I’ve always been
hooked on how
things work.
I’m Steve Quinlan. Give me a system and I’ll want to pull it apart, find the bit that’s quietly holding everything back, fix it, and watch the whole thing move. That instinct showed up early, in data: there’s a real satisfaction in spotting the pattern in the numbers that nobody else has seen yet.
That curiosity ran through a decade of experimentation work: testing ideas with real data before committing to them. It runs straight into AI today. The context is new. The discipline is unchanged. Not a pivot. A straight line.
These days I help established UK firms move from one-off AI experiments that work most of the time to systems that work on repeat, with the business at the centre of every key decision.
Different sectors,
same pattern.
For over a decade I ran testing and optimisation work across charities, banks, retailers and car makers. RNLI, Farrow & Ball, NatWest, Vitality, LV=, Alfa Romeo. The sectors changed wildly. The pattern stayed the same: find the win that proves what’s possible, and the bigger programme follows.
At the RNLI, getting +28% more website visitors to complete their donation was the moment a sceptical board agreed to fund the full programme. The number was small enough to argue with and big enough to matter. It showed what happens when you test things instead of guessing.
Doing that work for years teaches you more than how to run a test. It teaches product thinking, user experience, prioritisation, and how to talk to engineers, designers and commercial leads in their own language. That mix is what lets me spot where a system is quietly losing value before anyone else does.
So the discipline underneath all of it is simple. Start with the problem, not the tool. Measure what’s actually happening. Prove the smallest version before you commit the budget. Fix the biggest gap first.
The instinct to test, measure, and iterate does not switch off because the tools have changed.
Work with AI,
don’t bolt it on.
Most firms do the opposite. They bolt AI on. A ChatGPT subscription here, a custom chatbot there, a course someone attended last year. Each piece works on its own. None of it adds up to anything the business can feel.
Here is the thing the market keeps missing. You do not get more out of AI by buying more of it. You get more by being deliberate about which AI to use where, on which problem, in this firm, and by using evidence to decide.
I design AI systems the same way I designed experimentation programmes: start with the problem, measure what is actually happening, fix the biggest gap first. The business stays at the centre of every key decision. The aim is never the cleverest tool. It is the smallest change that makes the next job faster, better, or pointed at the right thing.
This is the thinking behind the AI Opportunity Roadmap.
AI Opportunity RoadmapBeyond the brief.
Two side projects say more about how I think than any case study.
ChatGPSteve.
ChatGPSteve is an interactive version of my CV. It’s a chatbot trained on 12 years of my career, so you can ask it what I’ve actually done instead of scrolling a timeline. I built it to test a simple idea: what does a CV become when you can have a conversation with it? It taught me as much about AI’s limits as its promise, and that thinking now shapes the client work.
Try ChatGPSteveThe bookshelf.
The bookshelf is where I keep the books that have shaped how I work. The Lean Startup, Testing Business Ideas, Hooked, and the product titles I keep coming back to. It’s a window into the ideas behind the method.
Browse the bookshelfI test my own ideas on myself first. Then I bring what survives to clients.
Where to next.
Three places to go from here:
Let’s talk.
If your firm has spent on AI and it has not yet added up, that is the problem I solve. Tell me what you have tried, and I will tell you where I would start.