Scoped tight. Delivered in the open.
The same discipline on every engagement, because the whole value is that our judgement can be trusted. Here is how we work, and the rules we bind ourselves with.
Frame the question
One conversation to establish what you actually carry, and whether we are the right people at all. If you do not need us, we say so.
Scope, priced to buy
A tight, fixed scope with a clear deliverable and a name attached. No open-ended retainers dressed as strategy.
Deliver, visibly
A finding a board can act on, evidenced against a public benchmark, that stands up when someone tests it later.
The rules we bind ourselves with
Our only asset is that our judgement has no axe behind it. The threat is not corruption, it is drift: a finding softened by a fortnight's delay, a comparison that quietly omits a client's preferred scheme. Nobody would ever notice, which is precisely why we make it mechanical.
- Competence for the role is something we assess through casework and judgement, not just a rubric, read alongside the wider international record (EN 18274, UNESCO, the OECD, the EU AI Act, NIST's AI Risk Management Framework, and ISO/IEC 42001) for the rest of the picture. Every benchmark we cite is publicly available, so it is one anyone can check.
- The only things sold under our name are our own: the advice, the free self-audit, and the Organisation Toolkit.
- If any relationship of ours could bear on a finding, we disclose it to the client in writing before we start, not buried in a policy PDF.
- Who we advise stays confidential. AI ethics work touches live, sensitive decisions, and naming a client would undercut the same discretion they are paying us for.
- Every finding has to survive being said out loud. Before we sign anything, we ask whether we would be comfortable explaining the reasoning, face to face, to everyone it affects, not just to the client who commissioned it. If it only survives being explained to the person paying for it, it is not a finding yet.