Responsible AI — Strategy, Governance & Delivery

I have been thinking about
responsible AI since 2018

David Viney

Before "agentic AI" was a phrase. Before the EU AI Act was drafted. Before most organisations had an AI policy. I have been building AI systems, confronting their governance implications, and developing frameworks to manage their risks — for longer than most people have been talking about it.

Book an MOT Explore MASTER-AI™

It was early 2018. I was building an observability tool at Arup — powered by what we were still calling Machine Learning. Joi was a simple Azure app that would log into our top business systems, perform a set of transactions, and log out again. She would flag if systems were performing well, degraded, or experiencing an outage.

Joi needed her own identity in the HR system and Active Directory. Her own paid licence for each system. Her own userID and password. She was, literally, the first artificial employee of the company — a new generation of synthetic beings who didn't get paid, didn't need to sleep, and was happy to work 24/7.

The question someone eventually asked — half-joking, nobody laughing — was: "How long will it be before Joi takes my job?" That question has never left me. Everything I have built and advised on in AI since then has been shaped by it.

Proprietary Framework
MASTER-AI
A practical six-domain framework for responsible AI governance — grounded in ISO 42001, the EU AI Act, and NIST AI RMF. Not process theatre. A working tool for boards and technology leaders who need to act before the regulatory window closes.
Model lifecycle Algorithm explainability Supply chain Trusted data Engaged humans Risk management
Explore the framework →
What this covers

Three things I do here,
each with real depth

Mode 1 — Consulting Mode 2 — Assurance Mode 3 — Delivery
01

AI Strategy & Fractional CAIO

Most organisations have an AI ambition problem, not a technology problem. The models exist. The tools are available. What is missing is sustained executive-level leadership — someone accountable for strategy, governance, and delivery outcomes at the leadership table.

As fractional CAIO I embed at the leadership level — owning the AI strategy and roadmap, designing and governing the AI trust framework, and building the board's confidence that AI is being adopted responsibly. The MASTER-AI™ framework is the methodology behind every engagement.

Mode 1 — Consulting
Global Engineering Consultancy · 10,000+ staff · Arup
Joi — The First Artificial Employee
Built and deployed Joi — an early agentic AI system that logged into Arup's top ten business systems every six minutes, performed a set of transactions, and reported on system health. Joi required her own organisational identity: HR record, Active Directory account, paid software licences, and individual data permissions. The governance questions this raised — about AI identity, accountability, and the human implications of synthetic workers — became the foundation for everything that followed.
First agentic AI system deployed at Arup · governance implications surfaced and addressed · foundation for MASTER-AI™ framework
02

AI Governance Audit & Standards Alignment

The EU AI Act comes into force in August 2026. ISO 42001 is already the international standard for AI management systems. Most organisations are significantly behind on both — and the gap between where they are and where they need to be is wider than they realise.

As independent reviewer I assess an organisation's AI governance posture against the relevant frameworks — identifying the gaps, quantifying the exposure, and producing a board-ready findings report with a prioritised action plan. I also advise on the standards themselves, working at the level of framework design rather than just compliance.

Mode 2 — Assurance
AI Governance Standards Body · Accounting Sector · UK & International · PPCS
Board Advisory — AIGAS Framework — PPCS
Non-executive board adviser to PPCS and the AI Governance and Assurance Standards (AIGAS) framework — contributing to the development of an AI governance standard for the accounting and professional services sector. Providing independent challenge on framework design, alignment with ISO 42001, the EU AI Act, and NIST AI RMF, and ensuring the resulting standard is practically implementable rather than theoretically correct.
Active board adviser · AIGAS framework under development · ISO 42001 and EU AI Act alignment assured
03

Agentic AI Platform & Transformation Delivery

Strategy without delivery is noise. I have built and deployed AI platforms at genuine enterprise scale — not proof of concepts, not pilots, but production systems used by tens of thousands of people generating documented commercial value.

As delivery lead I run AI transformation programmes from strategy through to production — technology selection, partner management, data governance, change management, and the responsible AI framework that ensures what gets built can be trusted. The six years between Joi and WPP Open represent the longest and most deliberate preparation for this work that I know of.

Mode 3 — Delivery
Complex, federated Media Business · 70,000 users · 110 countries · WPP plc
WPP Open — Agentic AI Platform Delivery
Platform Services Director responsible for building and deploying WPP Open — the world's first agentic AI operating platform for end-to-end marketing. A £250m programme delivered in partnership with NVIDIA, Google Gemini, IBM Watson, and Star. Deployed to 70,000 practitioners across 110 countries, with a responsible AI governance framework embedded throughout. Six years after asking the governance questions that Joi raised, this was the answer at enterprise scale.
$40m documented commercial value in year one · 70,000 practitioners · 110 countries · Major Business Win with Nestlé · Open placed at heart of WPP business strategy

The right starting point
is usually the MOT

Whether you need an AI governance audit before August 2026, a fractional CAIO to own your AI agenda, or a delivery partner for an agentic AI platform — the Technology MOT gives both of us a clear picture of your AI readiness first. Fixed price. Fixed scope. No ongoing commitment. The MOT pays for itself.

Book an MOT 30m Therapy Session

“As an experienced and astute leader he was able to look at challenges from multiple perspectives, and communicate effective solutions — ensuring that I understood the ‘Why’ as much as the ‘What’.”

— Steven Charlton, Strategic Transformation Leader, WPP