AI strategy, design and delivery
I help leaders turn promising AI ideas into useful capability their people can trust, own and extend. Sometimes that means an agent colleague for the team. Sometimes it means a new product, service or experience.
The paradox at the centre of this work
The people whose work is most exposed to automation are usually the people who understand that work best: every rule, every exception and every reason it is done this way rather than that. That knowledge is exactly what an agent needs to be useful.
Automation forced upon someone removes agency. Automation led by them expands their role: defining the work, setting the standard, judging the output and deciding what the agent must never be allowed to do. The work moves up rather than away.
The wheel did not simply lead to a better cart. It changed how people worked and made entirely new products, services and experiences possible. The point was never to improve the cart forever. It was to recognise what became possible once movement became easier.
I work the same way. I build the first capability alongside your team, then hand over the ability to improve it and create what comes next. Sometimes that means improving how people work. Sometimes it means creating something new for customers. Often it means both.
The most valuable thing I leave is the habit of asking: What is working? What is not? What is missing? And what has now become possible?
“Doing yourself out of your current job is not an ending. It is an ongoing occupation.”
Two kinds of work
Both start with the people affected. Both end with something useful running, a named owner and the capability to carry on.
We start with work worth improving: what slows people down, what depends on scarce knowledge and what only they can judge. I help define the role, build and onboard the agent, and establish how people and agents work together day to day.
AI makes propositions viable that were not viable a year ago. I carry the strongest opportunities from decision through design, development and delivery — holding the customer experience, commercial intent and technical shape together until people are using them.
What did not get cheaper
A tool can build what you describe. It cannot tell you that you described the wrong thing, that nobody will own it or that it will fail when real people and real risk arrive. The early decisions now matter more, because every later choice inherits them.
Judgment earned across the whole system
I started building software in 1999 and have worked hands-on through open source, SaaS, low code, no code and now AI. Each shift removed effort from the build. None removed the need to choose the right problem, design the right experience or make the result safe to operate.
I have been a software founder, product owner, delivery manager, enterprise architect and trainer, client strategist, innovation consultant and experience design lead. Most recently I was AI Strategy Lead inside a highly regulated insurer. I am TOGAF-certified.
Different titles, one job. Connecting purpose to decisions, decisions to design and design to working technology — then leaving the people closest to the work able to carry it forward.
A deliberately direct relationship
You work directly with me from the first conversation through to the decisions and details that shape delivery. No account manager and no rotating team.
I use my own agent workforce across discovery, design, architecture and delivery, and bring in trusted specialists when the work genuinely calls for them. That creates more capacity without fragmenting accountability. I work the way I am asking you to work.
Good AI work strengthens human capability. It does not ask people to disappear from the process.
The measure is simple. When I leave, are your people doing more valuable work than when I arrived — and can they carry on without me?
A useful first conversation
Bring the opportunity, the difficult decision or the workflow that will not improve by itself. We can work out whether AI belongs in the answer — and what a sensible first move looks like.
