Global Enterprise Feature

Active Shield

Four layers of proactive AI governance. Pre-flight compliance verification, SMCR decision rights mapping, Shadow AI discovery, and Annex IV auto-generation — all integrated into your audit trail.

Get Active Shield →Request a walkthrough

Pre-Flight Compliance Checks

Before any model goes to production, Audital runs six automated governance checks against your audit trail. Each check produces a PASS, WARN, or FAIL with the specific remediation required.

A model that scores 100% receives a cryptographically signed Pre-Flight Certificate — a timestamped document that proves governance was completed before deployment. Regulators can verify it independently.

Six Pre-Flight Checks

✓
SMF Holder Assigned
SMCR function holder linked to this model
✓
Approval on Record
APPROVAL_GRANTED event in the audit chain
✓
Risk Framework Mapped
At least one compliance framework applied
✓
Change Ticket Linked
Jira or ServiceNow ticket reference on record
✓
Version History Audited
Previous versions present in audit trail
✓
Incident Clearance
No unresolved incidents before deployment

SMCR Decision Rights Map

Map every material AI decision to a named SMF holder. Deployment, approval, risk sign-off, and incident response — each linked to the specific person accountable under SMCR.

The accountability map is visible on every model detail page and feeds directly into Pre-Flight checks. When the FCA asks who was responsible for a credit scoring decision on a specific date, you have the answer in three clicks.

Decision Types Tracked

—Deployment — SMF holder who approved production release
—Approval — Person who granted model sign-off
—Incident Response — Named accountable for incident remediation
—Risk Sign-Off — Person who signed off the risk assessment
—Monitoring — Ongoing monitoring responsibility holder

Shadow AI Discovery

As AI usage spreads across engineering teams, undeclared models and libraries appear outside any governance framework. Audital scans connected repositories for AI library imports and surfaces them as detections.

Every detection is tracked — OPEN, ACKNOWLEDGED, RESOLVED — and feeds into RegRadar. Shadow AI discovered in connected repos is automatically linked to your compliance risk exposure.

Libraries Detected

▸OpenAI SDK — GPT-4, DALL-E, Whisper
▸Anthropic SDK — Claude models
▸LangChain — Agent and chain frameworks
▸HuggingFace Transformers
▸PyTorch / TensorFlow
▸AWS Bedrock / SageMaker Runtime
▸Azure AI / Google Vertex AI

Annex IV Auto-Generation

EU AI Act Article 11 requires Annex IV technical documentation for all high-risk AI systems. Audital generates this documentation automatically from your audit trail — system description, risk classification, training data governance, human oversight, and change log.

Each report is hashed and timestamped. The SHA-256 fingerprint means you can prove to a regulator that the documentation was generated at a specific point in time from your actual audit data, not assembled retrospectively.

Annex IV Sections Generated

1. General description of the AI system
2. Risk classification and business purpose
3. Accuracy and performance metrics
4. Training data and data governance
5. Human oversight measures
6. Significant changes log
7. Post-market monitoring plan

Active Shield · Insurance

Governance documentation as an underwriting signal

We are in early discussions with Lloyd's market underwriters about recognising Active Shield certification outputs — Pre-Flight Certificates and Annex IV reports — as evidence in AI liability underwriting assessments. Firms with documented, cryptographically verified governance trails present a materially different risk profile to those without.

Global Enterprise

Active Shield is included in Global Enterprise contracts

£12,500/month · annual contract · dedicated infrastructure · founding team access · 99.9% SLA

Apply for Global Enterprise →