I started my career as a business analyst, moved into data science and AI, and stayed there. I grew from individual contributor to lead to team manager, building teams and delivering value. But always through a central function. I admired problems from a distance. I solved them across a divide.
That distance, I eventually realised, was the problem.
So I made a deliberate move. I stepped out of management — where your job is to help the people you manage solve problems that other people own — into an IC role inside a commercial business area. I don’t just own the technical solution anymore. I am now part of the team that has to make things real.
I had a gap. I thought I understood problems end-to-end. I thought my ability to ask good questions and think holistically was a genuine superpower. It was, when looked at through the lens of data and analytics. But that’s a narrow lens.
I’m reminded of an offsite from my time at HEINEKEN, in Egypt, where our then Chief Data & Technology Officer talked about “connecting the dots”. He said it was the most important thing we could do as D&T. I thought I understood what he meant. I was very wrong.
Sitting inside the business, you realise how much complexity sits behind the problems you used to solve from the outside.
I’m surrounded by people who live these problems — who see the world from a 30,000 foot view all the way down to a metre away. People who know what it takes to execute in the real world. That exposure is reshaping how I build. Slower in some ways. Better in the ways that matter.
I started noticing something else, too. A complex optimisation problem that would have taken a team one or two months to build properly — I could now translate into working code in a day, with a bit of back and forth with an AI tool. This was a signal. The automation of writing code had already begun, and it was only going to accelerate.
Agentic AI is closing the gap between the person who understands the problem and the person who builds the solution. That gap used to justify entire teams. It’s shrinking fast.
I’m being forced to rely more and more on agentic coding myself — there are too many problems to solve and too many of them are urgent. And the question I keep arriving at is simple: if I can do this, what’s stopping someone else? It won’t take a technical revolution. It will take a commercial manager who gets curious and builds something that used to require a central function and six weeks of work.
And so I find myself wondering if my skillset will become obsolete. When the cost of building moves towards zero and everyone becomes a builder, what does data science and analytics look like? Does the 90% that AI can do better make the 10% it can’t increasingly important? I don’t know.
I’ve been connecting dots my whole career. But I was doing it in my neighbourhood without ever exploring the city. I’m closer now but still not close enough for what’s coming.
This move came out of intuition and a desire to learn. It’s put me in a place where I don’t have the skillset — thinking operationally, seeing the full breadth of a business problem, is something I’m still learning. And the skillset I do have, the thing that makes me useful here right now, may not be needed in the same way in a year.
That’s uncomfortable.
But I’d rather be uncomfortable and close than confident and far away.