Practical thinking on AI adoption.
Field notes, frameworks, and perspectives from working directly on AI transformation — in commercial settings, in local government, and at board level. Published when there's something worth saying.
Why most AI pilots fail to scale
The issue is rarely the model. More often it's workflow design, operating discipline and governance. Organisations run successful pilots — tools that work, time saved, people engaged. Then nothing changes at scale. The pilot stays a pilot. Understanding why requires looking at what a pilot actually tests, and what it almost never tests: whether the organisation is ready to make AI a permanent part of how work gets done.
Read the insightTool → Assistant → Worker: a practical guide
A way to understand how AI capability matures inside organisations — and what leadership decisions need to change at each stage. Most organisations are at Tool. Some are at Assistant. Very few have reached Worker in any meaningful way. The framework helps leaders know where they are, what comes next, and what it actually requires.
Read the frameworkAI governance that actually works
Most AI governance frameworks are either too vague to be useful or so rigid they slow everything down. Good governance enables adoption — it doesn't obstruct it. This piece draws on the experience of designing and building a governance framework that went to board approval in a listed company, and what that process revealed about what boards actually need to understand.
Read the insightDesigning hybrid human–AI workflows
Where human judgement sits, where AI helps, and how to create the right balance between speed and control. The question most workflow redesigns skip: what does the human actually need to know, check, or decide — and what happens if they don't? A practical guide to designing workflows that are faster because of AI, not riskier.
Read the guideThe operating model question nobody is asking
Organisations are asking "which AI tools should we adopt?" The more important question is "how will our operating model need to change when AI becomes part of the workforce?" The answer involves accountability structures, management practices, capability requirements and leadership behaviours that most organisations haven't begun to think through. This piece makes the case for starting now.
Read the insightAI adoption in local government: what's different
AI adoption in the public sector isn't slower because people are resistant. It's slower because the governance environment is fundamentally different — and that's not a problem, it's a design constraint. Written from direct experience of overseeing AI adoption as a Cabinet Member, this piece explains what those constraints are and how to work with them rather than around them.
Read the insightWhat the work covers
The insights here are grounded in one consistent question: what does it actually take to make AI work inside real organisations?
Operational adoption
Why the gap between experimentation and adoption is where most AI programmes stall — and what it takes to close it
Governance design
What good governance looks like in practice: enabling rather than obstructing, and credible at board level
Workflow & capability
How work needs to change when AI is part of it — and what teams and leaders need to be able to do
Sector context
How AI adoption looks different in local government, financial services and professional services — and why that matters
Ready to talk about your situation?
If something here has prompted a question about your own organisation's AI adoption, a 30-minute conversation is a good place to start.