When regulators ask whether a decision can be reconstructed, most boards discover they have no evidence. Not documentation — evidence. The complete information picture that existed at the moment the system decided. I help boards in financial services, insurance, and regulated technology close the gap between adoption and defensibility.
For NEDs, Audit Committee Chairs, General Counsel and Chief Risk Officers · Financial services · Insurance · Regulated technology · London market
The most dangerous assumption in enterprise AI is that deployment equals governance. Boards approve AI investments, technology teams build and ship, and somewhere between the two, accountability disappears. I work in that gap — helping boards understand not just what their AI systems do, but what those systems knew when they did it, and whether the organisation can prove it under scrutiny.
I trained in astrophysics at Queen's University Belfast, which taught me to extract signal from noise in complex systems under uncertainty. I spent twenty-five years in enterprise technology leadership across insurance, aviation, financial services, and public sector — which taught me that most system failures are governance failures in disguise. I now focus on the gap every board overlooks: the ability to prove what your AI knew at the moment it decided.
Through Otopoetic Limited, I provide AI governance advisory built around two proprietary frameworks: the Governance Classification (A·E·C·R·M), which gives every organisation a precise, five-dimensional governance address, and the Digital Alibi, a forensic evidentiary standard for reconstructing AI decisions as they existed at the moment they were made — not assembled retrospectively when inquiry arrives.
I hold a Fellowship of the Royal Statistical Society and serve as Secretary of its Northern Ireland Local Group. I am a Fellow of the Royal Society of Arts, a Member of the Institute of Physics, author of four books, and am completing an academic working paper on AI governance frameworks at Lloyd's of London, now in peer review. I participate in the GLG and Guidepoint expert networks, with over 50 completed engagements spanning AI governance, cybersecurity, insurtech, and aviation technology.
Based in Belfast. Operating across the UK and Ireland. Writing at The Roche Review.
Twenty-five years in regulated-sector technology leadership — insurance, aviation, financial services — operating, not observing. The advisory work is grounded in a method built from delivery: operations stabilised before transformation begins, specialists engaged through diagnosis rather than mandate, and every engagement structured so the function is more capable when it ends than when it started.
Every engagement begins with triage. I map what runs, what fails quietly, and what is held together by individuals rather than process. I identify the two or three things that break when a key person is absent and fix those before touching anything structural. The transformation mandate only holds if operations do not visibly deteriorate while it runs. Stabilise, earn trust, then transform — the sequence is not optional.
I separate the two tracks structurally through portfolio governance. Quick wins route directly to operational teams without transformation gating. Foundation work — platform, architecture, operating model — gets a proper gate and a business case. One track keeps the function running while the other changes altitude and heading. Running them in sequence means the transformation consumes the function it is meant to improve.
Engineers, actuaries, underwriters, and data teams do not commit to platforms imposed on them. I surface the capacity lost to manual process and show them what that time would buy redirected into the work they were hired to do. I frame the problem in their language before proposing any solution. Mandate produces compliance. Diagnosis produces commitment. The distinction determines whether the change holds after the engagement ends.
Platform decisions, accountability chains, and governance structures are designed from the outset to operate independently. The portfolio governance board, the operating model, and the team capability remain when the engagement ends. The measure of a successful engagement is not what was built. It is what continues to run — at the same standard — without the person who built it.
Fractional CTO / CIO / CDO
Available for fractional executive mandates in insurance, financial services, and regulated technology businesses. The engagement begins with a 30-day triage. Reporting line, governance structure, and exit conditions are agreed at the outset.
A governance-focused NED brings something different from a technology-strategy NED: the ability to ask the question a regulator will ask, before the regulator does. What follows is what I bring to a board mandate.
Using Otopoetic's proprietary A·E·C·R·M framework, I locate your organisation across five independent governance dimensions — Accountability, Exposure, Control, Regulation, Maturity. The result is a precise address, not a traffic light. It maps directly to regulatory ceilings and defensibility gaps. Interactive self-assessment at otopoetic.com →
A structured forensic review establishing whether your organisation can reconstruct the complete information picture behind every material AI-assisted decision — as it existed at the moment it was made. Not retrospectively. Not from memory. This is the evidentiary standard regulators apply. Learn more at otopoetic.com →
A prioritised, board-facing action plan structured around your governance address. Identifies the gaps that carry the most regulatory and fiduciary risk, with clear accountability ownership and an implementation sequence mapped to your regulatory obligations.
Ongoing constructive challenge on AI governance posture, regulatory horizon, and accountability architecture. Ensures the board can answer the questions a regulator, litigant, or shareholder will ask — before they arrive. Includes regulatory monitoring and governance address updates.
I sat on the steering group of the study prepared by SAMI Consulting for the Department for the Economy and launched by the Economy Minister at AICON 2026 on 23 September 2026. It names post-quantum migration the clearest near-term commercial opportunity for Northern Ireland and recommends a specialised, application-led approach. Read the announcement · Read the report.
Most quantum discussions ask when the machine will arrive. The harder question for a board is whether a decision made today will still be defensible once the assumptions behind it have changed. The record, not the forecast, is what gets tested. The article sets out four elements worth minuting: assumptions, evidence base, alternatives considered and review triggers.
Keynotes, board sessions and foresight workshops that turn the NCSC milestones of 2028, 2031 and 2035 into minuted review triggers, with the reasoning on the record. See how to engage me.
Providing independent CTO-level oversight and constructive board challenge to a growing B2B technology company. Focused on technology strategy, governance frameworks, and ensuring alignment between investor commitments and operational delivery.
Governance, technology strategy, and KPI framework design for a not-for-profit delivering art-based programmes for neurodivergent individuals and young people. Leading phased technology architecture planning and board-level strategy documentation.
Boutique strategic technology advisory and technical due diligence consultancy. Serving PE firms, institutional investors, and legal counsel with systems thinking, AI governance, and forensic technology analysis.
Senior technology executive mandate for regulated-sector organisations that require operational leadership without a permanent appointment. Previous engagements have spanned insurance, financial services, and aviation, covering programme delivery, team build, vendor governance, and board-level accountability. The engagement begins with delivery, not discovery.
Member of the steering group for the Matrix NI study, prepared by SAMI Consulting for the Department for the Economy and launched by the Economy Minister at AICON 2026 on 23 September 2026. The study sets out where Northern Ireland should focus its quantum effort and names post-quantum migration as the clearest near-term commercial opportunity. Read the announcement.
Contributing to Reed's innovation strategy, providing practitioner insight on the intersection of AI, recruitment technology, and workforce transformation from a governance and advisory perspective.
Organising the regional programme for the Royal Statistical Society, connecting statisticians, data scientists, and researchers across Northern Ireland with the national professional body.
PE firms and institutional investors use expert networks to vet advisors before engagement. The record below is what they find.
Boards need to know their advisor speaks the regulatory language that will govern AI accountability. I advise boards on the following frameworks — not from a compliance checklist perspective, but from the question a regulator or litigant will actually ask.
High-risk AI system obligations under the EU AI Act require board-level defensibility evidence — not documentation of intent. These high-risk obligations were deferred from 2 August 2026 to December 2027 under the Digital Omnibus, but the deferral moves the conformity deadline, not the accountability standard — governance structures must be in place, not assembled afterwards.
Senior managers bear personal accountability for AI governance failures under SM&CR. The question is not whether the board signed off a policy — it is whether the senior manager can prove they understood, challenged, and owned the governance at the moment decisions were made.
DORA's third pillar requires contemporaneous decision evidence — not retrospective documentation. Organisations must demonstrate that governance structures were operational at decision time, not reconstructed in response to an incident.
NIS2 requires evidence of governance maturity and human oversight capability at system design and decision time. Organisations that can demonstrate their oversight architecture precedes deployment are in a fundamentally different position under inquiry.
Completing an academic working paper on AI governance frameworks at Lloyd's of London, including a NED Competency Framework and regulatory control matrix for the insurance market. Forthcoming on ResearchGate; currently in peer review with Paul McGee, co-founder of SITP London.
The NCSC expects a complete initial migration plan by 2028, migration of the highest-priority services by 2031 and completion across all systems by 2035. For a board the question is not whether it predicted the timeline. It is whether the assumptions, evidence base, alternatives considered and review triggers behind each quantum-related decision are on the record.
AI systems that process personal data in automated decision-making create specific obligations under UK GDPR and ICO guidance. Board-level accountability requires clear ownership of the data governance layer within the AI decision pipeline.
Why 2 August still matters in the boardroom: where the EU AI Act and FCA SM&CR converge →
An initial conversation to understand your governance address — where you sit across the five A·E·C·R·M dimensions. Not every organisation has the same gaps. This conversation determines what matters most and what the right engagement looks like.
A structured assessment using the Governance Classification framework. Establishes your precise governance address, identifies defensibility gaps, maps regulatory obligations to your specific sector ceiling, and tests whether your AI decisions can be forensically reconstructed. Delivered as a board-ready document.
A prioritised action plan with clear accountability ownership. For NED mandates, this becomes the foundation for ongoing constructive board challenge — ensuring the board can answer the question a regulator will ask before the regulator does. Includes Digital Alibi infrastructure design for organisations that need contemporaneous decision evidence.
These are the ideas I keep returning to — in boardrooms, in advisory work, and in the writing I publish at The Roche Review. They are the questions I believe every board deploying AI should be able to answer before a regulator asks them.
Eighty-eight hours of machine computation produced a claim about one of mathematics’ seven hardest problems. The Clay Mathematics Institute’s own rules require two years of published scrutiny before anyone may call it solved.
Read the articleThe gap between an incident and the record of who knew what, and when.
Read the articleNaming a champion is an announcement. Accountability is a record.
Read the articleArticle 50 and the FCA Senior Managers regime converged on 2 August. The board test is narrow: a named person, a dated record, a decision that predates the obligation. One agenda item for the next board meeting.
Read the articleA named external body, rules written before the claim, and a clock the claimant cannot shorten. Three questions make it a workable board test: attribution, the record and the clock.
Read the articleDirectors are rarely judged on foreseeing the future. The record of why a decision was reasonable when it was made is what gets tested. Four elements worth minuting, and the published NCSC dates to anchor review triggers.
Mastering AI Conversations to Lead the Future. An 11-chapter guide for leaders on the art and science of prompt engineering, covering practical exercises, case studies, ethical implications, and emerging trends in multimodal AI.
View on AmazonEmbracing the AI Revolution for a Future-Ready Workforce. A practical guide to building AI-capable teams, scaling AI across organisations, and embedding ethics at every level of workforce development.
View on AmazonBelfast's journey from Linenopolis to Tech Hub. Tracing the city's innovative spirit from Queen Victoria's 1849 visit through the Titanic era and the Good Friday Agreement to its emergence as a technology centre.
View on AmazonCork's Technological Odyssey. Charting Cork's journey from monastic settlement in 650 AD to a modern technology hub, weaving together pivotal moments in the city's history of resilience and innovation.
View on AmazonAI governance and executive risk, explored in depth. Twenty-seven issues published to date, covering board accountability, defensibility standards, agentic AI risk, and the regulatory environment under the EU AI Act. For boards that need to know more than their briefings tell them.
Browse all issuesA practitioner-focused working paper on AI governance frameworks at Lloyd's of London, including a NED Competency Framework and regulatory control matrix for the insurance market. Forthcoming on ResearchGate; currently in peer review with Paul McGee, co-founder of SITP London.
ResearchGate profileFor governance advisory through Otopoetic, request a confidential briefing at otopoetic.com.
Your enquiry has been received. I will respond within two working days.