
Publication gates verify
- Actionable steps and concrete examples
- Claim-to-primary-source mapping
- Article-specific depth and originality
- AI, advertising and uncertainty disclosure
- Duplication, orphan URLs and structured data
The scope of AI assistance and the conditions that stop publication are public.
Read the editorial & AI policy ↗