The Brand That Thinks
Part one of the Terra Incognita series.
I wrote recently about the six loops - the repeating scripts that have kept the AI and creativity conversation stuck for the past few years. Fear, hype, efficiency, exceptionalism, and the revolving door of “is AI coming for our jobs” panel discussions that leave everyone nodding and nothing moving.
This series is about what’s on the other side of those loops.
I’ve spent the past six months developing a POV on what I’ve been calling Terra Incognita - coordinates for genuinely new creative and strategic practice. Not productivity play. The stuff beyond it. What becomes possible when you stop treating AI as a production tool and start treating it as something to think with.
New coordinates. Already emerging in scattered, unnamed ways across different industries and practices. Each one represents a form of creative work that wasn’t possible before - and each one demands more human skill, not less.
This series takes each one in turn. So, let’s begin.
Somewhere in your organisation right now, there’s a PDF. It contains your brand guidelines - typography, colour palette, tone of voice principles, the do’s and don’ts. It was comprehensive when it was created. Addressed the platforms that existed at the time, the competitive landscape as it was understood, the cultural context of that particular moment.
It’s been gathering dust ever since.
Not literally, obviously. People still open it. Still reference it in briefs and creative reviews. But there’s a growing gap between what that document says and what the brand actually does. Between the brand as documented and the brand as experienced. That gap widens every day. The world moved on. The document didn’t.
I’ve worked with enough brands at this point to know this isn’t the exception. It’s the norm.
A brand today might need to show up in a TikTok comment, an AI-generated search result, a crisis response drafted at midnight, a customer service chatbot, a partnership announcement, a meme they didn’t make, a platform that launched six months after the guidelines were signed off. The surface area is essentially infinite. And the mechanism for maintaining coherence across all of it? A PDF and the hope that someone remembers what “on brand” feels like.
We’ve been designing brands like buildings. Fixed structures. Blueprints handed over, then constructed, then maintained. The assumption being that if you get the architecture right upfront, it will stand.
Brands don’t exist like buildings. Not anymore.
Let’s be specific about what’s actually failing.
The first layer is obvious. Documentation that captures a moment in time and then freezes. Brand guidelines written during a rebrand three years ago. Strategy decks that generated a lot of excitement in the room, then got shelved after the presentation. Principles abstract enough to mean anything and specific enough to mean nothing. “Be bold but approachable.” “Lead with purpose.” “Champion authenticity.” Fine. What does that actually mean when you’re being dragged into a political conversation you didn’t choose? What does “bold but approachable” look like on a platform whose entire vernacular shifted last quarter?
The second layer is harder. It’s the stuff that was never documented in the first place.
Every brand has an invisible architecture of judgment that lives exclusively in people’s heads. The marketing director who’s been there fifteen years and just knows. She can look at a piece of creative and tell you in seconds whether it’s right. Not checking against a guideline. She’s internalised decades of decisions, debates, rejected work, approved work, crisis responses, near-misses, and cultural shifts. She holds the logic underneath everything.
When she leaves, that logic walks out the door with her.
I keep coming back to this. Think about what actually gets documented in the life of a brand. The brief gets documented. The final work gets documented. Sometimes the rationale for the chosen direction gets documented, though usually in a sanitised, post-rationalised form that bears little resemblance to the actual messy process. But the reasoning that connects them - the judgment calls, the heated debates about whether this crosses the line, the work that got killed and why, the instinct that said “not this, not yet,” the understanding of which tensions the brand holds and which it doesn’t - all of that evaporates.
Every single time.
And the problem is accelerating. People change roles faster than ever. Average tenure in marketing keeps shrinking. The agency of record model that once provided some institutional continuity has fractured into project-based relationships where nobody stays long enough to hold the thread. Every new team, every new agency partner, every new hire begins from a position of guessing. They read the guidelines. Look at recent work. Do their best interpretation. And everyone’s interpretation is slightly different, because the document was never the real source of coherence. The real source of coherence was always the people.
And the people keep leaving.
This is brand decay. A slow, quiet erosion. The brand drifts. Not in any single catastrophic way, but in a thousand small inconsistencies that accumulate. Voice gets a little flatter here, a little more corporate there. Decisions that would have been caught by someone with twenty years of context slip through. The gap between what the brand is supposed to be and what it actually is grows so gradually that nobody notices until it’s already happened.
I’ve seen this at close range, across organisations of every size. The knowledge drain is staggering. One estimate puts it at tens of millions in annual productivity loss for a large enterprise - and that’s just the general organisational figure. For brands, where what you’re losing isn’t process knowledge you can stick in a handbook but taste and judgment and accumulated intelligence about how this specific entity navigates the world - the cost is incalculable, mostly because nobody’s measuring it. Nobody even knows what they’ve lost until the next crisis hits and the person who would have known how to handle it left eighteen months ago.
So the question becomes: how do you maintain coherence when the knowledge that holds it together keeps walking out the door?
Game designers figured this out years ago.
Not the brand version of it. But the structural equivalent.
No Man’s Sky doesn’t design 18 quintillion planets individually. It designs the system that generates them. The logic is fixed, the outputs are infinite. Every planet is different. But every planet is recognisably part of the same universe. Not because someone hand-checked each one against a set of guidelines. The generative logic itself does that work.
Minecraft doesn’t hand you a perfectly formed world. It hands you the rules that create worlds. Consistency without repetition. Identity without sameness.
This is what procedural generation does. You don’t design the things. You design the system that produces the things. The system holds the grammar. The grammar produces outputs that are always novel and always true to the whole. The underlying logic is sound.
What if we designed brands the same way?
Not a set of rules that humans check outputs against. A system that understands how the brand thinks - its logic, its grammar, its instincts, its boundaries - and can generate responses to contexts that haven’t been imagined yet. Holding the intelligence deeply enough to create, to evolve.
I want to be precise about what I’m arguing for here, because there’s already a partial version of this that exists and it’s not what I mean.
Coca-Cola and Diageo have trained models on decades of visual assets. Feed the archive in - every campaign, every design decision, every asset - and the model learns the visual language. It generates new work that feels like the brand without copying what came before. Same palette, same energy, same logic, completely new outputs. Faster. Cheaper. More consistent.
That’s a generative asset engine. It’s a production tool. It makes the design process more efficient, and it’s genuinely impressive at that. But it’s still the efficiency play. You’re still making stuff. You’re just making it faster.
That is not what I’m talking about.
What I’m talking about is something fundamentally different in kind, not just in degree. Not a system that produces brand assets. A system that holds how the brand thinks. Not the visual layer. The cognitive one.
How does this brand actually speak? Not the tone of voice document that says “warm, witty, confident.” How does it speak when it’s apologising? How does it argue? Where does it use humour and where does it pull back? What’s the difference between how it talks to a twenty-year-old on TikTok and how it speaks to a concerned parent in a customer service thread?
What tensions does this brand hold? Most interesting brands live inside contradictions they never fully resolve. Premium but accessible. Heritage but modern. Global but local. How do you talk about sustainability when your business model depends on growth? How do you champion diversity when your leadership doesn’t reflect it? How do you navigate a political moment when half your audience wants you to speak up and the other half wants you to shut up?
How does this brand behave when the context shifts beneath it? A cultural moment lands and suddenly the campaign you’re about to launch reads very differently. A competitor makes a bold move and your positioning feels a bit flat. A partner does something controversial and your audience wants to know where you stand. These situations happen constantly. And the response is never in the guidelines.
What are the things this brand won’t do? Every strong brand has territory it actively avoids. Some won't go near self-deprecation. Some won’t reference competitors. Some won’t touch political topics at all. These boundaries are rarely written down. They emerge from practice - discovered in the act of saying no. And the reasoning behind each ‘no’ is almost never recorded.
All of this - voice, tensions, how the brand navigates pressure, the invisible boundaries - is what I mean when I say how a brand thinks. And almost none of it exists anywhere. Some of it lives in the heads of people who’ve been around long enough. Most of it doesn’t even live there. It happened - in a meeting, a debate, a split-second judgment call - and then it was gone. Done and forgotten. The knowledge wasn’t lost, so much as it was never held in the first place.
So what does it actually look like to encode not just what a brand looks like, but how it thinks?
You start by feeding it everything. And by everything, I mean the stuff that has been strategically invisible until now.
Strategy decks, obviously. But also the debates that led to them. The presentation where the CMO pushed back and the direction changed. The three options that were rejected and why. The work that got killed - not just the final “no,” but the conversation about why it wasn’t right. Crisis responses, including the first drafts that got softened and the arguments about how far to go. The emails where someone said “this doesn’t feel like us” and tried to articulate what they meant.
You’re not creating a reference library. You’re training a system to understand how this brand makes decisions. What it leans toward. What it avoids. Where it takes risk. Where it doesn’t. The logic underneath the outputs.
I think of it as the difference between giving someone a recipe and giving them twenty years in a kitchen. The recipe tells you what to make. Twenty years in the kitchen teaches you how to improvise, what to taste for, when to trust your gut, how to save a dish that’s gone wrong. You can’t really write that down in a document. But you can - increasingly - encode it in a system that learns from the pattern of decisions over time.
There’s already a working example of this, and you’re probably interacting with it. Claude - the AI I use as my primary thinking partner - doesn’t check its responses against a rulebook. It holds what Anthropic calls a constitution - values, principles, and ways of reasoning that allow it to navigate millions of conversations nobody specifically prepared it for. It makes judgment calls. It holds tensions. It responds to contexts its creators couldn’t have anticipated, with a consistency that comes from deeply encoded logic rather than surface-level rules.
That’s the model. Not a perfect analogy - a brand isn’t an AI and an AI isn’t a brand. But the principle is the same. You encode how something thinks, and it can meet the world on its own terms.
And then the questions you can ask change completely.
Not: “Does this match the brand guidelines?”
But: “We’ve spent three years positioning ourselves around sustainability, but our supply chain still isn’t where it needs to be. A journalist is asking questions. How do we hold the tension between where we are and where we said we’d be - without sounding like we’re making excuses and without abandoning the position entirely?”
Or: “Our audience is splitting. The people who grew up with us want one thing. The people discovering us now want something different. Walk me through the ways this brand has navigated identity tensions before and what that suggests about how we hold both without losing either.”
Or: “We’ve been playing it safe for two years and the brand is going flat. Show me what it looks like if we push into more uncomfortable territory - and show me where the line is, based on every decision we’ve made about what we will and won’t do.”
These aren’t prompts. They’re the start of a dialogue. The system pushes back. You push back on the pushback. The thinking develops through the exchange - getting sharper, more nuanced, more honest about the trade-offs. Like the best strategist you’ve ever worked with, except one who holds the full weight of everything this brand has ever done, decided, debated, and rejected.
The immediate benefits are the obvious ones. Faster response times, more consistent output, less knowledge haemorrhaging through the revolving door. Efficiency gains. Fine. But efficiency is where every AI conversation goes to die, and I’m not interested in it here.
The interesting part is what changes about you. About what it means to be a strategist, a creative director, a brand thinker - someone who makes sense of systems and culture and markets and shifts for a living. And about the relationship you have with the brand itself.
Right now, a huge amount of brand work is custodial. Policing. Checking. Maintaining. Making sure things don’t drift too far from a document that was already out of date when it was signed off. The most experienced people in the building spend their time on quality control - catching deviations, correcting course, saying “not quite” in review meetings. It’s defensive work. Important, but defensive. And it wastes the most valuable thing those people have - their judgment, their taste, their ability to see what the brand could become rather than just protecting what it’s been.
When the brand can think, that custodial work dissolves. And what opens up is genuinely exciting.
You become an architect of intelligence. Not designing assets or campaigns but designing the system itself - deciding what gets encoded, how the brand’s logic should evolve, what new tensions need to be incorporated as the world shifts. You’re shaping how this brand thinks, not just what it produces. That’s a far more interesting job, IMO.
You become a philosopher of the brand. And I mean that seriously. When you can ask the system “show me how our relationship with sustainability has evolved over the last five years based on every decision we’ve made,” you start seeing patterns in your own thinking that were invisible before. Contradictions you didn’t know you were holding. Drift you didn’t notice. Evolution you hadn’t articulated. You develop a deeper, more honest understanding of what this brand actually is. And from that understanding, you can make genuinely strategic decisions about where it goes next.
You become an orchestrator of dialogue. The system pushes back. It offers perspectives you hadn’t considered, draws on history you’d forgotten, surfaces tensions you’d been avoiding. Your job isn’t to accept what it generates. It’s to argue with it, steer it, challenge it, develop your thinking through the exchange. The brand’s intelligence and your intelligence sharpen each other. That’s compounding in both directions.
And you become a builder of something that lasts. Every decision you make, every debate you have, every judgment call - it gets encoded. It accumulates. The brand gets smarter. When a new team member joins, they inherit not a PDF but a living intelligence they can immediately work with, talk to, learn from. The brand remembers what everyone before them built.
That’s a different legacy than a deck that gets shelved.
And the brand itself becomes something it couldn’t be before. It can hold more complexity without flattening itself. It can evolve continuously instead of lurching through expensive rebrands every few years. It can be different in different contexts while remaining deeply itself.
It compounds. That’s the word I keep coming back to. Every interaction makes it smarter. Every decision adds to its understanding. Every tension navigated enriches how it navigates the next one. The brand accumulates intelligence, not assets.
I’d be lying if I said this was clean or easy.
The first tension is political. In most organisations, brand knowledge is power. The people who “know” how the brand works, who hold the institutional memory, who can make the judgment calls - they have organisational influence precisely because that knowledge lives in them. Encoding it in a system redistributes that power. It democratises something that was a source of individual authority. Not everyone is going to welcome that, and pretending otherwise is just naive.
The second tension is about what gets lost. Judgment isn’t purely rational. It’s embodied. Intuitive. Sometimes contradictory. The brand director who “just knows” is drawing on something she probably can’t fully articulate - pattern, instinct, a kind of aesthetic intelligence that comes from being a person who exists in culture. Some of that transfers into a system. Some of it - the irreducibly human part - doesn’t. The system will be powerful. It won’t be the same thing. We need to be honest about what it gains and what it loses.
The third is dependency. Any system you build, you become dependent on. Technology changes. Vendor relationships shift. Models need retraining. You’re trading one form of fragility - knowledge trapped in people’s heads - for another - knowledge trapped in a system. The second is more durable. It’s not invulnerable.
And the fourth is about honesty. For this to work, you have to actually put the real stuff in. Not the heavily edited version. Not the version that got past legal. The actual debates. The actual reasoning. The actual messy and unresolved stuff the brand holds, including the ones nobody’s comfortable saying out loud. Most organisations aren’t built for that level of internal truth-telling. The brand book is, among other things, a document of aspiration - here is who we want to be. A brand that thinks requires something harder - here is who we actually are, including the parts we’re still figuring out.
The simplest way to put it is this.
We’ve been treating brands as static objects. Fixed things. Documents to be maintained, guidelines to be policed, assets to be managed. Andy Pearson, the creative director behind Liquid Death, said something last year (or the year before?) that stuck with me - that brands need to behave more like a character in a writers’ room. A living entity with a personality, a history, opinions, edges. Something that evolves through the writing, not something that gets checked against a brief.
He’s right. But I’d push it further. A character in a writers’ room still depends on the writers. What happens when the brand itself can hold the room? When it carries its own history, its own logic, its own sense of what it would and wouldn’t do - and you’re in dialogue with that?
That’s a brand that thinks.
Which doesn’t mean you hand the writers’ room over to it. A brand thinking in dialogue with itself and nobody else would end up going in circles - a snake eating its own tail. Same way AI trained on synthetic data collapses into noise. Nobody wants that. The human stays in the room. You still set the direction, still make the creative calls, still feed in the new thinking, the cultural shifts, the provocations the brand couldn’t generate on its own. But instead of starting every conversation from scratch, you’re starting from somewhere rich. Somewhere that already holds the complexity. And you’re steering it, shaping it, arguing with it - not policing it.
A document you check against versus a mind you think with.
The brands that work this out will be unrecognisable from the ones that don’t. Less for what they produce than for what they are. Living things that evolve, learn, hold contradiction, meet the world as it comes. Entities that get more interesting over time, not less.
The brands that don’t will keep updating PDFs. Keep losing their best people’s thinking to the revolving door. Keep flattening themselves to maintain a consistency nobody asked for. Keep arriving late to every conversation that matters.
The gap between these two kinds of brand is about to become unbridgeable.
And most of the industry hasn’t even noticed it’s opening.



Love this Zoe and its exactly aligned to a lot of what I've been thinking about. I was inspired by Arie De Geuss The Living Company and Wally Olin's The Corporate Personality- old books now but with AI it's now absolutely possible for companies to learn and brands to think.
I pitched a "procedurally-generated" view of brand to a major brand about two years ago, inclusive of a series of systems (now agents) that could help scan the horizon. They were excited, but the inertia proved too much. I feel so validated/less insane by your thinking!