I'm writing a series of articles on the changing shape of business, brand and humanity in the age of AI. I'm sceptical about anyone who says they have all the answers here — so think of this as thinking in public.
Everything is seemingly changing around us, at breakneck speed. AI is killing whole verticals on a daily basis if you believe our feeds. Design is dead. Figma is dead. SaaS is cooked. Or if your echo chamber sits at the other pole, AI is a bubble and it's all a sham. All the hyperbole is missing what's really changing.
The most interesting thing with AI for me right now, and to an extent blockchain before it, is the questions it forces us to answer about what is valuable to us as humans. From a market perspective yes, but on a deeper level too. What does it mean to be human? What gives us purpose? What does ownership and intellectual property really mean?
This is where the recalibration is happening. And as I think about what comes next for Airborne as an agency, for us as creatives and builders and particularly for brand as a discipline and business asset, it's through this lens that I'm trying to understand it.
Both ends of the bridge are changing
Brand has always been the humanisation of non-human things. A company is an abstraction — a legal entity, a strategy, a payroll, a balance sheet, some lines of code. Brand is the work of turning that abstraction into something a human can recognise, trust and feel something about. Personality wrapped around organisation. That's all it's ever been.
Which is why this AI moment is bigger than a tooling question. Brand has always had one job: connecting non-human entities to humans. The non-human end of that bridge is changing fast — that's the part everyone's talking about. The human end is changing too, and that part is barely in the conversation. What people value, what they trust, what ownership and attention and authenticity actually mean — all of it is moving.
That makes the failure mode obvious. It isn't AI replacing brand designers. It's overlaying old thinking onto new tools. Producing more polished, more professional, more abundant brand work — for a world that's stopped caring about the things that brand work used to be optimised to do.
This is what I've started calling programmable brand (a working title). Yes, it's about adapting to an AI production environment — encoding brand intelligence so AI can carry it forward at scale. But that's the easier half. The harder half is thinking carefully about what humans in this new world understand about themselves, and how brands operate inside that understanding.
Polished is (more or less) free now
Polished execution is almost free now. AI tools can produce a passable logo in seconds. Web copy that on the surface reads as ‘good’ can be smashed out in one prompt. Photography that looks professional can be summoned into existence for a few tokens. A lot of brand work shipping today already involves AI somewhere in the process. Most of it lands at ‘OK.’ Not bad. Just OK, in roughly the same way as everyone else's OK.
Which is exactly why the humanity captured in a brand — character, voice, the specific moves that make it itself — matters WAY more now, not less. Brand work for the last fifty years has run on one fairly universal model. A team digs into a business and synthesises its character into a personality. The end product is a set of guidelines designed to point human practitioners at work that feels like the brand and away from work that doesn't.
What changes now isn't the guidelines half. What changes is the practitioner half — because AI is starting to do parts of the practitioner's job. Not all of it. Not all of the seriously creative parts. But enough that a real question opens up. How do you keep what's valuable about the practitioner in the work, when AI is producing more and more of the output?
What are we actually encoding?
If the job is to encode judgement at scale — what are we actually encoding? The honest answer is the qualities that have always done the most work in brand, and have always been the hardest to write down. There's a long-running myth that a sufficiently detailed guidelines document can close the gap between a great designer and a competent one. It doesn't. The gap has always been made of things you can't fully write down.
Programmable brand as I'm starting to think about it is the project of approximating that combined ear inside a system. Not replacing the practitioner. Pointing at what they do, with enough specificity that an AI working from the intelligence in front of it produces work closer to what they'd have made than to the smoothed centre an unguided model defaults to.
Visual is the harder problem
Tone of voice has been encodable, in detail, for years. Visual identity is harder. The harder question is art direction at scale — photography choices, illustration commissions, the editorial sensibility that decides what gets shown and from what angle. Those are the parts that have always sat with the senior creative practitioner. And the parts current AI tooling handles least well.
Two versions of programmable brand are visible from here. The version worth building keeps the practitioner as creative director, augmented by a system that produces faster and against tighter brand intelligence than they could deliver alone. The version worth avoiding puts the practitioner in the role of inspector, signing off whatever the system happened to produce.
What keeps proving true
The artefact looks more like editorial than like a design system. Less a component library, more an annotated reference set — examples of what the brand is doing well, with notes on why, written so an AI can actually read them. The work is curation and explanation, not specification.
The brand intelligence has to be alive, not finished. A static guidelines PDF is the wrong shape for this. The system has to know what to do when it doesn't know what to do. Most current AI tooling guesses and ships. The version worth building defers, flags, or hands the moment back to a human — visibly, not silently.
As a studio, we have been working with AI in our process for several years. We have early experiments running on what programmable brand looks like in practice. We don't have it nailed. What we have is a thesis we're testing in the actual work, with actual clients, in real time — and a few observations which feel sturdy enough to share.
