For most of my career, a brand kit was a finished thing you could hold. It was the logo in every format someone might need, the hex values for the palette, the type and how to set it, the slide templates, and the component library that kept everything consistent. You delivered it, walked the team through it once, and trusted people to reach for the right file. The brand lived in a document, and the document sat still.
That version of the job is gone. The kit still matters, but it’s become the foundation of something much larger, because the brand has picked up a second audience. People aren’t the only ones reading your guidelines anymore. Language models now generate brand-bearing work directly, and they read your brand very differently than a person does.
Your brand has a second audience
When a person reads “use Navy for headers and keep generous spacing,” they fill in the rest with judgment. They know what generous means in this context, they notice when a layout feels crowded, and they make a hundred small calls the guideline never spelled out. But if you hand that same guideline to a model, it fills those gaps with the most plausible guess, and the most plausible guess is usually a few degrees off. It’ll pick a navy that’s “close.” It’ll set spacing that’s “fine.” The result LOOKS right at a glance and is quietly wrong in a dozen small ways, which is the most dangerous kind of wrong, because nobody catches it.
Handing your brand to a model is a different act than handing it to a designer. A designer needs principles and room to interpret them. A model needs the brand translated into constraints it can’t casually drift from: the exact tokens, the rules it can’t break, the patterns it should reach for. So the work goes beyond describing what the brand looks like. It means defining the decisions a machine can make on its own, the ones it can’t, and where a human still needs to step in.
The surfaces you never used to worry about
The old brand risks were predictable, because brand-bearing work came from a small number of places. A designer made the deck. A designer made the ad. Anything with the logo on it passed through someone who knew the system. That funnel is gone.
Picture a sales rep who needs a LinkedIn graphic to announce a partnership. They ask a model for it and get something back in seconds. The logo is right and the blue is close. But the headline is the line every other company in the category posts, the layout ignores the spacing system, and the image treatment could belong to anyone. Nobody flags it, because nothing looks broken… the brand just became a little less itself.
That’s the new shape of the risk. The work looks finished, and looking finished isn’t the same as being on brand. In the old world you worried about the handful of things you shipped. Now you have to account for the thousands of things other people will generate, in tools you don’t control, most of them never realizing they made a brand decision at all.
The work doesn’t end at handoff
Building the system is necessary, but it isn’t the whole job. A good designer sits with the thing after it ships and pays attention to how the organization actually uses it. That means teaching, and the lesson that matters most is perceptual: how to see when the brand is off.
Most people can’t tell when something is eight percent wrong. If you show them a slightly mistinted logo or a header in the wrong weight, they’ll feel like it’s fine, because nothing is obviously broken. A designer’s eye is trained to catch exactly that. Part of the job now is transferring some of that perception to everyone else: showing them what right looks like, how to notice drift, and what to do when it happens. The designer who can build that instinct in a team is worth far more than the one who only makes good files.
The same curiosity has to point back at the system. When a bad output slips through, fixing that one artifact is the small part. The real question is why the system let it through, and how to tighten the rule so the next thousand come out cleaner. That takes a different instinct than visual craft: a willingness to open the system, find the lane that was too wide, and narrow it.
Ship at 90, stop chasing 100
There’s an old instinct in design to hold work back until it’s perfect. You revise, you refine, you nudge the kerning one more time (just one more), and only then does it leave the building. But in this new world, that instinct quietly turns into a liability, for two reasons.
The first is volume. When work that carries the brand is generated across the company every day, you can’t hand-polish each piece to completion. There isn’t enough of you. The second is that completion is no longer a real state. The brand now lives across thousands of outputs that keep changing, so there’s no single finished artifact to perfect.
The better goal is to reliably hit around 90% on everything, fast… and let it go. This isn’t a license to care less. The care moves into a system that makes the average output safe and recognizable on its own, instead of into perfecting one artifact while a hundred others go out unattended. Without that system you get a scatter: some outputs land in the 90s, others fall into the 70s, and the 70s are what damage the brand. Stop polishing one piece to perfect. Raise the floor under all of them.
The hard part is the system
The brand kit hasn’t stopped mattering. You can’t build any of this without the logo, the palette, the type, and the components, and getting those right is still the foundation everything else stands on. What’s changed is that the kit is really just the floor of a much taller structure now. The job has grown from making the assets to designing how a whole system, people and machines together, reproduces the brand reliably and at speed, across surfaces you’ll never personally touch. And it has grown to include the people around it, teaching them to use the system, to see when it slips, and to care that it does. This doesn’t make designers less necessary. It makes their judgment harder to contain inside the old shape of the job.
The brands that hold together over the next few years will be the ones run as living systems with two kinds of users, not the ones with the prettiest guidelines. The brand kit was the easy part. The real work is building the system so that the easy thing to make is also the on-brand thing to make.