"AI can't do what we can do. At least not yet."
Visually seductive creative is dangerous.
A single render can do more damage to a project than almost any other device. The lure of a gossamer visual is a honey trap. You’ll fall for it, and worse still, your client will too. There is every chance it will hold you to ransom. And then it will break your heart.
This issue isn’t new. Managing creative expectations has always been a tricky path to navigate. For decades we have been presented with creative that is amazing. Concepts that look incredible. Many of these schemes are so extraordinary that they are close to impossible to deliver. Somehow, a hint at impossibility only adds to the allure.
AI isn’t helping. These diffusion models feed upon the crazy abandoned ideas of others. They take the most improbable elements from each and concentrate them into an intoxicating elixir of broken dreams. Curiously, this same elixir will likely win you the pitch.
So, can AI be reframed to become a more useful creative tool for our clients?
The answer is a resounding yes. The flip side is that it might not win you as many pitches. And here’s why.
AI can create superbly polished images. The quality of early-stage creative renders has never been as visually impressive. But AI can’t do what we can do. At least not yet. So, there is almost always a lack of real-world consideration to budgets, physics, and deadlines. What you see is rarely what you’re going to get.
Good design involves real world judgement; what to leave out, what to push, what to reconsider. This begins with design constraints. Think about weight, think about the weather, think about access, think about gravity, think about programme, think about safety. These things are never as exciting to creative teams as inventing new colours or manifesting a new emotion. But they matter and will endear you to your clients as the output you promise will be consistent with the output you deliver.
Therefore, the call to arms here isn’t to stop using AI to develop creative. These technologies are here to stay and there are thousands of other articles that present the case for and against these platforms.
But before falling in love with the render, refine the concept using these broad questions. Answering these yourself is fine, but it is sometimes worth prompting the AI engine to try answering them too.
- Where does the inspiration come from, has the engine simply plagiarised others IP?
- Is there precedent, has anyone ever done this before, and on this basis, is it truly creative?
- What might this thing be made from? And on this basis, will it come within budget?
- Is this achievable in the time available. Include for design development, procurement, fabrication, testing, transport, and installation?
- Can this thing actually be built? Where are the connections and the joint lines? Can we get it in the venue?
We salute those of you using AI, but like all good things, do it in moderation and never to the exclusion of other tools. Remember not to put your own judgement or sensibilities on hold; if the design looks impossible, it probably is. But perhaps most importantly, don’t forget that most considered creative doesn’t always win the pitch. Clients will always love the impossible.
Whilst we can’t promise your win rate will go up. We’re pretty sure what you promise at pitch will be deliverable.
And in terms of building a credible creative business, well, that’s priceless.
