
by Liam O'Dowd in
5 min read
In 2028, the world looks back in shock at a financial crisis caused by an AI-driven loop with no natural brake. That’s the story according to a fictional investor memo that caused investors to panic, but the question it asks is real: What happens to brands when the consumer disappears?
You may not have heard of Citrini Research, but you might have read the stories about the recently published Substack featuring a fictional memo written in 2028. In it, analysts take a retrospective look at a financial crash caused by increasingly productive and efficient AI.
Part prediction, part post-apocalyptic fanfic, the viral article caused actual ripples in the US stock market: the Dow dropped by 1.7 percent, the S&P 500 by 1 percent and the NASDAQ by 1.1 percent.
In this chilling glimpse from the near future, Citrini talk about the “Human Intelligence Displacement Spiral” – a negative feedback loop where AI agents have taken over much of the economy, especially in places where brands have monetised friction.
While the briefing reads like something from a sci-fi novel, and the story was widely rebuked, brands still need to pay attention – it was powerful enough to spook the financial markets after all. Brands exist because we don’t want to make decisions or are easily influenced. We pay a premium to avoid thinking. Agents don't. What does the future look like for these brands in Citrini’s world?
According to the memo, the first cracks begin to show with online food delivery apps like DoorDash. Instead of scrolling a list of algorithmically influenced restaurants and takeaways, consumers begin to opt for vibe-coded, personalised AI agents that find restaurants based on their specific tastes and price points, ordering directly from the business, eliminating the marketplace from the loop. Instead of a human forming habits with their platform of choice, an agent checks all the nearest businesses and orders from the cheapest and fastest. “Habitual app loyalty, the entire basis of the business model, simply didn’t exist for a machine,” the Citrini article says, reflecting on the death of similar platforms.
Other platforms like travel booking, insurance aggregators, and financial services providers soon follow. Agents are built on the supply side to serve the agents working on behalf of customers. “Once agents controlled the transaction, they went looking for bigger paperclips,” the report states, in a nod to the famous thought experiment. With price matching and aggregating optimised, the machines set out to eliminate fees, shifting from cards to crypto. The moats made from friction fall, with no natural brake, AI scales and scales, and before long, almost every decision a consumer makes is handed over to an agent.
While the internet disrupted and left new opportunities in its wake, the Agentic AI driven future painted by Citrini leaves nothing behind for humans. Countless brands exist because we are cognitively lazy, temporally constrained, and socially influenced. We are happy to pay a premium to platforms and apps to do our thinking and searching. We choose brands not just on price but on our cultural attachment – we know what a Nike person likes, what an Apple person signals. Our choice of supermarket says more about us than what is in our basket. So what happens when we hand over those choices to robots working on our behalf?
When McCann Erickson’s Bill Backer assembled 1,200 young adults in Rome to film Coca-Cola’s iconic ad “I'd Like to Teach the World to Sing”, the we was clear. Riding the wave of globalism and tapping into the counter-cultural spirit of togetherness, the brand cemented its place as the world's leading soft drink through loyalty to the idea of global harmony.
In Citrini’s 2028, there is no we. Brand loyalty is just another layer to be optimised away by agents that don’t care about the cultural signals from your choice of jacket or the subtle nods printed on your shopping bag. It doesn’t care about the app you navigate to subconsciously when you're hungover. Every choice becomes a parameter: price point, delivery speed, colour, size. All of the codes brands have built through advertising, design and psychology mean nothing to the robots.
In this world, brands risk losing meaning. Optimisation becomes inherently individual; agents maximise your outcome based on your parameters. The brands that survive Citrini’s 2028 build signals for machines, not humans. The we no longer exists. No water cooler moments, no cultural consensus about the signals, indicators and values brands project. Just an endless race to optimise data for agents that no longer interact.
Brands aren’t perfect, in fact, they’re manipulative. Designed to create an artificial demand to extract as much capital as possible. But they’re also part of our human infrastructure. They’ve become part of our shared, global reality. They let us communicate "I'm the type of person who..." without having to articulate the entire worldview that implies. If agents strip away brand selection from human choice, they remove a vocabulary we use to understand each other.
Even in the reality of 2026, we’ve already automated and optimised so much, but what do we lose when we automate the domain of choice? Humans are naturally inefficient. We are susceptible to a well-told story. Our rational brains are easily swayed by a smell, a memory or a guilty pleasure. The mere waft of coffee can make us beeline to our favourite caffeine dealer. Childhood nostalgia can have us picking up a box of sugary cereal even though we’re mid-diet. An expensive meal looks attractive when we’re budgeting because we’ve had a tough day. No agent can understand these uniquely irrational human choices.

generic products
Brands know that we’re susceptible to these flawed decisions, and they exist to monetise them. We become ourselves partly through what we choose to buy, wear, consume, and surround ourselves with. Not because products have inherent meaning, but because choosing them is a form of self-authorship, influenced by the brands that make them. When commerce no longer needs to account for these beautifully human quirks, when the economic system can route around our nostalgia, our aesthetics, our animal brain susceptibility, what happens to the infrastructure that creates and maintains these associations? If brands can't monetise our emotional resonance, they'll stop building it. If advertising can't sell stories to humans because agents make the purchasing decisions, the cultural production around consumption disappears. We might still have products perfectly optimised for our needs. But we'll have lost the apparatus through which products became meaningful to us.
The irony here is that in this world, we move towards the purest form of free market economics in our textbooks. Perfectly informed consumers, maximally efficient, untethered from social influence. Except it won't be us. It'll be our agents. And we'll be sitting behind them, no longer sure of who we are because we've outsourced the entire system in which our wants and needs are fulfilled.
Citrini’s memo is a viral think-piece based on a speculative future, not a roadmap to certain disaster. It's deliberately dystopian fiction, defined only by its viral moment. But it would be foolish to dismiss it in its entirety because the future it describes is plausible, and it’s close. Agentic commerce is already here, and while this exact 2028 might never arrive, something similar is inevitable.
What do we lose in this future, and can we notice before it’s too late? The entirety of modern branding is built around persuading humans to make a choice at the point of decision, but what does branding look like when humans hand over some or all of that choice to machines? What are brands for when they’re detached from our culture by agents?
Coca-Cola’s 1971 ad worked because there was a world to teach. A globally shared cultural moment enjoyed by a we that could easily be addressed. In 2028, there might not be a we, just billions of personalised agents humming to a different tune that none of us can hear.