← Back to the article

Audits

Seattle's office vacancy is at a record high and among America's worst.

Version 2. FactHub's auditors are automated reviewers, separate from the process that writes articles. Each checks a fixed, published list of things and grades the results with its own published rule. They can be wrong; an audit is not a human expert review. All auditors

Across versions

Version First Look
Version 2 5/5

First Look

5 of 5 checks passed · version 1 · claude-haiku-5-5 · 14 October 2026

A: Passes the first look: nothing obviously wrong. Minor notes may be reported as feedback without lowering the grade.

5 checks
  1. Pass Title matches the evidence

    Title's record-high claim rests on the downtown figure (about 36% against a roughly 21% prior record) and the ranking supports 'among America's worst'; the scope difference is a judgment call, not an obvious mismatch.

  2. Pass No obvious contradictions

    Finding Mostly true matches the answer (highest or second-highest depending on tracker) and the conclusion, which explains the ranking differences; the key points and evidence agree on the figures.

  3. Pass Plausible justification

    Conclusion gives a clear reason for Mostly true: the ranking depends on which tracker and geography is used, while the severity and direction hold across sources.

  4. Pass Sources are relevant

    Cushman & Wakefield national and local reports, Yardi Matrix, CommercialCafe, CBRE, Kidder Mathews, and Daily Journal of Commerce and Seattle Times coverage are all on topic for office vacancy in Seattle and the national comparison.

  5. Pass Common questions answered

    Five objections are present and each response addresses its question, including the San Francisco ranking, the downtown-only figure, recovery, broker bias and taxes.

Coverage: Read the full article payload (title, claim, answer, finding, conclusion, evidence, objections, sources with names/titles/descriptions/roles). Sources judged by name/title/description/role only, per First Look's rules; no source snapshots read and no web access used.