From casual use to repeatable workflows
Turn useful patterns into shared workflows, examples, and habits.
From AI uncertaintyto working systems
AI is on the agenda. The harder question is where it actually belongs.
I help organisations separate signal from noise, identify where AI can genuinely improve work, and turn the right ideas into workflows, internal tools, and software people can use.
My work sits between leadership conversations, delivery realities, and implementation detail.
Experience shaped across Shell · Tony Blair Institute for Global Change · Financial Conduct Authority
energy · public policy · financial regulation · healthcare IT · manufacturing · enterprise IT · software engineering · AI enablement · delivery · governance
The gap is not access to AI. It is translation.
Most teams already have tools. What they often lack is a way to turn possibility into judgement: which use cases matter, what risks are real, what should be tested first, and what needs to be built properly.
Without that translation, AI becomes scattered: private prompts, disconnected experiments, unclear ownership, and demos that do not survive contact with real work.
That is where I help: turning uncertainty into decisions, and decisions into systems that can actually operate.
From access to capability
Many organisations already give their people access to ChatGPT, Copilot, Claude, or internal AI tools. Access is not the same as capability.
I help teams learn how to use AI in real work: asking better questions, designing repeatable workflows, reviewing outputs, handling sensitive information, and knowing when not to use AI at all.
The goal is not to make people “prompt engineers”. The goal is to help them become better at their own work with AI as a practical support layer.
From impressive outputs to better judgement
Learn to question, verify, adapt, and supervise AI outputs.
From tool access to business value
Apply AI to real tasks, decisions, documents, and processes.
From anxiety to confidence
Provide guidance, guardrails, examples, and context so people know what is safe and useful.
Where I’m useful
I am most useful when the problem is not purely strategic and not purely technical — when the organisation needs someone who can connect both.
For leadership teams
I help turn AI uncertainty into clearer options, trade-offs, priorities, and decisions. The aim is not to sound innovative. The aim is to know what is worth doing, what is not, and why.
For teams learning to use AI well
Many teams already have access to ChatGPT, Copilot, Claude, or internal AI tools. I help them move from scattered experimentation to practical habits, repeatable workflows, safer usage, and better judgement.
For delivery teams
I help shape use cases into workflows, prototypes, and systems that can be tested, improved, and maintained. Good AI work has to survive contact with real users, real data, and real constraints.
For organisations with real constraints
I help keep ambition connected to data, security, governance, adoption, and the way people actually work. AI has to fit the operating environment, not just the demo.
How I work
The work usually starts before tools, platforms, or vendors. It starts by understanding the situation well enough to make better decisions.
Understand the real context
I start with the work itself: goals, constraints, users, stakeholders, data, systems, and the decisions people are trying to make. The aim is to build a shared understanding before recommending tools.
Separate signal from noise
I help identify where AI or software can create genuine leverage, where the idea is premature, and what assumptions need to be tested before time or budget is committed.
Shape the first useful step
That may be a workflow, prototype, internal tool, delivery plan, governance pattern, or implementation roadmap. The point is to make progress concrete enough to learn from.
Build or support implementation
Where useful, I can help design, build, review, or guide delivery — keeping the work practical, maintainable, and aligned with the organisation’s constraints.
Why this perspective works
Broad enough to see the system
I have worked across software engineering, enterprise IT, delivery, architecture, and product-shaped environments. That breadth helps me see how technology decisions affect people, process, risk, and implementation.
Technical enough to challenge assumptions
AI adoption often sounds simple until it touches data, integration, quality, security, maintainability, and ownership. I bring enough engineering depth to keep ideas grounded.
Senior enough to translate trade-offs
I am comfortable working with senior stakeholders where the task is to turn complexity into options, risks, priorities, and decisions — without flattening the technical reality underneath.
Practical enough to move the work forward
I do not stop at recommendations. I can help shape experiments, design workflows, build prototypes, support teams, and turn useful ideas into something operational.
Typical engagements
Some work starts with a single unclear question. Some starts with a team already experimenting. Some starts with leadership asking what AI should mean for the organisation. The format depends on the problem.
AI opportunity assessment
A focused engagement to understand where AI could create value, what constraints matter, and which first steps are worth pursuing.
Output: clearer priorities, practical use cases, risks to consider, and a realistic path forward.
AI coaching and workflow enablement
Practical coaching for teams that already have access to ChatGPT, Copilot, Claude, or internal AI tools, but need help using them well in their real work.
Output: role-specific examples, shared workflows, prompting patterns, safe-use guidance, and review habits.
Workflow prototype
A practical prototype that tests whether an AI-enabled workflow can reduce friction, improve quality, or create measurable operational value.
Output: something concrete enough to test, discuss, improve, or discard.
Internal tools and automation
Designing and building small, useful systems that connect tools, organise data, automate repetitive work, or support better decisions.
Output: working technology that saves time and reduces operational drag.
Leadership advisory and delivery support
Helping leaders and teams make informed technology decisions, prioritise work, and move from strategy to implementation with less ambiguity.
Output: better decisions, clearer trade-offs, and support through delivery.
Ideas, observations, and practical lessons from the field.
Writing on AI adoption, software engineering, LLM workflows, and the gap between impressive demos and systems people can actually rely on.
AI Adoption Is Not a Tooling Problem
Useful AI adoption starts by understanding the work, protecting craft, and turning scattered AI usage into operating models.
Read articleSkill issue? Probably. But that’s not the point.
An engineering rant from the post-hype side of AI-assisted development. AI makes you faster, then hands you the bill: more code, more drift, more review, and broken mental.
Read articleNot sure where AI fits yet? That is a good place to start.
You do not need a fully formed AI strategy before we speak.
Often the useful work starts with a messy question, a promising idea, a team under pressure, or a sense that there may be a better way to operate.
If the path is still unclear, that is usually the right time to talk.
Start a conversation