The world is being rebuilt on top of artificial intelligence. In a few short years, AI has become the layer that reasons, writes, codes, plans, and increasingly decides. It can do almost anything, except one thing. It does not actually know us. It has read everything humanity has ever written and still cannot tell you what a real person, a specific person, will do tomorrow.
That is the missing component. Every AI system being built right now is making decisions about human beings while being structurally blind to them, working from averages, assumptions, and the thin exhaust of online behaviour rather than from real, lived, validated human truth.
We started Brox to build that missing layer.
We believe every decision an AI system makes, and soon that will be most of the decisions in the world, is a better decision when it is informed by real humans. And the only way to do that at the scale AI now operates is to build high-fidelity digital twins of real people, replicas grounded in hours of real human conversation, who represent those people faithfully enough to stand in for them. Not data about humans. Humans, in the loop, at scale.
That is the human-truth layer. It is the part of the AI stack everyone is going to need, and it does not exist yet. So we are building it.
The future we are building
We are building the layer that keeps the age of AI tethered to reality.
Picture the next decade honestly. AI will design the products, price the services, write the campaigns, shape the medicines, and draft the policies. The intelligence to act is becoming abundant and almost free. But intelligence without understanding is just confident guessing at machine speed, and a world that decides faster than it understands is a world that gets things catastrophically wrong, faster.
The bottleneck of the AI era will not be how cleverly machines can think. It will be whether what they think is grounded in how real people actually feel, choose, and behave. That grounding is what we provide. We are building a world where any AI system, any company, any researcher, any agent, can consult a faithful, living model of real human beings before it acts, and get a true answer in minutes rather than the months and millions that human understanding has always cost.
Not a smarter survey. Not a language model guessing on your behalf. Genuine, validated, 1:1 replicas of real people, built from hours of real conversation, who reason the way those people reason and choose the way they choose. Ask them anything. Test a thousand ideas instead of three. Understand the one person behind the number, not just the number. Model how a recession, a price change, or a life event actually shifts behaviour, before it happens, before a single real person is touched by the consequences.
This is how human understanding stops being a luxury and becomes infrastructure. Our ambition is not modest and we will not pretend it is: we intend to build the human-understanding layer for the entire age of AI, the place every important decision on earth comes to be checked against reality before it is made. Tens of thousands of digital twins today, millions tomorrow, eventually a faithful model of humanity itself, working quietly behind every system that matters. More is possible here than almost anyone dares to imagine, and we are going to build all of it.
Why we specifically are doing this
We did not arrive here by accident. We have spent years deep inside the research and insights industry, on the inside of the very problem we are now solving, and we watched the same thing happen on project after project: people making enormous decisions on a thin, slow, expensive sliver of human understanding, and everyone simply accepting that this was as good as it would ever get. We knew exactly what was broken, we knew why nobody had fixed it, and the moment AI became powerful enough to actually do something about it, we knew precisely what to build. Understanding people is the oldest and hardest problem there is, and for the first time in history the tools exist to crack it at scale. Brox is us doing exactly that, every single day.
We are a small team that operates like a much larger one, spread across the US, the UK, Europe, Turkey, Japan and beyond, building digital twins of tens of thousands of real people and putting them to work for some of the most important companies in the world. We move fast because we believe speed solves problems that caution cannot. We trust evidence but we are not seduced by it, we have learned that the gut, when it is properly informed, is its own kind of intelligence, and that the consensus in the room is very often just the loudest guess wearing a suit.
We are doing this because the human-understanding layer of the AI age is going to be built by someone, and it matters enormously who. We would rather it was built by people who treat the humans behind the twins as something sacred, and who are slightly terrified of getting it wrong.
What we believe
We believe that in a world run on AI, the scarcest and most valuable thing will not be intelligence. It will be a true understanding of human beings, and whoever holds that faithfully holds the key to every good decision made from here on.
We believe real people are infinitely more nuanced and interesting than the averages they get reduced to, and that this nuance is exactly what everyone else throws away and we keep.
We believe understanding should be measured, not assumed. If we say a twin predicts a real human’s choice, we can show you the accuracy, because a claim you cannot validate is just a nicer-sounding guess, and the AI era will drown in those.
We believe speed is a moral good. Every week a good decision is delayed is a week of worse outcomes for real people, so getting to the truth faster matters more than almost anyone admits.
We believe in being trustworthy above being clever. Overconfidence we can forgive, in ourselves and in anyone else, but dishonesty about what our technology can and cannot do is the one line that does not come back.
And we believe the point of all of this is simple, to keep the human in the loop as the world automates, and to leave it a little better understood, and therefore a little better served, than we found it.
What we refuse to do
We will not pretend our digital twins are the real humans they are built from. They are the most faithful representation of real people ever built, and still a representation, and we will always be honest about where the edges are.
We will not let a decision that seriously affects someone’s life be handed entirely to a machine. However capable AI becomes, decisions about real people demand a real person who is accountable for them. We build to put human understanding inside the loop, never to remove the human from it.
We will not sell certainty we cannot stand behind. We would rather tell you what we do not yet know than dress up a weak answer as a strong one, because the whole value of a truth layer is that you can trust it.
We will not treat the people whose lives power this as data points. The humans behind every twin are represented with consent, care, and respect, because the entire thing falls apart, morally and practically, the moment they are not.
And we will not pretend the rise of this technology is painless. We are building it with our eyes open to what it means, because the people who should be trusted to build something this powerful are the ones who are honest about its weight.
What this means for our customers
It means you stop guessing, and you stop letting your AI guess too.
It means the question that used to take six weeks and a six-figure budget takes an afternoon, so you can ask the braver questions, test the wilder ideas, and kill the bad ones before they cost you anything real. It means every model, campaign, product, and strategy you build can be checked against how real people will actually respond, before you ship it. It means understanding not just what your customers will do, but why, person by person, with the reasoning laid bare, testing a thousand concepts as easily as five, and modelling the future instead of only measuring the past.
You still make the call. That is the point. We just make sure that you, and the AI systems you increasingly rely on, are deciding with the clearest picture of real human behaviour anyone has ever had.
What this means for our team and our investors
We are all going to become fabulously rich.
We are not going to be coy about it. If you build the layer that every AI decision in the world needs, and you build it first and you build it right, the upside is frankly absurd, and we would be lying (which we don’t do) if we pretended that wasn’t part of the appeal.
But the truth is simpler than that. We are doing this because it is the most interesting problem any of us has ever been let loose on, with people we actually like, on something that matters. So if we are right, we all end up somewhere with a very good view and a faint look of disbelief on our faces. And if we are somehow wrong, we will have spent these years building something gloriously audacious with brilliant people and having an absolutely tremendous time doing it.
What this means for the world
If real human understanding becomes fast, cheap, and accurate, then the most powerful technology ever built grows up tethered to the people it serves, rather than floating free of them.
This is the difference between an AI future that is done to people and one that is built around them. Medicines shaped around how patients actually live. Products built around what people actually need rather than what a boardroom or a model assumed. Services, policies, and ideas tested against real human reality before they are inflicted on real human lives. The waste of building the wrong thing, the cost of being misunderstood, the widening gap between what powerful systems think people want and what people actually want, all of it starts to close.
That is the world we are building toward, one where the most important component of the AI age is not the intelligence, but the humanity it is grounded in, and where the people on the receiving end of every big decision are finally understood before the decision is made, not after.