Why AI Can't Write Strategy (And What That Means for Brands)
Generative AI can produce a competent brand insight in seconds, but so can your competitor, from roughly the same prompt. Large language models optimise for the average, which means everyone converges on the same middle-of-the-road strategy. That isn't strategy anymore; it's decoration. Real strategic value now lives in what a machine can't do: decoding how culture actually works, the unwritten codes, rituals and gatekeepers that were never written down for a model to read. At Home, we build strategy from that cultural understanding, because a position earned inside a culture can't be bought with media budget or prompted into existence.
Generative AI has turned our industry on its head, not only creation and production, but strategy too. Give a model like Claude a decent brief and it returns a reasonable insight in seconds. Clients see that and reasonably ask whether the work can be done in less time. It can. But that creates a problem.
The generated insight is fine. The trouble is that a competitor feeding in roughly the same input arrives at roughly the same insight. LLMs search for the average, so the odds rise that everyone lands on the same average strategy. And you can’t really call that strategy. It’s decoration, a commodity with no added value. This is going to happen. Clients push on budgets, and that pressure pushes agencies to surrender to the machine.
So the question we asked ourselves was simple: how do you build a strategy a machine can’t build?
What the machine can’t do
The answer lies in what a model cannot reach: understanding how culture is structured and how it shapes behaviour. Culture runs on codes that are rarely written down anywhere; who the gatekeepers are, which unwritten rules apply, what works and what feels forced. An LLM cannot decode what was never written.
That is where the value sits now. Strategy that begins by understanding the cultural contexts where people find shared meaning, form their identity, set their norms and shape their preferences, and then works out how a brand can become a credible part of that. From that understanding, brands can build meaning inside chosen cultural contexts and, over time, convert that meaning into economic value.
That sounds simpler than it is. It forces us to take two words seriously that our industry tends to throw around carelessly: meaning and culture.
Culture as a web of meaning
The anthropologist Clifford Geertz gave one of the most vivid answers to what meaning is: culture as a web; the networks of meaning through which people order their experience, built from symbols, rituals, shared stories and norms. Not something people consume, but something they live inside.
Take a football World Cup. Streets turn a national colour, friend groups gather to watch, conversations spill into the next morning, and for a month people feel part of something larger. Nobody consumes the World Cup. They live in it.
For a brand, that’s where the value is. The web of culture moves slowly and sits deep, and that makes it defensible. A brand that earns a credible place there builds mental associations that outlast the next trend, a position a competitor can’t simply take over. No media budget can buy that value, and no model can generate it.
Why this is defensible in a way AI outputs never are
In a world where most marketing is forgotten within a quarter, an earned cultural position holds lasting value. A competitor can’t casually take it, and they certainly can’t prompt it into existence. A claim can be copied by a rival brand in a day; credible embedding inside a culture takes years.
That’s the real shift. As AI drives the cost of average insight to zero, the premium moves to the one thing it can’t replicate; the slow, earned, human work of becoming a genuine part of the cultures your audience actually lives in.
The price is time. The reward is a position no competitor takes over, no media budget buys, and no model generates. That’s what we build at Home, for brands that see the difference, and that still want a place of their own in culture ten years from now.
FAQ
Can AI do brand strategy?
AI can generate a competent insight quickly, but because language models optimise for the average, competitors using similar inputs reach similar conclusions. That produces interchangeable, commodity-level strategy rather than a defensible strategic position. The lasting strategic value lies in cultural understanding that models can’t access.
What is cultural strategy?
Cultural strategy is the practice of earning a credible place for a brand inside a specific cultural context, its codes, rituals and gatekeepers, and converting that earned meaning into long-term brand growth. It’s built on understanding, not on media spend, which makes it difficult for competitors to copy.
Why can’t a language model decode culture?
Culture runs on unwritten codes, who the gatekeepers are, which rituals matter, what feels authentic versus forced. Because these were never written down, there’s no text for a model to learn from, so an LLM can’t decode them.
Is AI useless for strategists then?
No. AI is a strong tool for speeding up research and generating baseline material. The point is that it can’t be the source of differentiation, because everyone has access to the same average output. Differentiation comes from cultural understanding layered on top.
