Thought Leadership
Enterprise Intelligence Insights
Research, briefings, and commentary on enterprise AI — written for practitioners, not academics. Published irregularly, when there is something genuinely worth saying.
- Featured
- Featured
Most AI failures are not model failures. They are management failures. The introduction to a series on what AI governance means when an organisation has to implement it, operate it and be accountable for it.
October 14, 2026
Article series
AI Governance in Practice
A series on what AI governance means when an organisation has to implement it, operate it and be accountable for it. Start with the introduction: its series box gives the reading order, and new articles join the series as they are published.
21articles
Articles
- ai risk, Insights
September 20, 2026
The worst documented AI failures were not unforeseeable. Six cases, the control each one lacked, and the audit question that would have exposed it.
- Insights, Regulation
September 21, 2026
“We’re not in the EU” is one of the most expensive assumptions in AI governance. What the Act is, why it reaches beyond Europe, and why every deployment needs a recorded scope determination.
- Insights, Regulation
September 22, 2026
ISO/IEC 42001 is the most consequential AI standard most executives have never read. Should governments mandate it? An auditor’s answer: not a blanket mandate, but a targeted one, and where the line falls.
- AI Governance, Insights
September 23, 2026
“We comply with all applicable regulations” is offered as reassurance. In AI it is a warning sign. Why ethics only becomes real inside operational process, and an introduction to the EGIM.
- AI Governance, Insights
September 24, 2026
Most of the AI in most organisations never went through a project or a procurement. Why banning it fails, and why the answer is to govern the ways AI gets in, not just a list of systems.
- Insights, Regulation
September 25, 2026
ISO/IEC 42001 certificates are turning up in AI vendor pitches. What one actually attests, what it does not, and the six things to check before you rely on it.
- Insights, Regulation
September 26, 2026
Article 27 of the EU AI Act asks certain deployers to assess a high-risk AI system’s impact on people’s rights before first use. Who it applies to, what it must contain, and how to start now.
- Insights, Regulation
September 27, 2026
The region has charters and policies. What it lacks is the machinery that turns them into accountability. A three-stage roadmap: lead by example, build the infrastructure, then enforce.
- AI Governance, Insights
September 28, 2026
Most AI governance guidance assumes a risk function, a legal team and an ethics board. Most organisations have none of them. What proportionate governance looks like when you are fifty people.
- AI Governance, Insights
September 29, 2026
Most organisations run on AI models they did not train and cannot open. What to ask for, what to put in the contract, and how to test and monitor a model you will never see inside.
- AI Governance, Insights
September 30, 2026
Policies, committees and registers show that AI governance exists. They do not show that it works. Four levels of evidence, the measures worth tracking, and five questions for your next leadership meeting.
- ai risk, Insights
October 1, 2026
When an AI-supported decision goes wrong, responsibility tends to vanish between many hands or land on the nearest person. What the law says, and how to design accountability before harm rather than after it.
- ai risk, Insights
October 2, 2026
An AI system that advises can be wrong. An AI agent that acts can be wrong at speed, at scale and with your credentials. How to set autonomy by risk, and the controls that keep agents within bounds.
- AI Skills, Insights
October 3, 2026
The EU softened the AI Act’s literacy duty in 2026. The risks it was meant to address have not gone anywhere. What changed, why it still matters, and what good literacy looks like by role.
- ai risk, Insights
October 4, 2026
The film version has glowing red eyes. The real record is less cinematic and more useful: what is documented, what is genuinely debated, and what governance can actually do about each.
Research
The report and the paper behind the series.
- Insights, Research Paper
October 5, 2026
Eighteen documented cases of AI bias and failure, a taxonomy of how AI systems fail, the regulatory picture as at October 2026, and six practices mapped to ISO/IEC 42001, the EU AI Act and the Ethical Governance Integration Model.
- Insights, Research Paper
October 6, 2026
Why compliance alone is not enough in high-stakes organisational environments, and the research case for a values-driven framework for AI governance.
More insights
- AI Strategy
October 7, 2026
The majority of AI governance initiatives collapse not from lack of policy, but from lack of operational backbone. Here’s what separates frameworks that work from those that gather dust on intranets.
- AI Strategy
October 8, 2026
Organisations treat these as competing demands. They aren’t they reinforce each other, and confusing them costs both money and political capital.
- AI Strategy, Insights
October 9, 2026
An endless stream of POCs and a shortage of production models. The causes are structural, not cultural and the fix is organisational, not technical.
- Featured
- AI Governance, Insights
Policies, committees and registers show that AI governance exists. They do not show that it works. Four levels of evidence, the measures worth tracking, and five questions for your next leadership meeting.
September 30, 2026
Articles
- AI Governance, Insights
September 30, 2026
Policies, committees and registers show that AI governance exists. They do not show that it works. Four levels of evidence, the measures worth tracking, and five questions for your next leadership meeting.
- AI Governance, Insights
September 29, 2026
Most organisations run on AI models they did not train and cannot open. What to ask for, what to put in the contract, and how to test and monitor a model you will never see inside.
- AI Governance, Insights
September 28, 2026
Most AI governance guidance assumes a risk function, a legal team and an ethics board. Most organisations have none of them. What proportionate governance looks like when you are fifty people.
- AI Governance, Insights
September 24, 2026
Most of the AI in most organisations never went through a project or a procurement. Why banning it fails, and why the answer is to govern the ways AI gets in, not just a list of systems.
- AI Governance, Insights
September 23, 2026
“We comply with all applicable regulations” is offered as reassurance. In AI it is a warning sign. Why ethics only becomes real inside operational process, and an introduction to the EGIM.
Articles
- Insights, Regulation
September 27, 2026
The region has charters and policies. What it lacks is the machinery that turns them into accountability. A three-stage roadmap: lead by example, build the infrastructure, then enforce.
- Insights, Regulation
September 26, 2026
Article 27 of the EU AI Act asks certain deployers to assess a high-risk AI system’s impact on people’s rights before first use. Who it applies to, what it must contain, and how to start now.
- Insights, Regulation
September 25, 2026
ISO/IEC 42001 certificates are turning up in AI vendor pitches. What one actually attests, what it does not, and the six things to check before you rely on it.
- Insights, Regulation
September 22, 2026
ISO/IEC 42001 is the most consequential AI standard most executives have never read. Should governments mandate it? An auditor’s answer: not a blanket mandate, but a targeted one, and where the line falls.
- Insights, Regulation
September 21, 2026
“We’re not in the EU” is one of the most expensive assumptions in AI governance. What the Act is, why it reaches beyond Europe, and why every deployment needs a recorded scope determination.
Articles
- AI Strategy, Insights
October 9, 2026
An endless stream of POCs and a shortage of production models. The causes are structural, not cultural and the fix is organisational, not technical.
- AI Strategy
October 7, 2026
The majority of AI governance initiatives collapse not from lack of policy, but from lack of operational backbone. Here’s what separates frameworks that work from those that gather dust on intranets.
Articles
- ai risk, Insights
October 4, 2026
The film version has glowing red eyes. The real record is less cinematic and more useful: what is documented, what is genuinely debated, and what governance can actually do about each.
- ai risk, Insights
October 2, 2026
An AI system that advises can be wrong. An AI agent that acts can be wrong at speed, at scale and with your credentials. How to set autonomy by risk, and the controls that keep agents within bounds.
- ai risk, Insights
October 1, 2026
When an AI-supported decision goes wrong, responsibility tends to vanish between many hands or land on the nearest person. What the law says, and how to design accountability before harm rather than after it.
- ai risk, Insights
September 20, 2026
The worst documented AI failures were not unforeseeable. Six cases, the control each one lacked, and the audit question that would have exposed it.
Enterprise Intelligence Briefing
A short email every few weeks with one genuinely useful observation about enterprise AI. No noise, no summaries of other people’s newsletters.
Work email