Publication Information

Editorial Policy: How Content Is Checked (2026)

AI Content Compass evaluates content through source support, reader usefulness, and the fit between a recommendation and its stated scenario. Product facts, editorial judgments, and illustrative examples should be distinguishable throughout a page.

Sources and Product Claims

Assess the recommendation through its evidence and stated reader scenario, rather than unsupported ratings or promises.

Current product claims should be checked against official feature pages, pricing terms, support documentation, and other appropriate primary references. A source establishes only what it actually says: plan access, billing period, version, geography, and required setup can change the meaning.

Merchant descriptions are evidence of documented capabilities and terms, not independent proof of universal performance. Articles should not turn a capability into a ranking guarantee or a customer case study into a promise for every publisher.

Source links should remain close enough to the relevant explanation to be useful. Important claims in tables, captions, metadata, and structured data need the same review as the main prose. A working citation alone does not establish that the wording is supported.

Comparisons and Recommendations

Comparisons should apply matching criteria to the products discussed. Recommendations should identify the audience and workflow that make the tool relevant. A product can be a sensible option for one publishing process and a poor fit for another.

The publication should not invent numerical ratings, output benchmarks, prices, testimonials, discounts, or results. Editorial use-case picks are qualitative judgments, with their reasoning explained. They should not be presented as measured universal rankings.

Examples and Visualizations

Illustrative examples explain a method using explicitly stated assumptions. They are not customer results or observed performance. Calculations should show enough information for readers to understand the units and arithmetic.

Numerical visualizations require verified values, labels, a meaningful scale, and a readable data equivalent. Where no suitable dataset exists, a process or relationship diagram can explain the topic without fabricated scores. Generated illustrations should not be presented as real screenshots or evidence.

Decision guide: editorial criteria, not measured performance scores.
Content TypeEditorial Standard
Product factAppropriate current source and retained restrictions
RecommendationExplained criteria and reader scenario
ExampleClearly labeled assumptions
ChartVerified values or labeled non-numeric process
CorrectionEvidence and all affected page elements

Corrections and Review Dates

A review date should reflect an actual review of the relevant content. Replacing a year token without checking the article is not an update. Product information can change after review, so readers should check current merchant terms before purchasing.

Corrections should identify the original statement, supporting evidence, and affected page elements. When a fact changes, related tables, FAQs, metadata, and conclusions should be reviewed too. See Contact for support routes and a correction-report helper.

Your Next Step
Explore Koala AI