Independent AI governance
Decide how far your AI goes.
Making AI a business decision, from the boardroom to every desktop.
You wouldn’t give a new employee more freedom, or direct contact with your customers, until they had shown they follow your processes. AI needs the same discipline. Our framework sets out what good AI governance looks like at each stage. Your board sets the principles, nothing moves up a stage without a decision, and we govern it with you month by month, so your business decides how far each use goes and what it costs.
AI board view
Quarter ended 30 September · for the board
Where your AI sits
The five measures
- ValueClaims handling and the finance close return most; customer service is below plan.
- RiskFour higher-risk uses rely on one model provider.
- AccountabilityThe two agents above approval have no named owner yet; owners are being named this month.
Built on recognised standards: your evidence is organised the way ISO/IEC 42001 and the NIST AI Risk Management Framework expect, and mapped to the EU AI Act where it applies to you.
Organisations tend to move through three stages, often several at once, and each asks more of the business.
What each stage asks of you →1 · AI helps your people work
What AI is in use that nobody approved, and what can it see?
2 · AI does the work
Who owns each agent, and is it worth what it costs?
3 · AI becomes your business
Do you own what your AI has learned, and could you prove it?
Start with the framework, then decide.
Each step earns the next: you see results in your own business before you commit to anything ongoing.
- 1 · 45 minutes, freeWalkthroughThe framework, and where your AI sits today.
- 2 · Your choiceChoose where to startSet your AI policy, see what AI is running, or bring a live agent into governance.
- 3 · Then extendProve it in one areaThe framework set up where AI matters most, then across the business.
- 4 · Month by monthGoverned with youAI Usage Oversight, or the AI governance service once AI does the work.
Two managed services, one route as your AI grows.
The framework decides which you need first. What you build in the first carries over into the second.
What we govern
AI runs across six layers, each with a different owner. We give you one view of all six.
AI runs through your people’s devices, your platforms, your business apps, your connectors and agents, your identity and data, and your contracts and spend. Each layer usually has a different owner, so nobody sees the whole.
Every decision and every agent, written down.
Alongside the board view, governance runs on short records anyone can check: what was decided and who owns it, and what each AI agent is allowed to do. Two examples follow.
After the decision workshops
Decision record
One page for each decision your leadership makes about AI, so anyone can see what was decided and who owns it.
Example
One approved assistant for everyone
With no approved assistant to use, staff had turned to their own AI tools, some connected to client files, so leadership approved one assistant for everyone and moved personal-account use onto it. The chief operating officer owns the decision, and the quarterly board view shows whether it is working.
When AI starts doing the work
Agent job description
An AI agent is software that completes tasks on its own. Each one gets a job description, like a member of staff, so its limits and its manager are clear.
Example
Supplier statement reconciler
The head of finance operations manages it. It reads supplier statements and proposes adjustments, and a person approves every one before it is posted.
Every sector has its own regulators, auditors and expectations.
Led by founders from cyber security and IT asset management, two disciplines AI governance draws on.
Ian Turnbull built and sold a cyber security business; Chris Gough built and sold a technology optimisation business. We are independent of the people who build and sell your AI.