AI Is Leverage, Not Magic
It is hard to be a business leader right now.
Every week brings a new AI announcement. A new tool that promises to transform your operations. A new headline about entire teams being replaced. Your board is asking questions. Your competitors are making noise. The pressure to move, to adopt, to not fall behind, is relentless.
I get it. I have spent thirty years building software, and even I find it difficult to separate signal from noise. So if you are a CEO or a board member trying to make sense of all this, I want to offer something that might actually help: a different way to think about what AI is.
AI Is Not a Productivity Upgrade. It Is Leverage.
There is an important distinction that gets lost in the hype. Most people talk about AI as though it is a better tool. A faster way to do the same things. That framing is wrong, and it leads to bad decisions.
AI is leverage. And leverage magnifies whatever foundations you already have.
If your architecture is solid, your governance clear, and your technical leadership strong, AI will accelerate you. It will help good teams build good systems faster. That part is real.
But if your foundations are weak? If your architecture is fragile, your processes unclear, or your technical leadership thin? AI will help you build fragile systems faster. It will generate production-grade code on top of shaky ground, and everything will look fine right up until it doesn't.
Think of it like installing a high-performance engine into a car with poor steering and weak brakes. You will not get places faster. You will just crash harder.
What AI Cannot Do
Here is something the AI evangelists will not tell you. AI is very good at answering the questions you ask. It is completely useless at surfacing the questions you do not know to ask.
That distinction matters enormously.
The hard part of building software has never been writing the code. It has always been understanding the problem well enough to know what to build. That requires experience. It requires judgement. It requires someone who has been through enough projects to recognise the patterns, the risks, the things that look fine on paper but fall apart in practice.
AI cannot do that. Not yet. Maybe not ever.
The people telling you that AI can do everything tend to be the people selling AI tools. Listen instead to the people who have actually built and shipped real products over real timescales. They will give you a more grounded picture.
This Is a Leadership Conversation
Over the next six to twelve months, a gap is going to open up between two kinds of organisations.
The first kind will rethink how software is led and governed. They will invest in technical leadership, clean up their foundations, and then use AI to accelerate from a position of strength.
The second kind will simply adopt tools. They will plug AI into their existing processes, celebrate the speed improvements, and watch their software moats dry up as competitors do the same thing. The strategic advantage they thought they had will erode because AI makes it trivially easy to replicate what used to be hard.
This is not a tooling conversation. It is a leadership one. The organisations that get this right will not be the ones that adopted AI first. They will be the ones that understood what AI actually is, and prepared their foundations before accelerating.
The Right Question
If you are leading an organisation right now, stop asking "should we use AI?" You should. Everyone will.
The better question is: do we have the foundations to use it well? Do we have the technical leadership to guide it? Do we have the governance to keep it on the road?
The leaders who navigate this well will not be the ones who moved fastest. They will be the ones who asked better questions first.
I have spent two decades building products, leading teams, and working out what actually matters when software meets reality. If you want a clear, commercially grounded conversation about what AI means for your business specifically, that is exactly the kind of work I do.