Subprocessor benchmark report appendix
This appendix makes the pilot benchmark auditable. It shows the point-by-point rubric breakdown for all 8 reviewed pages, the score distribution, and the most common public-page gaps from the sample.
This appendix scores public operational clarity, not legal compliance. Private trust-center content, contract terms, or customer-only notices were not used.
11.6 / 20
The pilot sample clustered in the thin-to-adequate middle, with no strong or risky outliers.
5 thin, 3 adequate
62.5% of the sample stayed in the thin band and 37.5% reached adequate.
Rights path and history
All 8 pages lacked a clearly visible objection path and a public archive or change log.
Score distribution
Band definitions follow the public methodology: 17-20 strong, 13-16 adequate, 9-12 thin, and 0-8 risky.
Most common public-page gaps
Criterion-by-criterion scoring appendix
| Company | Score | Public | Updated | Vendors | Purpose | Scope | Region | Notice | Rights | Action | History | Contact visible |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OpenAI | 12 / 20 | 2 | 2 | 2 | 2 | 2 | 2 | 0 | 0 | 0 | 0 | No |
| Superhuman | 11 / 20 | 2 | 0 | 2 | 2 | 0 | 1 | 2 | 0 | 2 | 0 | Yes |
| Inkeep | 10 / 20 | 2 | 2 | 2 | 2 | 1 | 1 | 0 | 0 | 0 | 0 | No |
| Pylon | 9 / 20 | 2 | 2 | 2 | 2 | 0 | 1 | 0 | 0 | 0 | 0 | No |
| WipRadar | 14 / 20 | 2 | 2 | 2 | 2 | 2 | 2 | 0 | 0 | 2 | 0 | Yes |
| Foundable | 14 / 20 | 2 | 2 | 2 | 2 | 2 | 2 | 1 | 0 | 1 | 0 | No |
| Cotool | 9 / 20 | 2 | 0 | 2 | 2 | 0 | 1 | 0 | 0 | 2 | 0 | Yes |
| AgentLattice | 14 / 20 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 0 | 0 | 0 | No |
`2` means clearly visible, `1` means partial or vague, and `0` means absent or too unclear to rely on from the public page. Contact visibility is tracked separately and does not change the 20-point score.
Correction path
If a reviewed page has changed since 2026-05-27 or if this appendix missed a public field that was visible on the review date, send the page URL and the factual correction to hello@noticekit.tech. NoticeKit can update future benchmark batches or append a correction note without changing the public scoring rules retroactively.