The Platform
Every AI system. Every control.
One platform.
Eight integrated modules that take your AI systems from inventory to regulator-ready assurance.
See every AI system. Own every risk.
A single, structured inventory of every AI system, model, and agent across your organisation.
- Register AI systems with full metadata — name, purpose, data inputs, risk classification, and owner
- Track AI agents and automated pipelines, not just models
- Maintain full lifecycle status: In Review, Approved, Deployed, Deprecated
- Link systems to business units, regulatory obligations, and data assets
- Full audit trail of every registry change
Governance that proves itself.
The Evidence Hub transforms governance documentation into verifiable, versioned, integrity-assured proof.
- Upload and version all governance artefacts: bias reports, security assessments, vendor agreements, and ethics approvals
- SHA-256 integrity hashing on every uploaded document — proving it has not been altered since upload
- Policy linkage: connect evidence items directly to specific controls and regulatory requirements
- Confidence Index: real-time view of evidence completeness per AI system
- Role-based access controls on all evidence artefacts
This is what separates Sentrify from every AI governance tool on the market. We don’t just collect documents. We prove they’re authentic.
Evaluation at the speed of AI.
Replace manual review cycles with automated, multi-dimensional evaluation across every AI system.
- Rule-based evaluation: deterministic checks across configurable rule libraries
- LLM-assisted evaluation: AI analysis of unstructured governance documentation — separate from TrustScore™ calculation
- Framework-aligned evaluation packs: pre-built test suites for ISO 42001, APRA CPS 234/230, TEQSA, and more
- Dimensions evaluated — Security, Fairness & Bias, Transparency, Data Governance, Regulatory Alignment
- Automated test plan generation based on the AI system's risk profile
- Full evaluation history with timestamped results and evidence linkage
One number. Full accountability.
TrustScore™ is the universal metric for AI trust — deterministic, weighted, and auditable.
- 0–100 score for every evaluated AI system
- Dimension-weighted scoring — weight dimensions by your organisation's risk priorities
- Risk-tier classification: Approved / Conditional / Restricted / Blocked
- Full score breakdown by dimension and control
- Score history and trend tracking across evaluation runs
- Export-ready for board reporting, regulator submissions, and assurance certificates
TrustScore™ is deterministic — the same inputs always produce the same score. No black box. No subjectivity. No surprises.
Audit-ready assurance on demand.
Generate regulator-ready assurance reports with a complete evidence chain — every decision, every artefact, every timestamp.
- Audit-ready assurance reports with full evidence chain and traceability metadata
- Integrity-hashed report artefacts — tamper-evident and immutable after generation
- Export for board reporting, regulator submissions, and assurance certificates
- Timestamped evaluation history included in every report
- Immutable, tamper-evident report record stored in Sentrify for regulatory review
Assist committee decisions with workflows.
Structured, auditable approval pipelines replace ad-hoc governance processes.
- Multi-stage approval workflows with configurable steps
- Ethics review pipeline with mandatory sign-off steps
- Role-based decisioning — admins, managers, and auditors each have defined workflow authorities
- Automated notifications and escalation on stalled approvals
- Full workflow audit trail: every decision, comment, and approval logged and timestamped
Humans endorse and approve — Sentrify provides the structured workflow, never the decision.
Every standard, versioned and immutable.
A domain-neutral evaluation engine built on versioned, integrity-hashed assurance packs — one platform, every standard.
- Domain-neutral evaluation engine separates core logic from framework-specific content
- Versioned assurance packs map controls to standards: ISO 42001 Annex A, APRA CPS 234/230, ASIC RG 273, Privacy Act ADM obligations, TEQSA, and NHMRC research ethics
- Published framework versions are immutable — integrity-hashed and locked after publication
- Every evaluation references the exact pack version it ran against, ensuring reproducible results
- Custom control pack builder for organisation-specific requirements (extends any base pack)
- Sector-specific packs for financial services, healthcare, government, and higher education
10+ frameworks · 6+ industries · AWS ap-southeast-2 data residency
Build the rules that govern your AI.
Author, test, and version custom evaluation rules without writing code — then deploy them to your live evaluation packs.
- No-code rule authoring interface for custom evaluation logic
- Test rules against existing AI system records before deploying
- Rule versioning and rollback across pack versions
- Share rules across teams or keep them organisation-private
- API extensibility for power users integrating with external governance tooling
Available to early-access customers — contact us to join the beta.
Ready to trust your AI?
See all 8 modules in action. Walk through the platform with our team and get your first AI system evaluated.