Governed AI transformation for regulated enterprises

AI ambition.
Institutional
proof.

We help boards and executive teams decide where AI belongs, redesign how the organisation works, and prove what changed.

Begin a conversation See the work

Structure. Judgement. Consequence.

Principal-led. Accountable through execution.

Explore the practice

01 / The situation

When activity stops becoming progress.

The AI strategy is producing activity, but the institution is not changing.
A major transformation has lost strategic coherence, and nobody can say precisely where.
Legacy systems have begun dictating product and operating choices.

These are questions about how the institution works. That is where we begin.

02 / Selected work

A portfolio is only as good as its decisions.

All selected work

Global insurance brokerage

Twenty-plus AI initiatives.
Seven regions.
One question: what matters?

A global marketing organisation with twenty-plus AI initiatives across seven regions, and no reliable way to say which of them mattered.

Read the engagement

Engagement summaryUnited States · Global

The constraint

Initiative volume had outrun the decision system. Nothing decided what deserved to exist, so everything did.

What we put in place

An operating model, governance rubric and change programme for the whole portfolio.

The decision enabled

Which initiatives to fund, challenge or stop, on stated grounds rather than on whoever argued hardest.

A decision rubric and enablement structure the marketing organisation runs with its own people.

Diversified conglomerateFrom asserted plans to examined evidence Operational risk practiceRisk guidance at the point of decision

03 / The method

One thread. From intent to evidence.

Governance holds when the decision to act, the authority to execute and the evidence of change remain connected. We design and hold that connection.

One accountable design

Decide
Establish what deserves to be funded, challenged or stopped.
Decision rubric
Redesign
Put authority, operating structure and adoption in place.
Operating model
Execute
Modernise and build around a business that keeps running.
Governed delivery
Prove
Show what changed against the institution’s own baseline.
Inspectable evidence

Authority and evidence, held together from the first decision to the final handoff.

04 / Working together

Three ways in. A clear end to each.

Each engagement is led by the principals, bounded in writing and priced to an outcome. Most relationships begin with a diagnosis.

012 to 4 weeks

Executive Diagnosis

A focused examination of the problem beneath the visible problem.

Explore the engagement 02

Institutional Design

The redesign of the machinery the diagnosis found wanting.

Explore the engagement 03

Governed Intervention

Accountability held through a consequential programme, from decision to evidence.

Explore the engagement

Authority stays with you

You retain operational authority, production access and the decision to go live. Those do not move, on any engagement. We remain accountable for the design holding together.

05 / The principals

The people in the room do the work.

Fifty-five years inside regulated institutions and building companies from zero. Direct involvement from diagnosis to handoff.

Zia Karim

Principal · Enterprise modernisation and governed AI

Thirty years in two halves: fifteen in investment-banking engineering, including a decade at Goldman Sachs; then fifteen across startups, fintech and enterprise technology — technology strategy inside Barclays' $800M Change-the-Bank programme, a product organisation restructured at Temenos. Builds governed-AI infrastructure and has published eight titles on institutional change.

Mushter M. Moin

Principal · AI and digital product strategy, adoption

Twenty-five years building digital products and the commercial operations around them: two companies founded, digital head of a $10B insurance division, five granted US patents. Leads AI strategy and adoption from Fortune 50 platforms to a conglomerate's planning cycle.

Meet the principals

06 / Published thinking

The reasoning is on the record.

Our perspective

Authority & accountability

Permission to Act

Permission to use a tool is not permission to act. The operational boundary that determines what AI may do, when it must stop, and who owns the consequences — with the evidence to defend each decision.

Clarkehouse PressSeptember 2026Explore the book

Economics & institutional value

The Mispricing of Intelligence

The seller’s revenue is not the buyer’s return. Why cheaper intelligence does not guarantee enterprise value, and how accountability, organisational design and capital allocation determine who earns a return on it.

Clarkehouse PressSeptember 2026Explore the book

A consequential question deserves a conversation.

Tell us what your institution is trying to change. You will speak directly with the principals who would lead the work.

Begin a conversation