Contributors

Daniel Urruela, Senior Product Designer. Walked the full guide on desktop and mobile, including the readiness check and the state tracker, and returned a prioritized design pass: reserve Nolte red for actions and keep scores and data categories in black, so the real buttons read; fill the home page card grids so no cell renders empty; add an on-this-page rail and previous and next links to the long parts; give the readiness results a clear how-did-I-do, what-to-fix-first, what-next hierarchy; make the pillar bars and state tiles carry a text label, not color alone; and correct the home copy that still described a ten-question tool that stored nothing, when the check now runs twenty-five questions and logs an anonymous score.

Kaio Pedreira, Senior Product Engineer. Walked the nine controls against the exhibit references and live estates and confirmed the artifact claims and evidence map hold, then sharpened the build: corrected the cost estimate to three to five engineer-weeks; named the approval-gated prompt-promotion pipeline as a legitimate artifact when prompts must stay in a SaaS (Control 2); added the timer-based config hot-reload exposure (Control 3); added the approval and deny asymmetry, found inverted in the wild, to Control 5; and allowed a gated pre-release evaluation job so per-commit model spend is not forced (Control 6). Also settled the scoring-scale wording across the methodology and the readiness check, and shipped the accompanying code fixes.

James Addison Moore, Corecreek. Shared an interactive NAIC readiness prototype and the thinking behind it, which shaped the evolution of the readiness check into a guided, step-by-step assessment: per-pillar scoring, concrete gaps mapped to the controls that close them, and an optional tailored memo.

Dulce Hernández Cruz, Senior Product Engineer. Flagged the redundant "Exhibit B" before "checklist" references; proposed a model retirement_date on the register (Control 1) with a CI rule that fails as a vendor deprecation date nears, so a retired model forces a planned, tested migration instead of a silent swap; and clarified the Control 4 replay assertion to compare the decision and customer-facing fields, not the raw response or confidence, which vary between runs.