§ ICoverFOLIO 001 / 006

01/Evidentiary AI infrastructure

Prove your AI
was governed.

Every AI-assisted decision becomes evidence, sealed against . When a reviewer asks, your record is your evidence.

Live · 47 events / day · representative practice

app.moirai.health/dashboardVIEW 01 / READINESS

Good morning

Sydney Radiology Group

Tier B+

Review-Ready

Verified

14:32:08Annalise CXR · Decision pack sealed

14:31:42Aidoc PE Detection · Pending senior review

14:30:15Harrison CT Brain · Compliance check passed

14:28:51Annalise CXR · Drift alert acknowledged

14:27:09Lunit INSIGHT · Tool registered + verified

87/ 100

Readiness

AI Risk Register

ToolRiskStatusScore

Annalise CXRMediumreadyA

Harrison CT BrainHighreadyA

Aidoc PE DetectionHighreviewB+

§ II Your workflow

02Your workflow

What you do.

Three steps. Twenty minutes to first sealed record.

01

Register your AI tools

Add every AI system your practice deploys. Moirai auto-classifies risk level and maps regulatory requirements.

AI Risk RegisterVIEW 02 / TOOLS

ToolRiskStatusScore

Harrison.ai CT BrainHighreadyA

Annalise.ai CXRMediumreadyA

Aidoc PE DetectionHighreviewB+

3 tools registered2 mapped

02

Log every AI decision

Each time AI assists a clinical decision, log it in under 60 seconds. Moirai seals it with a hash chain and trusted timestamp. The result is an immutable evidence record.

Decision PackDP-2026-0347

DP-2026-0347Sealed

1

AI Identified

Annalise CXR v3.2

2

Human Accountable

Dr S. Chen

3

Context Preserved

CT Chest, PE query

4

Reasoning Documented

Accepted as-is

5

Integrity Verified

SHA-256 chain valid

hash: a3f2b8c1d4e5...f0a1

03

Export sealed proof

One-click audit packs for insurers, regulators, and boards. Cryptographically sealed, independently verifiable, review-ready.

Audit Pack Export2026 / Q1

Compliance Policy Pack24 pages

Risk Register12 pages

Evidence Trail48 pages

Cryptographically sealed

TGA readyRANZCR readyInsurer readyBoard ready

See it in action. No signup required.

See it in the live demo

§ III Your exposure

Your exposure

Every ungoverned AI decision is a liability.

Medico-legal frameworks were not designed for AI-assisted diagnosis. Without a documented record, every decision is an undefended exposure.

The premise

0+ FDA-authorised AI/ML medical devices[4].

The verdict

Zeropurpose-built governance platforms[5].

Until now.

Calculate your exposure

AI-assisted decisions per week100

10100250400500

Estimated annual exposure

$260,000

Model. $1M+ avg AU negligence claim, 0.5% AI governance incident rate. Avant Medical Indemnity data[6].

Sources

[4]FDA Software as a Medical Device (SaMD) registry, October 2024.
[5]Market scan, FDA + TGA registries, October 2024. No purpose-built clinical AI governance platforms identified at time of survey.
[6]Avant Medical Indemnity claims data, 2018 to 2024. Average AU negligence claim $1M+. Per-decision incident rate 0.001% based on published settlements.

§ IV What Moirai contains

03What Moirai contains

Everything you need to prove it.

One platform. Five governance domains. Every artefact an insurer, board, regulator, or auditor would ask for.

AI Risk Register

Central inventory

12 tracked

ToolRiskStatusScore

Harrison.ai CTLowReadyA

Annalise.ai CXRMediumReviewB+

Aidoc PE DetectionHighAttentionC

0

/ 100

RANZCR Ch.9

Evaluated 09:42 AEST · 14 Mar

Evidence Vault

SEALED

Immutable record of decisions and artefacts.

risk-assessment-q1.pdf12 Mar 2026

policy-ai-usage-v3.pdf26 Feb 2026

audit-pack-annual.pdf15 Jan 2026

TGARANZCRInsurer

Monitoring & Alerts

Live

Harrison.ai model update requires re-validationURG2h ago

Q2 policy review deadline in 5 daysREV1d ago

Monthly compliance report generatedINFO3d ago

§ V The pipeline

05The pipeline

What Moirai does.

What happens between AI inference and a sealed audit pack. Five stages, each timestamped, each hashed, each immutable.

Fig. 01Pipeline. Five-stage evidentiary record.

01

AI finding

Model inference

t₀ · raw output

02

Clinician reasoning

Why-Bar captured

t₁ · free text

03

Integrity hash

SHA-256

t₂ · 64 hex

04

Sealed record

Timestamp · signature

t₃ · immutable

05

Audit pack

Exportable dossier

t₄ · PDF · JSON

01

AI finding

Model inference

t₀ · raw output

02

Clinician reasoning

Why-Bar captured

t₁ · free text

03

Integrity hash

SHA-256

t₂ · 64 hex

04

Sealed record

Timestamp · signature

t₃ · immutable

05

Audit pack

Exportable dossier

t₄ · PDF · JSON

§ VI Why this exists

04Why this exists

Not another compliance product.

Built on an evidentiary framework developed through clinical practice, regulatory analysis, and medicolegal research. Currently under peer review at JMIRO.

01

Three years of clinical research.[1]

Built from inside the workflow, not bolted on.

02

One evidence framework.[2]

CAIOS, currently under review at JMIRO.

03

Zero precedents.[3]

First purpose-built clinical AI oversight platform.

First sealed record20min

Clinical research3yr

Evidentiary requirements40+

CAIOS domains covered5/5

Designed for Australian radiology groups deploying TGA-registered AI diagnostic tools. Read the CAIOS standard

Sources

[1]CAIOS development log, 2023 to 2026. Internal authorship records.
[2]JMIRO submission, March 2026. Currently under peer review.
[3]Market scan, FDA + TGA registries, October 2024. No purpose-built clinical AI governance platforms identified.

How the chain holds

Every record, hash-chained to the one before it.

Tampering breaks the chain. Verification is a public function — no account, no auth, no telemetry. Click any record below.

12 most recent recordsVerify a hash →

Make this your last ungoverned decision.

Start building your evidentiary trail today. The first sealed record takes twenty minutes.

Aligned with

Prove your AI was governed.