Design & AI

What if brands stopped confusing consistency with uniformity?

Consistent brand experiences remain essential. But they are not created simply by making every touchpoint look as similar as possible. This understanding comes from a more monologic brand world in which communication was developed, rolled out, and controlled centrally. Generative AI intensifies the question. Images, texts, snippets, and countless variations are created in seconds for different target groups, markets, and channels. Those who try to control this diversity solely through more and more rules ultimately risk scaling one thing above all: uniformity and monotony.

The more interesting question is therefore: How can brands create coherence in order to enable consistent brand experiences even under these conditions?

Coherence does not mean that everything looks the same. Rather, it means that different forms of expression are recognizably derived from the same brand identity. A brand can, and must, evolve, respond to contexts, and be able to surprise. Without becoming arbitrary. For brand design, this is more than a semantic nuance. It shifts the task: from the consistency of individual outputs to the coherence of the system behind them.

AI is the stress test for the brand

Generative AI does not create this problem. It reveals what was already there. A brand whose identity essentially consists of a logo, color values, typography, and a brand manual runs into difficulties as soon as content is no longer produced exclusively by a well-coordinated design team. Experienced designers interpret a brand. They know when a deviation is still appropriate, which elements need to work together, and when something may be formally correct but still not “on brand.”

AI can reproduce rules, process references, and imitate stylistic characteristics. What it does not automatically understand is the distinctive interplay that makes a brand unmistakable. The real danger of AI is therefore not inconsistency. It is averageness.

When the same models process similar aesthetic references and more and more brands use the same generative possibilities, distinction becomes a scarce resource.

Real Distinctive Assets are becoming more important, not less

Distinctive Brand Assets are therefore by no means becoming less important. On the contrary. Color, typography, shapes, imagery, sound, or motion are becoming important anchors in an increasingly variable content landscape. But we must not understand them as a collection of isolated constants. A color alone does not make a brand. Nor does a logo in the corner of an AI-generated visual.

What matters is the interplay: Which assets appear together? What hierarchies are there? What can change, and what cannot? And how far can a design element be varied before it loses its brand character?

This is precisely where the systemic task of brand design begins: Distinctive Assets must be designed systemically.

The brand as a system

The traditional brand manual primarily documents what a brand should look like and how it should be applied correctly. In a generative world, we also need to define how it can behave. This does not make guidelines and governance obsolete. On the contrary: the more people, systems, and technologies work with a brand, the more important clear governance becomes. But this governance cannot consist of formulating increasingly detailed rules for more and more use cases.

We therefore need to think more in terms of principles in addition to rules. Rules remain indispensable where assets need to be protected, standards adhered to, or clear boundaries drawn. But they scale poorly when they are expected to anticipate every conceivable application. Principles take over where interpretation begins: they provide guidance for decisions without fully predetermining their outcome.

The crucial question, therefore, is not rules or principles. Rather: What is non-negotiable? Where are clear boundaries needed? What is allowed to vary? And which principles ensure that this variation still results in something characteristic of the brand?

This is how systems emerge in which constants, variables, relationships, and degrees of freedom are deliberately defined. Systems that provide guidance without anticipating every possible output.

Initial developments show where this is heading. Coca-Cola, for example, is working together with Adobe on “Project Fizzion.” The AI-powered design system translates traditional brand guidelines into intelligent, adaptive assets and aims to make brand-compliant design more scalable across brands and markets.

A look at Spotify is also revealing. With “Encore,” the brand has deliberately built its design system as a decentralized “system of systems.” At its core is a shared foundation of color, typography, motion, spacing, writing, accessibility, and design tokens, in other words, the core that creates recognizability across products and platforms. Different systems for web and mobile, as well as local design systems for specific applications, build on this foundation. Spotify for Artists, for example, can require its own navigation patterns or tables without losing its connection to the overarching system.

The principle behind this is remarkably simple: Not every application has to be the same, but every application must remain traceable to the same core. Shared foundations ensure coherence, while decentralized systems enable variation where the product, platform, or target group requires it.

Coca-Cola and Spotify approach this from different directions, but point to the same development: brand identity is increasingly becoming part of the infrastructure in which content and experiences are created. This also changes the role of design systems. Good systems do not just organize components and their correct application. They define constants, principles, and degrees of freedom.

A good system, therefore, does not just describe what a brand looks like. It also defines how far it can go.

From controlling output to steering the system

This is precisely where the decisive shift in perspective lies. Given the amount of content that can be created in the future, it will become increasingly challenging to centrally review and manage every individual output. The design task therefore shifts further upstream: we need to design the conditions under which brand-specific output can emerge in the first place.

Governance does not disappear as a result. It shifts from controlling as many individual cases as possible to systematically steering the conditions under which the brand is created. This also increasingly means moving from appearance alone to meaning: What defines this brand at its core, and how can that core be translated into ever-new forms of expression?

The perspective shifts from the Brand Manual towards a BrandOS. Not as an even more extensive guideline, nor as a replacement for governance, but as an infrastructure in which Distinctive Assets, rules, principles, and degrees of freedom are anchored in such a way that people and machines can work with them. The decisive quality of such a system lies in enabling variation without losing coherence.

In the future, we should therefore assess brand consistency in a more differentiated way: not solely according to whether every touchpoint follows the same design rules, but according to whether a consistent brand experience emerges across those touchpoints. Whether we recognize the brand, especially when it does something we have not seen before.

AI is the toughest stress test for this. And that is precisely why it is a major opportunity for brand design: it forces us not merely to design brand identities, but the systems behind them.