The Picken Governance Institute™
Liverpool City Region · United Kingdom

Boards are judged in financial terms.
AI governance is not.

AI Governance Direct closes that gap. The Picken Diamond® turns board-level AI governance into a scored assessment, a priced liability, and a signed verdict that names who is accountable — in the language a board, an auditor and a regulator already use.

What this practice does

The instrument

The Picken Diamond® assesses AI governance across five pillars and returns a Board Confidence Score, a quantified value at risk, and a board liability verdict.

It applies established financial theory in reverse: instead of pricing an asset, it prices the governance failure attached to one.

The corpus

Nineteen working papers apply the framework to specific sectors — clinical AI in NHS trusts, SEND assessment in local authorities, cyber accountability, synthetic media, biometrics, professional services.

Each is public, dated and citable. The argument in full is set out in The AI Fairness Machine.

The engagement

Fractional governance director, board assurance review, or a single scored assessment with a written verdict.

Forty years of board-level finance sits behind it, most of it spent explaining risk to people who had to sign for it.

The instrument family
What state are we in?
The Picken Diamond® — five-pillar assessment
Board Confidence Score
Rose / Buttercup / Pudding / Critical — lifecycle position
What is it costing us?
Governance Dividend — value retained
Governance Decay Premium — value destroyed
PAAPM+ — governance-adjusted expected return
What would failure cost?
PAVR — value at risk, 95th percentile
GALC — governance quality to liability
ALES — the resulting exposure
Who carries it?
PABLM — the signed board verdict
PAQOS — the Single Signatory Theorem
Where does it spread?
PACB — the AI Conveyor Belt · PADMG — boards and non-executive directors · PALIA — general and life insurers · PACSA and PACES — cyber accountability

And what it does not measure. The taxonomy names seven areas the framework does not yet price — training data provenance among them — and declines to build instruments for any of them until each has had the same empirical grounding as the rest. A measurement system that claims complete coverage is the one to distrust.

Register
Company
AI Governance Direct Limited — No. 16860497, England and Wales
Trade mark
The Picken Diamond® — UK00004378626, classes 35 and 42
Framework paper
doi.org/10.5281/zenodo.21807014 — the complete taxonomy, free to read
Book
The AI Fairness Machine — Terence Picken

Named accountability, published identifiers and a traceable audit trail are what the framework asks of its clients. It would be a poor advertisement not to publish its own.

Terence Picken — Founder and Director
Free AI Governance Health Check · five minutes · no sign-up

The free AI governance health check: would yours survive an ISO 42001 readiness review?

Fifteen questions across the five pillars, scored on the same evidence rubric used in board-level assessment. No documents to gather, and nothing leaves your browser.

Answers stay on your device · no tracking · no database

0 of 15 answered
Assessment result

What this does not tell you

This assessment places you in a band. It does not tell you what the gap is worth. The full Picken Diamond® assessment returns a Board Confidence Score, a quantified financial exposure at the 95th percentile, and a written verdict naming the director who carries it.

For a short readiness note on your answers, put your email address below. It opens a message in your own mail programme with your results already filled in.

The five-pillar structure and the five-point evidence rubric are published in the taxonomy paper. The calibration by which pillar scores resolve to a Board Confidence Score and a quantified exposure is proprietary and is not applied here.

Now tell me where it is wrong

This instrument is published so it can be argued with.

A governance framework that has never been challenged is not a framework, it is an opinion with a diagram. If a band looks wrong for your organisation, if a question misreads how your sector works, or if the ISO 42001 mapping does not match what your auditor asked for — I want to know, and I will say so publicly when you are right.

Post it on LinkedIn where others can see the exchange, or send it privately. Practitioner challenge is how this gets better, and there is no version of it where I benefit from a comfortable answer.

If your board has deployed AI and cannot say what it would cost when it fails, that is the finding.