Editorial policy

Paz.ai publishes product, methodology, and educational content about AI shopping visibility. This page explains the standards we hold that content to: primary sources, named accountability, evidence, proprietary research disclosure, corrections, and dates.

Published 2026-08-15. Last reviewed 2026-08-15.

Primary-source policy

Claims about platform behavior and specifications are sourced from official primary sources: platform documentation, official specifications, and peer-reviewed research. We link to the source so a reader can verify it.

Primary sources describe platform behavior. They do not prove that a specific optimization causes a Paz citation. Any causal visibility claim relies on Paz's own published methodology and data.

We do not present a vendor marketing page as a primary source for a factual platform claim. Where a platform's own documentation is the source, we link to that documentation directly.

Author and reviewer accountability

Each methodology, guide, and research page names a responsible team as author and a reviewer. Product claims are checked against documented product capabilities and cited primary sources.

Pages that state what Paz does are reviewed before publication. Pages that state platform specifications are checked against the linked primary source.

Vendor claim, Paz observation, and external fact

We separate three kinds of statements and label them so a reader knows what kind of evidence backs each one.

  • Vendor claim: a statement a platform makes about itself. We attribute it to the platform and link to where the platform says it.
  • Paz observation: something Paz measured, such as a found rate or a position trend. We label it as a Paz observation and publish the sample size, query set, engine set, geography, collection dates, and limitations.
  • External fact: a statement from an independent, cited source such as peer-reviewed research or official documentation. We link to the source.

Evidence and proprietary research

Where Paz publishes proprietary research, such as a visibility finding or a catalog-completeness observation, we publish the sample size, query set, engine set, geography, collection dates, and limitations. Without those, we remove the number rather than publish it unqualified.

We do not claim a fixed attribute-count enrichment (for example, "from 5 to 47+ attributes") without a published methodology and a downloadable aggregate table. Catalog optimization proposes attribute improvements for merchant review; it does not publish a fixed enrichment count.

We do not publish self-awarded "best" rankings or aggregate-rating markup for Paz. A self-authored page should make Paz eligible for an answer by defining the category and showing a verifiable fit, not by declaring itself best.

Publication and review dates

Each methodology, guide, and research page shows a publication date and a last-reviewed date. When a page changes materially, we update the last-reviewed date.

Pages that state platform specifications show a last-reviewed date for each cited source. The date shows when the cited claim was last checked.

Sitemap lastmod is generated from meaningful content changes where feasible rather than a broad deployment timestamp.

Corrections

If a claim is inaccurate or a source has moved, we correct the page and update the last-reviewed date.

To report an error, email editorial@paz.ai. We review reports against documented product capabilities and cited primary sources.

Related