AIGR · Institutional AI governance

Ratings

EvidenceReviewDecision

AI Governance Ratings

Artificial Intelligence Governance Ratings built from governed evidence.

AIGR AI Governance Ratings express the governance maturity and evidence confidence of a defined AI system within a fixed scope, jurisdiction, evidence period and methodology version. The published Artificial Intelligence Governance Rating remains separate from the underlying analytical score so the executive signal stays interpretable without hiding the evidence record behind it.

Rating scaleAIGR-100 → 40
Not ratedAIGR-NR
Evidence standard4 tests
Decision controlCritical gates

The score supports analysis. The rating communicates the governed outcome.

Separating the analytical measure from the published designation allows reviewers to work with granular evidence and domain performance while boards, buyers and risk teams receive a stable governance signal.

Analytical layer

AIGR Governance Score

A controlled 0–100 analytical measure used for diagnostics, reviewer analysis, remediation planning and trend monitoring. It is not the published rating.

Rating layer

AIGR designation

A categorical designation from AIGR-100 to AIGR-40, or AIGR-NR when scope, evidence or review conditions do not support an opinion.

Seven designations, plus Not Rated.

The scale is categorical. AIGR-100 is not 100%, a probability or a credit score. A designation is read together with its scope, evidence period, open conditions and methodology version.

Designation Meaning Executive interpretation
AIGR-100 Exemplary governance Leading maturity and evidence confidence
AIGR-90 Advanced governance High confidence with limited material gaps
AIGR-80 Established governance Governance operating with defined improvement areas
AIGR-70 Developing governance Material improvement required
AIGR-60 Emerging governance Substantial remediation required
AIGR-50 Limited governance High governance concern
AIGR-40 Critical governance concern Urgent review and remediation
AIGR-NR Not Rated Evidence, scope or review conditions are insufficient

A severe control failure cannot be averaged away.

The analytical score may show broad maturity, but an open critical gate can still hold the governed outcome. Gate logic is reserved for failures material enough to affect deployment, accountability, rights, safety, security, privacy or the integrity of the evidence itself.

Illustrative decision stateCONDITIONAL

AIGR Governance Score
72 / 100
Critical gates
1 open
Evidence complete
86%

Open gate: required evidence is not available for reviewer validation.

A rating only means something when the boundary is explicit.

Parameter Fixed for the assessment Why it matters
Entity Legal entity and accountable business unit Identifies governance ownership
System Named AI system, version and intended use Prevents the opinion transferring to unrelated systems
Lifecycle Development, validation, deployment or production state Evidence expectations change by lifecycle stage
Jurisdiction Applicable operating and regulatory context Obligations differ across markets
Evidence period Review window and effective date Keeps the opinion time-bound and reproducible

Evidence has to prove the control is operating—not merely described.

01

Relevance

The artifact addresses the requirement actually being reviewed.

02

Currency

The artifact reflects the system and period in scope and has not been superseded.

03

Traceability

The artifact has an accountable owner, source, date and version.

04

Sufficiency

The evidence demonstrates operating effectiveness rather than intent alone.

Validated

Evidence meets the requirement.

In review

Evidence exists; reviewer determination remains open.

Not applicable

Excluded only with a recorded scope rationale.

Blocker

A material failure that can independently hold the outcome.

The rating is designed to remain governed after publication.

01

Affirm

Current evidence continues to support the existing designation.

02

Upgrade / downgrade

Material improvement or deterioration changes the governance opinion.

03

Conditional

A material condition or critical gate limits the decision state.

04

Suspend

An incident, material change or evidence-integrity concern requires review.

05

Withdraw

The scope no longer applies or the opinion can no longer be supported.

06

Not Rated

AIGR-NR applies when the evidence or conditions do not support a rating.

Published framework · 2026

Artificial Intelligence Governance Ratings framework.

Read the complete ratings framework covering the category thesis, architecture, governance lifecycle and rating scale. Proprietary formulas, weights, thresholds and calibration logic remain controlled.

Download framework PDF

Important boundary

AIGR is a governance rating, not a certification.

An AIGR rating does not constitute regulatory approval, legal advice, an audit opinion, a compliance attestation, a credit rating, investment advice, or a guarantee of safety or performance. Read the Ratings & Methodology Disclaimer.

Enterprise assessment

Establish governance readiness for a defined AI system.

Request an Assessment