The Ownership Problem

Understanding what talent is available has always been a consideration when selecting technology. Pick a route with a narrow pool and you either can’t get the people or they are very expensive.

AI is starting to change that picture.

The ability to build something is now available to everyone. AI agents can work across multiple frameworks, languages and platforms without needing months of onboarding. You can direct work at a level that would have required a team of specialists not that long ago. The expertise to execute is becoming less of a bottleneck, which means the old question of “can we hire for this?” is losing some of its weight.

That sounds like progress. And in many ways it is. But there is a flip side that not enough people are talking about.

Ownership used to be about the code. Could someone maintain it? Debug it? Keep it secure? Reasonable questions, and for a long time the right ones. But AI is increasingly capable of all of that. What matters now is the layer above the code. The decisions about what gets built, why, and to what standard. The ability to spot when something looks right but is not. The willingness to stop work that should never have started.

That is not a technical skillset. It is a leadership and judgement problem. And it is getting harder to solve, not easier, because most of the industry is still treating AI as a productivity tool rather than a workforce that needs managing.

When nobody genuinely owns what AI produces, you get bloated platforms full of features nobody can explain. You get technical debt that accumulates silently because the person directing the agents does not fully understand what they built. You get teams maintaining systems they did not design, burning out not from overwork but from the particular exhaustion of fixing things that should never have shipped. And you get digital waste - code, infrastructure, features - that serves no real purpose but still consumes energy, money and attention.

We talk a lot about sustainability in how we build digital products. Every unused feature, every redundant service, every system that exists because it was cheap to create rather than necessary to have, that all has a cost. It runs on servers. It uses energy. It adds complexity that slows down the work that actually matters. When you can build at AI speed without the governance to match, that waste multiplies fast.

The hiring question used to be about technical skills. Now it is about something harder to screen for. Can this person set a quality bar and hold it? Can they make the call to stop building something that should not exist? Do they know the difference between output and progress?

The part that should perhaps concern us is that if the next generation of technologists is no longer learning by writing the code themselves, how do we build a pipeline of people who can do this? That instinct for when something is off. That experience-led judgement. Those skills have always been developed through doing the work. If AI is doing the work, we need to think seriously about how we develop those capabilities through other means.

The ability to build is now within reach of almost anyone. That is genuinely exciting. But the ability to own what gets built, to make sure it serves the people it is meant for and does not just add to the noise, that is where the real value sits. The organisations that work this out now will have a significant advantage over those still catching up later.

If you are rethinking how your team is set up for this, we should talk. Learn more about our Digital Strategy and Delivery service.