EVEglyphDesign · Public rubric
Enterprise is not a pricing tier. It is a substrate-accountability commitment. When a vendor sells the enterprise tier, the vendor accepts accountability for the whole stack it delivers — including the foundation models it did not build.
EgD-GCF-001 · Reference model
The buyer-side instrument. Eight sovereignty control domains, a jurisdictional register anchored to the EU AI Act, EU Data Act, GDPR, DORA and the major national regimes, four veto conditions, a 16-point scoring sheet, contract language hooks, and ten patterns to flag in an incumbent vendor's AI proposal.
EgD-GCF-002 · Positioning note
BCG, Gartner, Bain, McKinsey and NIST publish the frameworks every board has already seen. Five are adoption frameworks; one is a governance framework. None scores the counterparty. One page on the gap, and how EgD-GCF-001 pairs with the NIST AI RMF rather than competing with it.
Practitioner kit · EgD-GCF-003 to 005
The kit for advisers running the assessment for their own clients: a forty-one question vendor questionnaire, a fillable scoring workbook, a worked example against a named vendor, a client-facing one-pager, and a licence permitting use in paid engagements, unmodified, with attribution.
Sign up as a practitioner — free, no assessment, and it puts you on the public register with your LinkedIn, X and YouTube handles so the work can be shared across all of them in one pass.
Template
The versioned roster every enterprise-tier AI platform should publish, so customers know which model substrates sit behind the surface they bought.
Template
The periodic disclosure that turns failure handling into something auditable rather than conversational — named patterns, correction status, remediation.
Briefing
Who signed, who did not, and how the letter interacts with this rubric.
An AI platform vendor claiming enterprise tier should publish its foundation-model roster, publish its alignment overlay, publish failure-mode disclosures at real cadence, provide substrate-selection controls, accept liability commensurate with the tier claim, publish an operator-reachable escalation path, and fund a red team on substrate honesty. The full statement is in the repository README.
The rubric runs at product tier, not at persons. It does not adjudicate individual founders, executives, or employees at any vendor.