AI governance

AI governance that lets you say yes to AI

AI is an advantage when it is governed. We build the policy, guardrails and oversight that let your people use it with confidence, built to ISO/IEC 42001.

Led by one of our principals, who holds

  • ISO/IEC 42001 Lead Implementer
  • ISO/IEC 42001 Lead Auditor
Risk management Cyber security AI governance

What we do

AI governance services

For organisations already using AI, or about to, that need to show a board, a customer or a regulator that it is under control.

ISO/IEC 42001 AI management system

A gap assessment against ISO/IEC 42001:2023, then the AI management system itself: policy, roles, risk and impact assessments, and controls across the life of each AI system. Then an internal audit.

AI policy and guardrails

A responsible-use policy your people will read, a register of the AI already in use, rules for approving new tools, and the guardrails that keep sensitive information out of them.

AI risk and impact assessments

Each AI system assessed for what it could get wrong and who it could affect, using the NIST AI Risk Management Framework, and classified against the EU AI Act where that law reaches you.

Oversight and audit

The steering committee, the centre of excellence and the reporting line to the board. Then audits of how AI is used, measured against your own policy.

What you get

AI your board can answer for

When someone asks who approved an AI system, what it does with personal information and who checks it, there is a written answer.

  • An AI policy your people will read
  • A register of the AI in use, and who owns each system
  • A risk and impact assessment for each system
  • A governance structure with named roles and a line to the board
  • An internal audit report against ISO/IEC 42001

Standards and law

Four sets of rules this work answers to

We help you work out which of these reach you, alongside your lawyers. This page is general information, not legal advice.

  • ISO/IEC 42001:2023The international standard for AI management systems. An organisation can be certified against it.
  • NIST AI RMFThe AI Risk Management Framework: voluntary, in four functions. Govern, map, measure and manage.
  • EU AI ActEuropean law that can apply to organisations outside the EU when their AI systems, or the output of them, are used there.
  • Privacy Act 1988A privacy policy must say when a computer program uses personal information to make, or play a substantial and direct part in, decisions that significantly affect people. This applies to organisations covered by the Act from 10 December 2026.

How it runs

From gap assessment to the certification audit

An ISO/IEC 42001 project runs in four stages. We do the first three with you. The fourth belongs to someone else, by design.

  1. Gap assessment

    We compare what you do now with what the standard asks, and rank the gaps by risk.

  2. Build

    We build the AI management system with your people: the AI policy, the roles, the risk and impact assessments, the controls and the records.

  3. Internal audit

    Someone who did not build the system audits it the way a certification body will: one of us, or your own auditor with our support.

  4. Certification audit

    An accredited certification body audits you in two stages and issues the certificate. We stay beside you through both.

Common questions

AI governance questions, answered

What is ISO/IEC 42001 (ISO 42001)?

The first international management system standard for artificial intelligence, published in December 2023. It sets out how an organisation that builds or uses AI should govern it: policy, roles, risk and impact assessment, and controls across the life of each AI system.

We already have ISO/IEC 27001. How much of ISO/IEC 42001 is done?

More than you might expect. The two standards share a clause structure, so the management system you already run carries over. What is new is specific to AI: the AI policy, the impact assessments and the AI controls. We build those onto what you have.

Does the EU AI Act apply to an Australian organisation?

It can. The Act reaches organisations that place AI systems on the EU market, and organisations outside the EU whose AI output is used in the EU. What it asks depends on the risk class of each system, and its obligations apply in stages. We help you form a view on whether you are in scope, and your lawyers confirm it, before you spend anything on compliance.

What does the Privacy Act ask about automated decisions?

From 10 December 2026, a privacy policy must disclose when a computer program uses personal information to make, or play a substantial and direct part in, a decision that could significantly affect someone. If your organisation is covered by the Act and AI makes or shapes decisions like that, your privacy policy needs to say so. Exemptions can apply, so check with your lawyer.

Related

AI governance stands on security

Start with a conversation

Tell us what you’re working on. We’ll tell you whether we can help, and how.

Contact us