Talk to a security expertFree technical auditFree audit

AI governancenoun
Knowing every AI system you run, who owns it, what it can touch, and which tests back the decision to run it.
Review AI governanceSee an example register

Example register

One list everyone can read.

Filter it the way a reviewer would. Each row has an owner, what the system can do and when it was last tested.

Showing 8 of 8 · example data

SystemOwnerWhat it can doLast testedStatus
Support assistantCustomer serviceReads tickets, drafts replies12 AugApproved
Invoice agentFinance operationsPays invoices under a limit3 SepApproved with limits
Code review botEngineeringComments on pull requests21 JulApproved
Meeting notesVendor toolNo ownerRecords and summarises callsNeverNeeds an owner
Sales email writerVendor toolSalesDrafts and sends emails30 JanDue for review
Claims assistantClaims operationsDrafts claim replies14 JunModel changed
CV screeningVendor toolNo ownerRanks job applicantsNeverNeeds an owner
Demand forecastFinancePredicts stock levels2 MayApproved

Four questions your board will ask.

  • What AI are we running?

    Including the AI features inside tools you already pay for.

  • Who owns each one?

    A real team that answers for it.

  • What can it touch?

    The data it reads and the actions it can take.

  • Why is it safe to run?

    Test results attached to the decision, not a slide.

From policy to proof.

We help you turn written rules into checks that run at every release, with evidence a reviewer can follow.

Write rules people can follow. Short rules your teams can apply, like “customer data only goes to approved AI tools”.

Turn each rule into a release step. Before a system goes live, its owner, data and tests are checked against the rules.

Test before you approve. We test the risky parts: hidden instructions, data leaks and actions that need a person.

Keep a record a reviewer can read. Each decision links to its tests and its owner, so nobody chases screenshots.

Questions we hear.

Do you certify us?

No. We don't certify. We give you records and test results an auditor or regulator can use.

Can this support the EU AI Act or ISO/IEC 42001?

We can line up your register and evidence with the frameworks you report against. Your legal team decides what each one requires of you.

Does AI inside vendor tools count?

Yes. We help you find it, give it an owner and decide what data it may see.

How do we keep the register current?

Set triggers. A new model, tool or data source brings a system back for review.

Where do we start?

With one live AI system. We connect it to its owner, checks and release decision, then use it as the template for the rest.

AI governance & assurance

Make your AI governance usable.

Start with one live system. We'll connect it to its owner, its checks and its release decision.

Review AI governanceCompliance & assurance