The metric

The AI Trust Index.

One 0–100 number for the board — a credit rating for your AI estate. Security 50, Compliance 30, Value 20 — every input traceable to evidence, and benchmarked against your sector peers.

See it on your estate
AI Trust Index
0
● MEDIUM · 65–84
▲ +11 vs last quarter
AI Security weight 50%86 / 100
Measured across Proxy, Probe, Connector and the Recon questionnaire — the whole estate, not just the gateway.
AI Compliance weight 30%70 / 100
APRA, DPDP, ISO 42001 and NIST mapped continuously from live telemetry, exported as PDF + JSON.
AI Value weight 20%40 / 100
FinOps cost attribution and token-budget fences — the CFO’s question, answered beside the CISO’s.

Composite 0–100 per tenant. Bands: 85+ good · 65–84 medium · 45–64 high risk · <45 critical. A decision engine, not a dashboard.

BAND

85+ · Good

Demonstrable control across security, compliance and value. Audit-ready.

BAND

65–84 · Medium

Material gaps in one or more pillars. A prioritised remediation path is generated.

BAND

<45 · Critical

Ungoverned AI in active use. Immediate board-level attention required.

How you compare

A credit rating for your AI.

Benchmarked against anonymised peers — so the board sees not just a number, but where you sit in your sector. Peer benchmarking is always-on, for every tenant.

0sector median · 61100

Above the financial-services median of 61 — and trending up +11 this quarter.

what moves your scorerecommended
Enforce the PII egress guardrail
+6
Close the CPS 234 §32 incident gap
+4
Set token budgets on 3 over-spend teams
+3
Bring 2 shadow-AI tools under policy
+2
Methodology

How the score is computed.

Transparent and auditable by design — no black box. Each pillar is measured from live signals, weighted, and rolled into the composite.

50

AI Security

Proxy enforcement, Probe findings, Connector audit signals and the Recon 330-question assessment — combined into the security pillar.

Largest weight
30

AI Compliance

Continuous control coverage against APRA, DPDP, ISO 42001 and NIST AI RMF, mapped from telemetry.

Evidence-linked
20

AI Value

Cost efficiency and token-budget adherence — value realised per dollar of AI spend.

FinOps pillar
FAQ

Frequently asked questions

What is the AI Trust Index?
A single 0–100, board-level score for your AI posture — one number that rolls up how secure, compliant and cost-controlled every AI in your estate is.
How is the AI Trust Index calculated?
It is composed of three weighted pillars — Security (50%), Compliance (30%) and Value (20%) — and every input is traceable back to evidence.
Is the score benchmarked against peers?
Yes. Your AI Trust Index is benchmarked against sector peers, so the board sees not just your number but where you stand.
What is dual-confidence scoring?
Inputs come from two sources: attested (from assessments and questionnaires) and verified (from live gateway telemetry). Verified evidence carries more weight, so the score reflects what is actually happening, not just what was claimed.
How often does the score update?
Continuously — verified inputs come from live telemetry and the score recomputes as posture changes, so it does not go stale between audits.
Can I show the AI Trust Index to my board or auditor?
Yes. It is built to be board-ready (one number, peer benchmark) and auditor-ready (every input links to evidence, mapped to your frameworks).
What improves the score?
Enabling guardrails, closing compliance gaps and tightening cost controls all raise it — and the platform recommends what to enable next for the biggest lift.
Does the Trust Index map to compliance frameworks?
Yes. The Compliance pillar maps to 26 frameworks across six jurisdictions, so the score reflects APRA, EU AI Act, ISO 42001, NIST AI RMF, DPDP and more.