AI driver or AI driven

November 14, 2026

AI driven work in IT makes a lot of really hard things very easy. It can extract huge amounts of data, convert it into a visual representation that would takes days to develop and spit it out in a matter of minutes.

It can have people with surface-level knowledge of an area or problem spit out double digit design documents in a matter of minutes.

It can create beautiful prototypes from really rough designs compared with loosely described use cases - also just in a matter of minutes.

It can do just about anything in just about no time.

But is it actually making us achieve more valuable work?

Who’s in the driver seat?

AI driven development and AI driven “just-about-anything” has made it very easy to spit out what only months or years ago would have taken a whole team months or years to accomplish in mere minutes. It can make people output seem “super-human” - which in many cases it is.

But does it actually make us achieve more of what matters?

I have started to seriously doubt it.

I am sure that AI will only be getting better and better and that the future looks very fragile for all us meat-proxies making a living in front of computers.

Are we being driven round in circles

But currently I believe that we being driven around by AI more than we are in control of where AI is driving.

I have seen and spent so much time on output either in form of documents, prototypes and visualizations that looks and are extremely impressive. They take on a lot of time looking through a lot of time impressing everyone and then ends up being the solution that goes around looking for a problem.

Meaning that I see close to half of what AI is used for is giving extremely impressive output that should never have been created in the first place. Long design documents that someone has to sit down read through, come up with an opinion about and in some cases spent exhorbitant amount of time defending against it as being the right solution.

Or data visualizations that are pretty to look at, can pull in all sorts of data from different sources and link it together, swallow up clever peoples minds on how to enrich even further, when in reality what it shows could have been derived and understood well enough from existing sources and now has ended up being a source of truth that again requires too much time being spent on defending or trying to enrich even further - when the task of creating it in the first place was never a task on anyones backlog.

It is not what can be created it is what should

What many people recognized and has known for a very long time is that IT in general and development especially it was never a question of what can be created or spent time on. It was always a question of what should.

An under-funded IT department paired with a over-funded business department shows this brilliantly. It always ends up with a huge mess of applications - all solving actual business needs - but each being developed by either semi-capable business-developers or external consultancies solving the assignment they are presented with to the specification required.

But no one ever looks at the big picture trying to figure out what are things we not only can solve but things we should solve. What is our core business and how do we build, protect and evolve that in a sustainable way while outsourcing or buying things that are outside of that.

Maybe sometime in the future it won’t matter and every business is controlled and maintained entirely by AI, but where we stand currently we should seriously consider whether we are in the driver seat or whether we are being driven around.

AI is now good enough that it can solve almost any task we throw at it - also the ones that should never have been solved in the first place. Nothing is free there is always an opportunity cost.