AI audit vs rubric — Spocket
An independent Workers AI LLM scored Spocket 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 | 86 | 0 | -86 |
| Product range | 85 | 70 | -15 |
| Access & onboarding | 65 | 40 | -25 |
| Support track record | 82 | 65 | -17 |
| Store integrations | 85 | 60 | -25 |
| Overall | 76 | 48 | -28 |
What this means: Large disagreement — investigate. The LLM read the published signals very differently from the deterministic rules.
Median per-dimension |Δ| = 21.
Spocket's pricing is opaque, requiring signup to view pricing. Business transparency is low, with no publicly listed company or audited statements. Review score is missing, so it's scored as 0. Product range is estimated at 7M+ SKUs, but this is not a direct match to the rubric. Access is restricted, requiring signup to browse. Support is mostly positive, but not consistently so. Integration is limited to 4 platforms.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.