Column: Without a foundation, visual cacophony prevails

Everyone wants AI-generated imagery these days. But hardly anyone has the foundation for it. These tools are like selling roof tiles to people who don’t have a house. Because before brands start thinking about generative AI, they need to answer an uncomfortable question: What actually defines our visual identity?

Image: Courtesy of

I have a problem, no, rather a need: it's like a compulsion to maintain my autonomy at certain moments. It's not about crucial things, but about small things in life, such as an IKEA shelf or a tomato sauce.

I can't bring myself to read the instructions at IKEA, it just doesn't work, and every time I reach the eighth screw and look at the crack in the wood, I think that it can't be right. The same thing happens to me regularly with tomato sauce, I've never followed a recipe before, it's unique every time, and unfortunately often bad, if I can tell from the looks on my children's faces as they push the Pasta Napoli to one side with that look of rejection and incomprehension.

At IKEA, I now usually have things delivered with an assembly service and when it comes to pasta, I now only use ready-made sauce.

Psychologically, this phenomenon is called reactance, an inner motivation to restore a threatened freedom. I'm not sure why I feel my freedom is under threat, but it might be interesting to investigate further. As long as it stays with such small things, the consequences are manageable. If I were a carpenter or chef, it would be a different story. The impact on business would be devastating.

In the corporate image world, however, reactance is the order of the day. I know this because I have often been involved. It wasn't my reactivity that was the problem, but the briefing specifications from the company itself. Because if you don't know exactly what you want even after reading it for the fourth time, and here I'm excluding obtuseness as a reason, the last option is to simply do what you think. Sometimes it worked, sometimes not. The feedback? Usually nothing ... silence. Or a "Mmh ... yes ... ok" without any further explanation. And how could they - if nobody knows what the brand is visually about, nobody can say whether the result is right.

A system that suited everyone

If we are honest: It suited everyone, except perhaps the brand managers. But you rarely got to see them during production. The agency, who preferred creative freedom to tight specifications. The photographer, who preferred to do his thing rather than work through checklists. The marketing team, who didn't have time for another basic document. Everyone benefited - well, almost everyone, except the brand.

I've seen a lot over the years and have already given the mood boards names. There was the "untraceable", which you simply couldn't find because it either didn't really exist or the client just didn't like it. Or the "Egalo", one page, four pictures and three adjectives. The "Penible" was particularly laborious, 47 pages with countless details but no clear line.

A study of 598 marketing managers shows the extent of this: less than half have a style guide at all, and of those who have one, only half actually use it. At the same time, 82 percent feel insecure when making visual decisions, which is a toxic combination for consistent brand management.

To make this clear once again: Images are not decoration, they are the container for brand codes and emotions. And yet, in reality, it is often pretended that the image is the icing on the cake of a good headline. Nice to have, but something is enough. The problem is that a lead agency used to ensure visual consistency. Those days are over. Today, five service providers and three internal teams work in parallel, all with the best of intentions, all "by feel". The result is visual cacophony.

The empty promise

Tool providers are now bursting into this situation with their promises. "Generate your visual world in seconds." "Brand-consistent visuals at the touch of a button." "AI that understands your brand." That sounds tempting. And it's dangerous nonsense. Not because the tools are bad. But because they presuppose a foundation that most companies have never laid. They sell roof tiles to people without a house.

Prompts, i.e. the text instructions to the AI, work better the more clearly the task is defined. A prompt that is not based on clear guidelines and simply somehow describes a brand ("authentic, emotional, credible") delivers exactly that: somehow results. Generic, interchangeable, ineffective. If you don't tell it precisely what your brand is all about, it will deliver the average of the internet. And the average is slop - or looks like the competition. The problem is not the AI. The problem is that we can't give it any useful instructions because we ourselves don't know what actually constitutes our visual world. We never wrote the instructions. And now we wonder why the tomato sauce doesn't taste good.

The inconvenient truth

Despite all the euphoria about the new possibilities of generative AI and its promise to revolutionize communication quickly, widely and cheaply: Before we discuss AI tools, we need to talk about foundations. About visual worlds worthy of the name. About guidelines that are more than three adjectives on a PowerPoint slide. About expertise that enables people to make good visual decisions - and recognize bad ones. The solution is neither complicated nor new. It just requires us to follow the sequence: First understand what makes your brand visual. Then develop guidelines that are precise enough for both humans and machines to work with. Then build up the expertise to assess the results. Internally. Not delegable. Generative AI is not the problem. It is the mirror that shows us what we have always failed to do: Treating visual worlds as what they are - a strategic tool.

For those who understand this, AI becomes an accelerator. For everyone else, it remains an expensive disappointment machine, like my tomato sauce without a recipe ... in most cases anyway.

The author

Per Kasch began on stage, then moved behind the camera and photographed international advertising campaigns and visual worlds for over 15 years. Today, with SightEffect, he advises companies such as Fraport, AXA, Sanitas and Radeberger on their visual strategy - in particular through AI-ready guidelines and measurable expertise.

More articles on the topic