Equipment Insider //Heavy equipment intelligence Account

Data policy / Version 1.0

Every claim keeps
its lineage.

AI may accelerate discovery and extraction. It does not become the source, invent missing values, or silently resolve contradictions.

01

Source hierarchy

Official OEM product pages, manuals, spec sheets, service and safety materials are preferred for machine facts. Dealer listings, auction results, rental listings, owner communities, videos, and forums serve different evidence purposes and remain labeled.

02

Claim-level provenance

Material facts retain source URL, observed time, source text, extraction method, review state, and confidence. Conflicting claims stay in review instead of being averaged into false certainty.

03

AI-assisted research

Research agents discover records, collect permitted pages, extract candidate claims, normalize units, and flag conflicts. Publication requires deterministic validation or human review appropriate to the claim’s risk.

04

Scores and sentiment

The 1–5 Owner Score comes only from qualified explicit reviews. The 0–100 Market Sentiment signal comes from separately qualified public evidence. A reviewed observation can publish from the first data point, but sample size and confidence always travel with the number. Scores are never inferred from specifications or marketing language.

05

Corrections

Anyone may challenge a factual record with evidence. The original value remains visible until review is complete. Accepted corrections update the record, provenance, and audit trail; they do not erase legitimate historical observations.

06

Rights and licensing

Public facts are distinguished from copyrighted expression, private submissions, and licensed data. Raw source captures and private proof are access-controlled. Commercial feeds, bulk exports, and derived products require rights review before release.

Submit evidence or a correction →