AI audit vs rubric — AliExpress
An independent Workers AI LLM scored AliExpress 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 | 72 | 20 | -52 |
| Business transparency | 85 | 80 | -5 |
| Shipping clarity | 70 | 65 | -5 |
| Public reviews | 80 | 75 | -5 |
| Product range | 95 | 92 | -3 |
| Access & onboarding | 95 | 90 | -5 |
| Support track record | 65 | 65 | 0 |
| Store integrations | 72 | 40 | -32 |
| Overall | 79 | 66 | -13 |
What this means: Strong agreement — the rubric appears well-specified. Median per-dimension |Δ| ≤ 5.
Median per-dimension |Δ| = 5.
Pricing transparency is low because pricing is not visible without signup. Business transparency is high because the public parent company is listed and audited statements are available. Shipping clarity is moderate because delivery windows are published per region, but origins are not. Review score is moderate because the weighted average of public review sources is 4.3-4.7 stars. Product range is high because 10M+ SKUs are available. Access is high because a free plan is available with no signup required to browse. Support is moderate because mostly positive feedback is available. Integration is low because only one platform is natively integrated.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.