For PR & Communications
Trust claims without evidence are PR debt.Evidence without narrative is missed opportunity.
Good Proof gives PR and Communications teams verifiable evidence for AI trust narratives — stakeholders can check, not just trust.
Verifiable trust claims • Crisis-ready evidence • Stakeholder confidence signals • Show-don't-tell verification
Not a certification. Scope-limited verification.
PR & Comms 60-Second Brief
Problem
Public trust claims about AI safety lack verifiable evidence — stakeholders can't distinguish real controls from marketing.
Control
Status Link provides a live, externally-verifiable signal that can back public statements.
Output
Evidence Pack and verification status available for crisis response, press inquiries, and stakeholder briefings.
Boundary
Provides verifiable evidence for trust narratives. Does not create or manage the narrative itself.
Is this a fit?
Best fit
- Organisations making public claims about AI safety, governance, or responsible deployment
- PR teams needing defensible evidence when AI incidents generate press attention
- Comms leads building stakeholder trust narratives backed by verifiable controls
Not fit
- Teams seeking PR strategy, messaging, or media relations services
- Organisations wanting AI safety marketing without operational controls
Done means: every public AI trust claim is backed by a verifiable Status Link.
Trust Narrative Outcomes
Verifiable trust signal
Public claims backed by live Status Link — stakeholders can check, not just trust
Crisis-ready evidence
When incidents hit press, Evidence Packs provide immediate, factual response material
Stakeholder confidence
Board, investor, and regulator briefings grounded in verifiable operational data
Show-don't-tell narrative
Replace trust claims with trust evidence — verification status speaks louder than messaging
How Verification Backs Your Narrative
1
Organisation deploys verification gates on AI decision classes
2
Each gated action produces a verifiable Status Link and Evidence Pack
3
PR can reference live verification status in public statements
4
Stakeholders can independently verify the claim via Status Link
5
Crisis events produce evidence trails for factual response
Failure Semantics
Explicit condition → state → action mapping
| Condition | State | Action |
|---|---|---|
| Verifier timeout | NOT_VERIFIED | Block/Escalate |
| TLS failure | NOT_VERIFIED | Block/Escalate |
| Domain mismatch | NOT_VERIFIED | Block/Escalate |
| Malformed response / signature failure | NOT_VERIFIED | Block/Escalate |
| WITHDRAWN | WITHDRAWN | Block |
| NEEDS_REFRESH | NEEDS_REFRESH | Escalate/Review |
| Out-of-scope / expired | NOT_VERIFIED | Block/Escalate |
What a Stamp Proves
Proves (within scope)
- Verification status at decision-time
- Scope boundaries for gated AI actions
- Authority reference and signer identity
- Evidence chain available for external review
- Fail-closed semantics on verification failure
- Change detection and re-verification triggers
Does not prove
- AI model accuracy, bias, or output quality
- Whether organisational intent is trustworthy
- Brand reputation or public perception
- Comprehensive AI safety across all systems
- Regulatory compliance or certification
For public claims: Status Link = verifiable current state. Evidence Pack = decision-time record stakeholders can audit.
Status Triggers
NEEDS_REFRESH
- AI model version update or retraining
- Policy or governance scope change
- Public commitment or disclosure update
- Key rotation or credential refresh
- Incident opened affecting stated controls
WITHDRAWN
- Confirmed control failure contradicting public claims
- Material incident requiring public correction
- Scope boundary breach in verified actions
- Regulatory finding invalidating stated controls
- Critical vulnerability in verification pipeline
Evidence for Stakeholders
Fields per Gate Decision
evidence-fields.json
stamp_idaction_classdecisionstatusverified_atexpires_atverifier_domainsignerversionrequest_idreason_codefailure_mode
Designed for transparency reports, board packs, and stakeholder briefings.
Fits Into Your Comms Function
Evidence source, not narrative replacement
Corporate communications and media relationsInvestor relations and ESG reportingCrisis management and incident responseRegulatory affairs and public policyBrand safety and reputation monitoring
What PR & Comms Teams Ask
Can we claim 'AI-verified' in press releases?
You can state that specific AI action classes are gated by externally-verifiable Status Links with fail-closed semantics. Claims should be scope-bounded — Good Proof verifies within defined action classes, not across all AI systems.
What happens when a journalist asks 'prove it'?
Status Links are independently verifiable. A journalist can check the verification status of a specific decision class without needing access to internal systems. Evidence Packs provide the detailed audit trail.
How does this help during an AI incident?
Evidence Packs provide factual, timestamped records of what controls were in place at decision-time. This replaces speculation with evidence in crisis communications and prevents the 'we had controls' claim without proof.
Does this help with investor/board confidence?
Yes. Board packs and investor briefings can reference live verification status and operational evidence rather than aspirational governance statements. The data is auditable and exportable.
"Our AI [ACTION CLASS] decisions are externally verified at decision-time with fail-closed semantics."
Scope-bounded claims only. No blanket safety or compliance statements.
"[ACTION CLASS] verification status is externally checkable via Status Link (No Stamp → No Claim)."
30-Day Sprint
One trust narrative backed by live verification, crisis-tested evidence, and stakeholder-ready output.
1
Scope & narrative mapping
Define 3 AI action classes, map to public trust claims, establish evidence requirements
2
Integrate gate
Deploy verification gates, wire Evidence Pack output, establish Status Link endpoints
3
Crisis simulation
Test incident response with Evidence Packs, stakeholder verification flow, press inquiry drill
4
Go-live report
Trust narrative evidence pack, board-ready materials, stakeholder verification guide, rollout plan
Risk Reduced vs Residual
Risk reduced
- Public trust claims without verifiable evidence
- Crisis response delays from missing operational data
- Stakeholder confidence erosion from unsubstantiated safety claims
- Regulatory scrutiny of marketing-without-controls
Residual (out of scope)
- AI model accuracy, bias, or output quality
- Public perception and brand reputation
- Media narrative and editorial framing
- Comprehensive safety across all AI systems
Designed to make trust claims verifiable, not to manage public perception.
Assurance Governance (Optional by Lane)
Mind Chill Guardians
Human finality for high-profile exception cases
Conflict checks and reviewer rotation
Multi-review thresholds for publicly-referenced lanes
Minimal disclosure with auditable traceability
Ready to back your trust narrative with evidence?
Definition of done: every public AI trust claim is verifiable via Status Link.
Every trust claim without evidence is a reputation risk waiting for the next headline.
Scope-limited verification. Not a certification.