Standards
Editorial Policy
Our rules for independence, sourcing, conflicts, review and the boundary between observation and advice.
Editorial independence
Research questions, market prioritisation and published conclusions are set by the AI Search Asia editorial team. Commercial relationships, contributed tools, data access arrangements or speaking engagements must not determine inclusion, scoring, prompt selection or headline findings.
When access to an interface or dataset materially assists the work, we describe the arrangement and what it did not buy—coverage, ranking placement or favourable interpretation.
What we publish
We publish methods, planned samples, market briefs, governance guidance and—when release gates are met—observation summaries and datasets. We distinguish observation (what an interface returned) from inference (what we think it means commercially) and from advice (what a specific organisation should do).
Sourcing standard
Material factual claims should link to primary or authoritative sources where feasible. Vendor statements are attributed as claims, not verified facts, unless independently supported. AI-generated text is never used as evidence for a factual conclusion about market behaviour—we observe answer systems; we do not ask one to testify about another.
- Primary sources preferred for regulatory, statistical and corporate structure claims
- Secondary sources allowed for context if their role is explicit
- Paywalled or inaccessible sources cited with access date and limitation note
- Translations labelled; no undisclosed machine translation of evidentiary passages
Conflicts and external links
Relevant affiliations, commissioned work or shared infrastructure with cited organisations will be disclosed on the affected article. External links—including to related projects such as 852GEO or AI Search Lab—are selected for reader utility at the point of context. We do not operate a sitewide reciprocal-link network or exchange favourable coverage for links.
Review process
Research reports and benchmark registrations receive methodological and editorial review before publication. Every article displays an authoring desk and updated date. Named contributors appear with role and disclosure where appropriate.
AI-assisted production
We may use assistive tools for drafting, formatting or translation workflow. No published factual claim about observed answer-system behaviour relies solely on AI output without underlying observation records. Assistive drafting is subordinate to human review and the sourcing rules above.
Reprints and citation
Short quotations with link and attribution are welcome. Bulk reuse of tables, coding definitions or dataset excerpts requires permission via the contact page so we can confirm version and correction status.