RadmisAIOS Book a demo

Governance

AI your risk owner can sign off. And your team will still use.

The controls are built into the platform, so nobody has to remember to apply them. This page shows what they are and what they are mapped to.

The five questions

What a risk review asks. And where the answer is kept.

Most AI projects stall because nobody can answer these with a record. The platform keeps the record for each one.

  1. Who owns it?

    One entry per AI system: its owner, purpose, risk rating and review date. A check finds any co-worker or workflow that is not registered, registers it and raises a ticket.

    Kept inThe AI Register

  2. What may it do?

    A co-worker is an account with an access level, on the same access tree as your people. It reaches your data only through two controlled entrances, and only through calls you have registered.

    Kept inAccess levels and connectors

  3. What did it do?

    Every change to audited data is recorded by the database itself in a chain of SHA-256 hashes, so a change made outside the screens is recorded too. Each AI turn is kept with its request, its response, the tools it called and its sources.

    Kept inThe audit trail and conversations

  4. Who checked?

    A workflow stops at its approval step until a person decides, and the decision is recorded with their name. A workflow that mails what an AI wrote cannot be saved unless an approval stands on every path.

    Kept inWorkflow approvals

  5. How do we stop it?

    The kill switch stops the connectors, the models and the classifier from one screen. Each flip is recorded and mailed.

    Kept inThe AI Kill Switch

Why it passes review

Automate what repeats. Reason where it needs judgement.

The less of the work a model does on its own, the less there is to defend.

  • Most of the workIf it repeats, automate it.

    An AI helps build the automation quickly. Once it is built, it runs without one.

  • The edge casesIf it needs judgement, reason.

    Keep the model for the cases your existing systems cannot cover.

  • AlwaysGuardrails around both.

    Access levels, human approval and an audit trail apply to all of it. The compliance report comes out of the same record.

Your data

AI stays outside. Your data stays in.

  • Where you run it

    On your premises or in your cloud, next to the data it works on.

  • Local models first

    Personal data can be handled by a model on your own hardware, so it never leaves your network.

  • Sanitiser

    Before text goes to an outside model, personal details are replaced with labels. They are restored in the answer. The log keeps counts, never values.

  • Data inventory

    Personal data in your tables is found and listed, so you know what you hold and where.

  • Personal data requests

    Find, export, erase or restrict one person's data, with the request on a clock.

  • Retention

    Old rows are removed on schedule, with a destruction log and holds for records that must be kept.

Registers that come with the platform

  • AI Register
  • Test sets and scheduled reviews
  • Providers and models
  • Risk profile and controls
  • Policies and training
  • User consents
  • Settings changes, with a reason
  • Records holds

Coverage

Mapped, not certified.

Thirteen regulations and guidelines, point by point, against the code on 28 October 2026. We show what is not met yet, because your reviewers will ask.

MetIn partNot yet

  • All thirteen191/365
  • APRA CPS 23020/35
  • APRA CPS 23411/14
  • Privacy Act 198817/22
  • OAIC APP guidelines11/12
  • DTA AI policy v2.17/19
  • DTA impact assessment tool20/42
  • DTA assessment guidance6/18
  • NSW AI Assessment Framework7/21
  • NSW AI Operational Policy v1.119/41
  • NSW Circular DCS-2026-022/5
  • NIST AI RMF 1.026/52
  • EU AI Act12/42
  • GDPR33/42

Ask for the full map, point by point

For your reviewers

Five questions to ask any AI vendor. Including us.

Can our auditor verify the audit log without your help?

Yes. The trail is a hash chain written by the database itself. A change made outside the screens is recorded too, and a check names the first broken link.

If an agent is tricked by a prompt, what can it write, delete or read?

Only what its access level opens, through calls you have registered.

When AI builds a screen, is any code generated?

No. The AI describes and the platform builds.

Where does personal data go before it reaches the model?

To a local model first. Before an outside model, through the sanitiser where it is set.

How long before the first governed use is running?

About an hour to install. A first register can be on screen the same day.

Bring your risk owner to the demo. We will show the record.

We walk through the five questions on the live platform and take your reviewers' questions as they come.