AI has not changed what we make. It has changed how fast we can be wrong.

The honest position on AI in our industry sits somewhere between the two loud ones. It is not going to design your reveal. It has already quietly changed how a project gets from a conversation to a drawing.

Where it genuinely helps

Iteration count. In the first two weeks of a concept, the constraint has always been how many ideas you can visualise before someone has to commit. That number has gone up sharply. We can put twelve directions in front of a client where previously we would have offered three, and kill nine of them in a morning.

That matters because bad ideas are cheap to kill early and expensive to kill late. The value is not that the tools produce the answer. It is that they let you get to a no faster.

Then there is the unglamorous end. Site survey notes transcribed and searchable. Content variants for social cut without tying up an editor for two days. Documentation, schedules, risk assessments drafted rather than started from nothing. None of this is exciting. All of it takes hours out of a week.

A render nobody can rig is not a concept. It is a promise you have made on someone else’s behalf.

Where it does not

It does not do structural engineering. It does not do load calculations, wind loading on an open-fronted stand, or the mapping maths behind an anamorphic install. When we built the anamorphic LED cube for MG at Goodwood, the illusion held from every angle because someone worked out the geometry against real sightlines. There is no prompt for that.

It does not call a show. It does not know that the third camera position blocks the VIP walkway, or that the client CEO will not stand where the visual works.

The actual risk

The risk is not that AI takes creative jobs. It is that it makes it trivially easy to present an image of something that cannot be built, to a client who reasonably assumes it can. The gap between a beautiful frame and a buildable structure has never been wider or easier to hide.

Which is an argument for the same thing we have always argued for. The people imagining it should be sitting with the people who have to make it stand up, from the first week. That was true before any of this. It is more true now.