Live demo of the Axiom Audit Portal. Data is synthetic but representative of a real assessment.
Recommended Assessment Outcome
BLOCK
Residual Non-Conformity Index
MODERATE
Score: 1.31
(Unwitnessed ÷ Total) × Severity × Staleness
Passing
39
of 103 requirements
Findings
8
open findings
EU AI Act risk tier determines which conformity obligations apply. Select the classification that matches this AI system.
AI systems in Annex III areas that pose significant risks to health, safety, or fundamental rights. Requires third-party or internal conformity assessment.
Obligations
Full conformity assessment required: risk management (Art. 9), data governance (Art. 10), logging (Art. 12), transparency (Art. 13), human oversight (Art. 14), accuracy & robustness (Art. 15), post-market monitoring (Art. 72).
Annex III Categories
Loading risk register...
Loading checklist...
AI Governance
| Procedure | Family | Title | Severity | Verdict | Last Witnessed | Verify |
|---|---|---|---|---|---|---|
| AI-GRD.1 | AI | Guardrail Enforcement — Required Safety Filters Active Required Guardrails: 1Active Guardrails: 0 Open 7d | CAT II | FAIL | 6/17/2026 | Verify → |
| AI-HITL.3 | AI | Overseer Identity Capture - Reviewer Identified Per Decision Identity Capture Required: 0.5Identity Captured: 1 Open 76d (overdue) | — | FAIL | 4/8/2026 | Verify → |
| AI-EXPL.1 | AI | Explanation Generation Explanation Required: 0.5Explanation Generated: 1 Open 74d (overdue) | — | FAIL | 4/10/2026 | Verify → |
| AI-GRD.2 | AI | Content Safety Filter — Output Classification Passed Clean Expected: 0.5Refusal Detected: 1 Open 73d (overdue) | — | FAIL | 4/12/2026 | Verify → |
| AI-FAIR.2 | AI | Fairness Calibration Min Calibration Score: 0.5Actual Calibration: 1 Open 72d (overdue) | — | FAIL | 4/13/2026 | Verify → |
| AI-EXPL.2 | AI | Confidence Scoring Min Confidence Score: 0.5Actual Confidence: 1 Open 71d (overdue) | — | FAIL | 4/13/2026 | Verify → |
| AI-GOV.3 | AI | Approved Model Registry Registry Required: 0.5Registry Active: 1 Open 70d (overdue) | — | FAIL | 4/14/2026 | Verify → |
| AI-TOOL.1 | AI | Tool Call Witnessing Tool Called: 1Latency (ms): 45Result: 1 Open 7d | — | FAIL | 6/17/2026 | Verify → |
| AI-DATA.3 | AI | Training Data Statistics | — | UNKNOWN | — | — |
| AI-DATA.4 | AI | Training Data PII Lifecycle | — | UNKNOWN | — | — |
| AI-GOV.6 | AI | AI Risk Management Scope Definition | — | UNKNOWN | — | — |
| AI-GOV.7 | AI | AI Governance Resource Allocation | — | UNKNOWN | — | — |
| AI-RISK.1 | AI | AI Risk Identification and Categorization | — | UNKNOWN | — | — |
| AI-IR.1 | AI | AI Incident Response Capability | — | UNKNOWN | — | — |
| AI-IMPACT.1 | AI | AI Societal Impact Assessment | — | UNKNOWN | — | — |
| AI-RAG.2 | AI | Context Relevance Scoring | — | UNKNOWN | — | — |
| AI-MDL.5 | AI | Weight File Integrity | — | UNKNOWN | — | — |
| AI-MDL.6 | AI | Adapter Stack Attestation | — | UNKNOWN | — | — |
| AI-MDL.7 | AI | Quantization Attestation | — | UNKNOWN | — | — |
| AI-SKILL.2 | AI | Memory Context Binding | — | UNKNOWN | — | — |
| AI-SKILL.3 | AI | Reward Model Binding | — | UNKNOWN | — | — |
| AI-HW.1 | AI | Hardware Runtime Attestation | — | UNKNOWN | — | — |
| AI-HW.3 | AI | TPM Platform Attestation | — | UNKNOWN | — | — |
| AI-TRUST.2 | AI | Trust Credential Presentation | — | UNKNOWN | — | — |
| AI-CHAIN.2 | AI | Chain Trust Degradation | — | UNKNOWN | — | — |
| AI-FAIR.3 | AI | Bias Audit Witnessing | — | UNKNOWN | — | — |
| AI-ENV.2 | AI | Dependency Manifest Attestation | — | UNKNOWN | — | — |
| AI-BASE.1 | AI | Agent Behavioral Baseline | — | UNKNOWN | — | — |
| AI-LIC.1 | AI | License Provenance | — | UNKNOWN | — | — |
| AI-REDTEAM.1 | AI | Adversarial Test Campaign | — | UNKNOWN | — | — |
| AI-MULTI.1 | AI | Multi-Agent Delegation | — | UNKNOWN | — | — |
| AI-INCIDENT.1 | AI | Incident Reporting | — | UNKNOWN | — | — |
| AI-PERF.1 | AI | Performance Metrics | — | UNKNOWN | — | — |
| AI-ROBUST.1 | AI | Robustness Testing | — | UNKNOWN | — | — |
| AI-CYBER.1 | AI | Cybersecurity Attestation | — | UNKNOWN | — | — |
| AI-AUTO.1 | AI | Automated Decision Notification | — | UNKNOWN | — | — |
| AI-DUALUSE.1 | AI | Dual-Use Model Classification | — | UNKNOWN | — | — |
| AI-SUPPLY.1 | AI | Supply Chain Risk | — | UNKNOWN | — | — |
| AI-PMM.1 | AI | Post-Market Monitoring | — | UNKNOWN | — | — |
| AI-AUTO.2 | AI | Autonomous Generation Depth | — | UNKNOWN | — | — |
| AI-AUDIT.2 | AI | External Timestamp Attestation | — | UNKNOWN | — | — |
| AI-METAGOV.1 | AI | Governance Infrastructure Attestation | — | UNKNOWN | — | — |
| AI-METAGOV.2 | AI | Governance Layer Registration | — | UNKNOWN | — | — |
| AI-METAGOV.3 | AI | Policy Downgrade Detection | — | UNKNOWN | — | — |
| AI-METAGOV.4 | AI | Circular Dependency Check | — | UNKNOWN | — | — |
| AI-METAGOV.5 | AI | Governance Authorization | — | UNKNOWN | — | — |
| AI-METAGOV.6 | AI | Emergency Override Attestation | — | UNKNOWN | — | — |
| AI-METAGOV.7 | AI | Governance Sync Verification | — | UNKNOWN | — | — |
| AI-METAGOV.8 | AI | Attestation Purity Verification | — | UNKNOWN | — | — |
| AI-ENG.1 | AI | Design Generation Provenance | — | UNKNOWN | — | — |
| AI-ENG.2 | AI | Simulation Validation | — | UNKNOWN | — | — |
| AI-ENG.3 | AI | Safety-Critical Review Gate | — | UNKNOWN | — | — |
| AI-ENG.4 | AI | Material Specification Compliance | — | UNKNOWN | — | — |
| AI-ENG.5 | AI | Design Revision Chain | — | UNKNOWN | — | — |
| AI-ENG.6 | AI | Fabrication Release Authorization | — | UNKNOWN | — | — |
| AI-FIN.1 | AI | Agent Transaction Witnessing | — | UNKNOWN | — | — |
| AI-LCM.1 | AI | Agent Lifecycle Witnessing | — | UNKNOWN | — | — |
| AI-TOOL.2 | AI | Tool Permission Attestation | — | UNKNOWN | — | — |
| AI-JUR.1 | AI | Cross-Border Inference Routing | — | UNKNOWN | — | — |
| AI-COST.1 | AI | Resource Consumption Witnessing | — | UNKNOWN | — | — |
| AI-DEL.1 | AI | Delegation Tree Witnessing | — | UNKNOWN | — | — |
| AI-CAP.1 | AI | Capability Attestation | — | UNKNOWN | — | — |
| AI-AUTO.3 | AI | Autonomy Level Transition | — | UNKNOWN | — | — |
| AI-CLR.2 | AI | Clearing Fidelity Attestation | — | UNKNOWN | — | — |
| AI-MDL.2 | AI | Model Version Tracking — Version Identifier Recorded Version Required: 1Version Recorded: 1Context: 1 | — | PASS | 4/8/2026 | Verify → |
| AI-LOG.1 | AI | Log Retention Compliance - Minimum 180-Day Retention Verified Min Retention (days): 1Actual Retention (days): 1Context: 1 | — | PASS | 4/9/2026 | Verify → |
| AI-MDL.4 | AI | Feedback Loop Control - Training Data Isolation From Biased Outputs Threshold: 1Measured: 1Delta: 1 | — | PASS | 4/11/2026 | Verify → |
| AI-GRD.3 | AI | PII Redaction PII Redaction Required: 1Redaction Active: 1Context: 1 | — | PASS | 4/11/2026 | Verify → |
| AI-FAIR.1 | AI | Bias Disparity Measurement Max Disparity Ratio: 1Measured Disparity: 1Context: 1 | — | PASS | 4/13/2026 | Verify → |
| AI-DATA.1 | AI | Training Data Provenance Documentation Required: 1Documented: 1Context: 1 | — | PASS | 4/14/2026 | Verify → |
| AI-DATA.2 | AI | Training Data License Compliance License Check Required: 1Verified: 1Context: 1 | — | PASS | 4/12/2026 | Verify → |
| AI-HITL.2 | AI | Human Override Event Tracking Override Logging Required: 1Logging Active: 1Context: 1 | — | PASS | 4/8/2026 | Verify → |
| AI-GOV.1 | AI | AI Acceptable Use Policy Policy Required: 1Policy Active: 1Context: 1 | — | PASS | 4/14/2026 | Verify → |
| AI-GOV.2 | AI | Employee AI Training Training Required: 1Training Complete: 1Context: 1 | — | PASS | 4/9/2026 | Verify → |
| AI-GOV.4 | AI | Shadow AI Incident Response IR Procedure Required: 1Procedure Documented: 1Context: 1 | — | PASS | 4/13/2026 | Verify → |
| AI-GOV.5 | AI | Third-Party AI Vendor Assessment Vendor Assessment Required: 1Assessments Complete: 1Context: 1 | — | PASS | 4/12/2026 | Verify → |
| AI-ID.1 | AI | Agent Identity Assertion Identity Required: 1Identity Present: 1Context: 1 | — | PASS | 6/17/2026 | Verify → |
| AI-ACC.1 | AI | Agent Access Control Witnessing Access Attempted: 1Within Scope: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-REV.1 | AI | Anchor Revocation Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-SEC.1 | AI | Adversarial Threat Detection Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-SEC.2 | AI | Input Validation and Sanitization Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-RAG.1 | AI | Context Retrieval Provenance Threshold: 0.94Measured: 6Delta: 2 | — | PASS | 6/17/2026 | Verify → |
| AI-SKILL.1 | AI | Skill Manifest Attestation Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-TRUST.1 | AI | Trust Verification Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-CHAIN.1 | AI | Multi-Agent Chain Handoff Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-SAFE.1 | AI | Safe State Transition Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-CHR.1 | AI | Agent Charter Registration Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-VIO.1 | AI | Policy Violation Record Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-ENV.1 | AI | Runtime Environment Attestation Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-MARK.1 | AI | Content Provenance Marking Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-SBOM.1 | AI | AI Bill of Materials Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-CONSENT.1 | AI | Data Subject Consent Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-DRIFT.1 | AI | Model Drift Detection Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-AUDIT.1 | AI | Audit Log Integrity Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-TRANS.1 | AI | Transparency Disclosure Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-WATERMARK.1 | AI | Watermark Verification Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-DPIA.1 | AI | Data Protection Impact Assessment Threshold: 1Measured: 1 | — | PASS | 6/10/2026 | Verify → |
| AI-INF.1 | AI | Inference Provenance — Prompt/Response Hash Capture Hash Required: 0.99Hashes Present: 43Context: 2 | — | PASS | 6/17/2026 | Verify → |
| AI-INF.2 | AI | Inference Latency — Response Time Within Threshold Latency Limit (ms): 1Actual Latency (ms): 1 | — | PASS | 4/20/2026 | Verify → |
| AI-MDL.1 | AI | Model Weight Integrity — Deployed Hash Matches Approved Hash Required: 1Hash Present: 1 | — | PASS | 4/20/2026 | Verify → |
| AI-INF.3 | AI | Inference Volume - Hourly Rate Governance Volume Limit: 1Actual Volume: 1Context: 1 | — | PASS | 4/10/2026 | Verify → |
| AI-MDL.3 | AI | Model Drift Detection Drift Tolerance (%): 1Measured Drift (%): 1Context: 1 | — | PASS | 4/9/2026 | Verify → |
| AI-HITL.1 | AI | Human Review Completion Review Required: 1Review Completed: 1Context: 1 | — | PASS | 4/12/2026 | Verify → |
Loading findings...
Loading requests...
Loading interviews...
Same seed + same population = same sample. Reproducible for peer review.
Open
9
Overdue
7
Closed
56
| Procedure | Severity | Status | Days Open | Milestone |
|---|---|---|---|---|
| AI-GRD.1 | CAT II | OPEN | 7d | 8/16/2026 |
| AI-TOOL.1 | — | OPEN | 7d | 8/16/2026 |
| AI-GOV.3 | — | OPEN | 70d | 6/13/2026 (overdue) |
| AI-EXPL.2 | — |
Models Witnessed
8
Total Inferences
178
Pass Rate
69%
Complete inventory of AI agents witnessed in this system. EU AI Act Articles 13 (Transparency) and 14 (Human Oversight) require disclosure of agent identity, capabilities, and operational boundaries.
Agents
17
Models
8
Tools Invoked
15
Violations
3
| Agent | Models | Inferences | Pass Rate | Clearing | Violations | Active Period |
|---|---|---|---|---|---|---|
▶ — | claude-haiku-4-5 | 35 | 51% | L0 Analytics | 0 | Mar 31 — Apr 20 |
▶ — | gpt-4o | 35 | 74% | L0 Analytics | 0 | Apr 2 — Jun 10 |
▶ — | gpt-4o-mini | 28 | 57% | L0 Analytics | 0 | Mar 31 — Apr 20 |
▶ — | claude-sonnet-4-6 | 24 | 58% | L0 Analytics | 0 | Mar 31 — Apr 20 |
▶ — | demo-compliance-model-v1 | 12 | 83% | L0 Analytics | 0 | Apr 8 — Apr 15 |
▶ — | claude-3.5-sonnet | 7 | 100% | L0 Analytics | 0 | Jun 10 — Jun 10 |
▶ doc-processor | gpt-4o-mini | 7 | 100% | L2 Sensitive | 0 | Jun 17 — Jun 17 |
▶ support-bot | claude-sonnet-4-6 | 5 | 80% | L1 Standard | 1 | Jun 17 — Jun 17 |
▶ fraud-orchestrator | gpt-4o | 4 | 100% | L1 Standard | 0 | Jun 17 — Jun 17 |
▶ compliance-agent | a2a69af70d1b9be7 | 4 | 100% | L3 Classified | 0 | Jun 17 — Jun 17 |
▶ rag-assistant | gpt-4o | 3 | 100% | L1 Standard | 0 | Jun 17 — Jun 17 |
▶ moderator | claude-haiku-4-5 | 3 | 100% | L1 Standard | 0 | Jun 17 — Jun 17 |
▶ research-coordinator | claude-opus-4-6 | 3 | 100% | L1 Standard | 0 | Jun 17 — Jun 17 |
▶ data-analyst | gpt-4o | 3 | 33% | L2 Sensitive | 2 | Jun 17 — Jun 17 |
▶ — | claude-opus-4-6 | 2 | 0% | L0 Analytics | 0 | Apr 19 — Apr 19 |
▶ web-searcher | claude-sonnet-4-6 | 2 | 100% | L1 Standard | 0 | Jun 17 — Jun 17 |
▶ risk-assessor | gpt-4o | 1 | 100% | L1 Standard | 0 | Jun 17 — Jun 17 |
Loading monitoring data...
Multi-agent decision lineage with cryptographic proof at every handoff. Hover nodes for human-readable context. Click to drill down.
Chains
8
Total Nodes
35
Agents
10
Violations
2
Daily Merkle rollups anchor all witness records into a tamper-evident tree. Click any date to inspect its anchors, then click a fingerprint to generate an inclusion proof.
Days Sealed
2
Anchors Sealed
55
Latest Root
0b239145...abab
| Date | Anchor Count | Merkle Root | |
|---|---|---|---|
Get your own audit portal with cryptographic evidence, Annex IV checklist, and assessment tools. Free tier includes 90-day anchor retention and up to 1,000 inferences.
Every finding in this report is backed by a cryptographic SWT3 Witness Anchor. Anchors can be independently verified at any time using the public verification endpoint.
Axiom is a neutral evidence platform. It does not grant certifications, issue conformity assessments, or replace the professional judgment of a qualified auditor, Notified Body, or Authorizing Official. Scores, gate decisions, and recommended outcomes are computed deterministically from machine-gathered evidence and are provided for informational purposes only. The final authority on conformity or compliance rests with the designated assessment body and their professional standards.
SWT3 Protocol, Witness Anchor format, UCT Registry, and AI Witness clearing methodology are Patent Pending (USPTO Provisional Patent Applications filed 2026). © 2026 Tenable Nova LLC. All rights reserved.
| OPEN |
| 71d |
| 6/12/2026 (overdue) |
| AI-FAIR.2 | — | OPEN | 72d | 6/12/2026 (overdue) |
| AI-GRD.2 | — | OPEN | 73d | 6/11/2026 (overdue) |
| AI-EXPL.1 | — | OPEN | 74d | 6/9/2026 (overdue) |
| AI-4.2 | — | OPEN | 75d | 6/9/2026 (overdue) |
| AI-HITL.3 | — | OPEN | 76d | 6/7/2026 (overdue) |