AI audit vs rubric — CJDropshipping

An independent Workers AI LLM scored CJDropshipping against the same published rubric. The deterministic rubric result is our canonical score. The LLM's result is shown here as a sanity check — never mixed into the scoring formula.

Dimension Rubric LLM Δ (LLM − Rubric)
Pricing transparency 84 20 -64
Business transparency 40 25 -15
Shipping clarity 92 85 -7
Public reviews 88 0 -88
Product range 95 92 -3
Access & onboarding 95 90 -5
Support track record 80 80 0
Store integrations 85 75 -10
Overall 81 63 -18
What this means: Moderate disagreement — rubric bands may need tightening. Median per-dimension |Δ| is between 5 and 15. Median per-dimension |Δ| = 8.5.
Pricing transparency is low because the source price is not visible. Business transparency is low because the company is not publicly listed and audited statements are not available. Review score is 0 because the hint is below the threshold. Product range is high because the catalog size estimate is 10M+. Access is high because a free plan is available without signup. Support is high because the hint is above the threshold. Integration is high because 5 platforms are supported.
This is the LLM's own explanation, not editorial commentary from SupplierSpy. The LLM result is a sanity check on the rubric — never mixed into the scoring formula.